Feed for discussion General in project Filesystems.

post122402: Re: Distributed file system in QNX

2023-10-17T11:41:28Z

No it's not a cluster FS, I'm not aware of us having a cluster fs available on QNX(someone else may know better).

So as I said I think nfs is the closest we have to your described use case, that is readily available.

post122401: Re: Distributed file system in QNX

2023-10-17T09:54:10Z

Thanks.

QNX ver is 7.1

NFS is the closest one, however, does it offer a back-up mechanism - means primary storage data is synchronized with another secondary backup in real time?

post122400: Re: Distributed file system in QNX

2023-10-17T09:24:31Z

You should specify for which qnx version.

nfs is probably the closest to what you seem to want.

http://www.qnx.com/developers/docs/7.1/#com.qnx.doc.neutrino.user_guide/topic/fsystems_NFS.html

Though it likely depends on your security requirements are. nfs isn't what I would call secure.

post122398: Distributed file system in QNX

2023-10-17T08:33:26Z

This is to check availability of filesystem in QNX that can be accessed from multiple nodes in a distributed environment and store data to a primary storage device that is connected to anyone of the nodes.

Additional requirement is to store data to a backup storage device that is connected to another node ( than primary storage) that can synchronise and replicate data from primary storage simultaneously.

Do we have support of this use case in any of the QNX supported filesystems? Thanks for your response.

post121891: Re: fs-qnx6 and QNX 6.5

2022-08-01T12:46:41Z

Or the system disk is not yet mounted when mounting qnx6 filesystem so /lib/dll is not accessible.

Good to know you solved your problem.

post121890: Re: fs-qnx6 and QNX 6.5

2022-08-01T12:43:48Z

yes, it was set in the boot image:

LD_LIBRARY_PATH=/proc/boot:/usr/lib:/lib:/lib/dll:.

so I can't figure out what's wrong .. maybe the env. variable is not accessible when I run

the mount ?

...anyway, i've solved including the library fs-qnx6.so in the boot image....now it works fine

Mario

post121889: Re: fs-qnx6 and QNX 6.5

2022-08-01T12:32:07Z

Hi,

Is /lib/dll listed in LD_LIBRARY_PATH ?

post121876: fs-qnx6 and QNX 6.5

2022-07-18T13:36:03Z

Hi all

recently I've try to use a qxn6 fs on a  QNX6.5

the device (/dev/hd1t179) has been formatted using qnx6 ff:

mkqnx6fs /dev/hd1t179               all ok

Try to mount:

mount -tqnx6 /dev/hd1t179  /mnt/qnx6

mount: cant'mount /mnt/qnx6 (type qnx6)

mount: possible reason: can't access shared library

note that the fs-qnx6.so is on system disk /lib/dll.

What's is wrong?

I do not see any messages explaining the real trouble here...

Thank

Mario

post121661: UUID

2021-12-15T11:59:49Z

Hello,

Is it a fuction to call from C/C++ or any other way to generate a UUID for qnx.

I have notice the create_uuid /target/qnx7/usr/include/uuid.h but have problems using it for programs I try to create.

post121572: Re: QNX 6 power fail-safe file system and compact flash cards

2021-08-25T01:01:00Z

We also observed some failure by using EMMC and QNX6FS.

We made a t178 partition and use qnx6fs mounting on boot. Then we only write SLOG and DUMPER to it. When we do power-on/power-off test on it, low percentage the FS corrupt: when next time boot-up, OS can not mount it successfully. Then we manually umount/mount it, the FS is only read-only. After we format it and remount, all are OK but data lost.

Anything wrong?

By Using RENESAS R-CAR GEN2.

Best regards.

Johnson

post121471: Re: QNX4 fs using QNX7.1

2021-06-29T12:33:23Z

Thanks, John: I know that qnx4 has been discontinoed, but my problem is to

have access to qnx4 partition using a qnx7.1 system, for legacy purposes.

Copy a qnx4 data over a qnx6 fs using a 6.5 system should be a problem,

I can do it, but is should be complicated to do in some installations

Compatibility does not seem a priority issue for qnx :-)

Mario

post121470: Re: QNX4 fs using QNX7.1

2021-06-29T08:54:26Z

QNX4 fs support was removed in QNX7.0 back in 2017. I don't see it getting added back.

http://www.qnx.com/developers/docs/7.0.0/#com.qnx.doc.qnxsdp.migration/topic/filesystems.html

QNX6 fs is supported by 6.5, 6.6, 7.0. 7.1

http://www.qnx.com/developers/docs/6.5.0SP1.update/index.html#com.qnx.doc.neutrino_utilities/f/fs-qnx6.so.html

Alternatively FAT32, EXT2, and NTFS are also possible.

http://www.qnx.com/developers/docs/6.5.0SP1.update/com.qnx.doc.neutrino_sys_arch/fsys.html#DOSFSYS

http://www.qnx.com/developers/docs/6.5.0SP1.update/com.qnx.doc.neutrino_sys_arch/fsys.html#NTFS

http://www.qnx.com/developers/docs/6.5.0SP1.update/com.qnx.doc.neutrino_sys_arch/fsys.html#EXT2

post121469: QNX4 fs using QNX7.1

2021-06-29T06:47:17Z

Hi, I'm workink on a system that uses both 6.5 and 7.1 versions.

For legacy purpose we have a QNX4 formatted partition on a sd card shared among

6.5 and 7.1.

The problem is that in the 7.x version the QNX4 fs is no more supported: why?

I've not found any way to access a QNX4 partition from 7.1 ...

is it possible to have back the support for qnx4 also in 7.x?

Thanks

M. Sangalli

post120859: Re: Filesystems source code request

2020-07-22T14:07:01Z

Hi,

this is the page we use: http://theloop.ott.qnx.com/departments/marketing/source_request/index.html

Don

post120420: Re: Creating Writable filesystem in qnx6.5

2020-03-31T10:04:41Z

Why is your OS readonly ?

post120256: Re: Creating Writable filesystem in qnx6.5

2020-02-13T07:49:15Z

To create a new user/password, you have to redirect /etc/passwd and /etc/shadow to a writeable location.

More generally, when you need to change system settings :

- Copy /etc (the full content) to a writeable location.

- Redirect /etc to this location.

- Do the modifications you need.

- Copy modified files into your BSP sources.

- Regenerate BSP.

- Reboot with new BSP.

This creates a "static" configuration used by all machines using the BSP.

If you need to be able to modify system settings dynamically, create (a) link(s) in your BSP init script to automatically redirect necessary files/directories at startup.

post120250: Re: Creating Writable filesystem in qnx6.5

2020-02-12T10:39:44Z

Thanks for your prompt answer.

I Already redirected the key files in sshd_config file to '/fs0p1/ssh' folder, and the ssh keys(rsa, dsa) created in this folder which is writable. ssh client can now access to the session but because i cant be able to set a password (for ssh) i am not able to login to the target.

It would be greatly appreciated if you give me a hand

thanks in advanced,

post120248: Re: Creating Writable filesystem in qnx6.5

2020-02-11T10:28:57Z

To be able to do this, no need for a writeable filesystem.

Redirect /etc/ssh/ to a writeable area (USN key, SD card, /var ...).

Generate the required files. Copy these files in your BSP and regenerate your BSP..

You can then login with SSH.

If you want dynamic management of SSH, you have to permanently redirect /etc/ssh to a writeable media. Permanently means add the redirection in the boot sequence.

Nicolas

post120246: Why QNX7 does not see all the RAM of the PC?

2020-02-10T14:48:59Z

I use ADLE3800PC with the following parameters:

CPU: Intel Atom E3827 2 ядра, 1,75 GHz, 1MB;  memory: 4GB DDR3L.

I run the trial OS QNX7.0 on this PC.

I am using BSP_x86_br-700_be-700_JBN89, the boot file is x86-generic.bin.

I managed to start the OS. Standard applications gles1-gears, gles2-gears, gles3-gears,

sw-vsync are also start. My heavier qt applications do not start.

After looking at the free memory using the pidin info command:

-------------------------------------------------------------------------------------------

# pidin info

CPU:X86 Release:7.0.0 FreeMem:413MB/495MB BootTime:Apr 01 04:32:49 GTM 2019

   Actual resident free memory:418Mb

Processes: 29, Threads: 121

Processor1: 197294 Intel 686 F6M1JS9 1749MHz FPU

Processor2: 197294 Intel 686 F6M1JS9 1749MHz FPU

-------------------------------------------------------------------------------------------

I saw that QNX does not see all the PC RAM (496 MB instead of 4 GB).

The same problem with Kontron CP3003-SA, with parameters:

CPU: Core i7  4 ядра,  2,1 GHz, 4MB; memory - 8GB (2x 4GB) SODIMM DDR3 with ECC (instead of 8GB QNX saw only 511MB).

With boot file from BSP_intel-x86_64-APL-abl_br-700_be-700_JBN89, it was not possible to start QNX7 on these platforms, an Invalid Image OS error occurs.

How to make QNX see the full amount of RAM?

post120245: Creating Writable filesystem in qnx6.5

2020-02-09T12:30:02Z

I am trying to enable SSH login, but as far as i found on the net, for this purpose i need a writable file system e.g. RW access to "/etc/ssh/" but my built OS image is read-only.

would you please help me to build a writable file system inside QNX Neutrino 6.5

Thanks in advanced.

post120101: Filesystems source code request

2019-12-11T07:59:13Z

Hi,

Any way to get access to the sources ?

Regards,

Nicolas

post119993: QNX trusted disk

2019-09-26T09:09:24Z

Has any one experience with creating a QNX trusted disk image and mounting it.

Can you please share some steps to begin with.

I have a qnx 7 environment installed in vmware virtual box (x86_64).

post118836: Re: RE: QNX 7 Ramdisk Caching Question

2018-05-19T19:21:25Z

David,

Thanks for looking into the buffer size.

Maybe the size hasn't changed but the write behavior has in some subtle way. Data would not get hung up in the stdio buffer forever and never flush. At some point it was getting written even if it was a few seconds delay. Now it remains in there forever until the buffer fills up. We've been using this logging behavior for over 10 years and there is no way we wouldn't have noticed it.

Regardless, I'll add that line into our code. It won't be at program start but rather will be in a class that is responsible for logging. That may mean it gets set a few dozen times per application due to class instantiation but I assume that's not a problem.

Tim

post118835: RE: QNX 7 Ramdisk Caching Question

2018-05-19T18:26:15Z

The stdio buffer size is 1024 bytes, and hasn't changed since at least 6.3.0.

________________________________

From: David Sarrazin

Sent: Friday, May 18, 2018 6:03 PM

To: general-filesystems

Subject: Re: QNX 7 Ramdisk Caching Question

Tim,

That sounds more like fwrite buffering, so MyApp isn't even sending the data to the FS. The redirect causes studio to become fully buffered, which means it will only flush to the filesystem once the internal stream buffer is full, or an fflush() is called. This is POSIX behaviour so will not have changed since 6.3.2. I'll check if the buffer size changed between that version and 7.0.

If you can change the code, adding

setvbuf(stdout, NULL, _IOLBF);

In your program startup will force stdout to flush after every newline, instead of when the buffer is full.

David

Sent from my BlackBerry - the most secure mobile device

From: community-noreply@qnx.com

Sent: May 18, 2018 5:27 PM

To: general-filesystems@community.qnx.com

Reply-to: general-filesystems@community.qnx.com

Subject: Re: RE: QNX 7 Ramdisk Caching Question

David,

Thanks for the fast reply!

> These include deferring I/O to the "disk" as long as possible.

Is there a way to turn this off? I'd like to get the same behavior I have under 6.3.2. I only used all the extra options in the wiki page because it said it was the fastest way to write data. I don't mind slightly slower writes (it is RAM after all so it's already lightning fast) if they occur immediately.

I can 100% reliably produce the problem where data is not available in the RAM log.

Normally I start my app as ./myApp >/fs/ram/myLogFile to redirect stdout to log file.

If I just start ./myApp I see everything printed immediately as the code executes.

If I do the redirect to the log file on the RAM disk I do not get anything in the file until a certain amount of data is logged. It appears at a glance to be probably pretty close to that 512 value (bytes?). So roughly every 512 characters sent to the file tends to get the next block write to occur to update the file.

So would changing that cache size from 512 matter or is there some other option to look at?

Tim

> Tim,

>

> devb-ram looks and behaves like a real disk, as far as the io-blk layer and

> filesystem are concerned. The io-blk internal ramdisk which you are now using

> , is understood by the io-blk caching layer to be a transient RAM disk, and

> some optimizations are made. These include deferring I/O to the "disk" as

> long as possible.

>

> With that said, if you write to a file, and then read the file back, you

> should absolutely see the data that you wrote, immediately. The power-safe

> filesystem itself still does not realize that the underlying device is a RAM

> disk, and doesn't take any special action.

>

> David

_______________________________________________

General

http://community.qnx.com/sf/go/post118833

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com<mailto:general-filesystems-unsubscribe@community.qnx.com>

post118834: Re: QNX 7 Ramdisk Caching Question

2018-05-18T22:22:26Z

Tim,

That sounds more like fwrite buffering, so MyApp isn't even sending the data to the FS. The redirect causes studio to become fully buffered, which means it will only flush to the filesystem once the internal stream buffer is full, or an fflush() is called. This is POSIX behaviour so will not have changed since 6.3.2. I'll check if the buffer size changed between that version and 7.0.

If you can change the code, adding

setvbuf(stdout, NULL, _IOLBF);

In your program startup will force stdout to flush after every newline, instead of when the buffer is full.

David

Sent from my BlackBerry - the most secure mobile device

From: community-noreply@qnx.com

Sent: May 18, 2018 5:27 PM

To: general-filesystems@community.qnx.com

Reply-to: general-filesystems@community.qnx.com

Subject: Re: RE: QNX 7 Ramdisk Caching Question

David,

Thanks for the fast reply!

> These include deferring I/O to the "disk" as long as possible.

Is there a way to turn this off? I'd like to get the same behavior I have under 6.3.2. I only used all the extra options in the wiki page because it said it was the fastest way to write data. I don't mind slightly slower writes (it is RAM after all so it's already lightning fast) if they occur immediately.

I can 100% reliably produce the problem where data is not available in the RAM log.

Normally I start my app as ./myApp >/fs/ram/myLogFile to redirect stdout to log file.

If I just start ./myApp I see everything printed immediately as the code executes.

If I do the redirect to the log file on the RAM disk I do not get anything in the file until a certain amount of data is logged. It appears at a glance to be probably pretty close to that 512 value (bytes?). So roughly every 512 characters sent to the file tends to get the next block write to occur to update the file.

So would changing that cache size from 512 matter or is there some other option to look at?

Tim

> Tim,

>

> devb-ram looks and behaves like a real disk, as far as the io-blk layer and

> filesystem are concerned. The io-blk internal ramdisk which you are now using

> , is understood by the io-blk caching layer to be a transient RAM disk, and

> some optimizations are made. These include deferring I/O to the "disk" as

> long as possible.

>

> With that said, if you write to a file, and then read the file back, you

> should absolutely see the data that you wrote, immediately. The power-safe

> filesystem itself still does not realize that the underlying device is a RAM

> disk, and doesn't take any special action.

>

> David

_______________________________________________

General

http://community.qnx.com/sf/go/post118833

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com<mailto:general-filesystems-unsubscribe@community.qnx.com>

post118833: Re: RE: QNX 7 Ramdisk Caching Question

2018-05-18T21:46:02Z

David,

Thanks for the fast reply!

>  These include deferring I/O to the "disk" as long as possible.

Is there a way to turn this off? I'd like to get the same behavior I have under 6.3.2. I only used all the extra options in the wiki page because it said it was the fastest way to write data. I don't mind slightly slower writes (it is RAM after all so it's already lightning fast) if they occur immediately.

I can 100% reliably produce the problem where data is not available in the RAM log.

Normally I start my app as ./myApp >/fs/ram/myLogFile to redirect stdout to log file.

If I just start ./myApp I see everything printed immediately as the code executes.

If I do the redirect to the log file on the RAM disk I do not get anything in the file until a certain amount of data is logged. It appears at a glance to be probably pretty close to that 512 value (bytes?). So roughly every 512 characters sent to the file tends to get the next block write to occur to update the file.

So would changing that cache size from 512 matter or is there some other option to look at?

Tim

> Tim,

>

> devb-ram looks and behaves like a real disk, as far as the io-blk layer and

> filesystem are concerned.  The io-blk internal ramdisk which you are now using

> , is understood by the io-blk caching layer to be a transient RAM disk, and

> some optimizations are made.  These include deferring I/O to the "disk" as

> long as possible.

>

> With that said, if you write to a file, and then read the file back, you

> should absolutely see the data that you wrote, immediately.  The power-safe

> filesystem itself still does not realize that the underlying device is a RAM

> disk, and doesn't take any special action.

>

> David

post118832: RE: QNX 7 Ramdisk Caching Question

2018-05-18T21:27:59Z

Tim,

devb-ram looks and behaves like a real disk, as far as the io-blk layer and filesystem are concerned. The io-blk internal ramdisk which you are now using, is understood by the io-blk caching layer to be a transient RAM disk, and some optimizations are made. These include deferring I/O to the "disk" as long as possible.

With that said, if you write to a file, and then read the file back, you should absolutely see the data that you wrote, immediately. The power-safe filesystem itself still does not realize that the underlying device is a RAM disk, and doesn't take any special action.

David
________________________________________
From: Tim Sowden(deleted) [community-noreply@qnx.com]
Sent: Friday, May 18, 2018 5:14 PM
To: general-filesystems
Subject: QNX 7 Ramdisk Caching Question

I have a question about how devb-ram and caching interact with the power safe filesystem.

Here's some background:
On our 6.3.2 systems we create a RAM drive with the following commands:
devb-ram ram capacity=100000 &
mount -t qnx4 /dev/hd1t77 /fs/ram

to create a 50 meg RAM drive where we write run-time log files (not intended to survive reboot obviously). When our applications write to a log file on the RAM drive it gets placed there immediately (seen via editing / catting the file).

Now on QNX 7 after reading this
http://www.qnx.com/developers/docs/7.0.0/#com.qnx.doc.neutrino.utilities/topic/d/devb-ram.html
we are creating a RAM drive with the following commands:
devb-ram ram capacity=1 blk ramdisk=50m,cache=512k,vnode=256 &
mkqnx6fs -q /dev/ram0
mount -t qnx6 /dev/ram0 /fs/ram

However now when our apps write to the log files there is often nothing there immediately. It seems to take a lot of data pushed to the file in order to see earlier writes appear which makes the log file unreliable because stale data may not be written ever (even killing the app doesn't flush writes to the file).

So some kind of caching is now occurring which I want to eliminate. Is it the cache=512 (that says it's supposed to be the min to io-blk) or is it something in the power safe file system that I need to enable for the RAM drive?

Note that our apps don't call flush() (either in 6.3.2 or 7.0) after doing a write so it's not related to that or I hope QNX 7 doesn't require a flush() for every write to get immediate writes.

TIA,

Tim

_______________________________________________

General
http://community.qnx.com/sf/go/post118831
To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post118831: QNX 7 Ramdisk Caching Question

2018-05-18T21:14:52Z

I have a question about how devb-ram and caching interact with the power safe filesystem.

Here's some background:

On our 6.3.2 systems we create a RAM drive with the following commands:

devb-ram ram capacity=100000 &

mount -t qnx4 /dev/hd1t77 /fs/ram

to create a 50 meg RAM drive where we write run-time log files (not intended to survive reboot obviously). When our applications write to a log file on the RAM drive it gets placed there immediately (seen via editing / catting the file).

Now on QNX 7 after reading this

http://www.qnx.com/developers/docs/7.0.0/#com.qnx.doc.neutrino.utilities/topic/d/devb-ram.html

we are creating a RAM drive with the following commands:

devb-ram ram capacity=1 blk ramdisk=50m,cache=512k,vnode=256 &

mkqnx6fs -q /dev/ram0

mount -t qnx6 /dev/ram0 /fs/ram

However now when our apps write to the log files there is often nothing there immediately. It seems to take a lot of data pushed to the file in order to see earlier writes appear which makes the log file unreliable because stale data may not be written ever (even killing the app doesn't flush writes to the file).

So some kind of caching is now occurring which I want to eliminate. Is it the cache=512 (that says it's supposed to be the min to io-blk) or is it something in the power safe file system that I need to enable for the RAM drive?

Note that our apps don't call flush() (either in 6.3.2 or 7.0) after doing a write so it's not related to that or I hope QNX 7 doesn't require a flush() for every write to get immediate writes.

TIA,

Tim

post118695: Re: UDF "filesystem limitation" documentation

2018-03-22T16:02:50Z

I checked with the developers, and fs-udf.so does support links. We'll update the docs.

post118693: Re: UDF "filesystem limitation" documentation

2018-03-22T15:57:08Z

For context:

This came up because QNX7 dropped support for the fs-cd.so driver.

We could use a loopback mounted ISO9660 w/ RR extensions as a nice readonly filesystem, containing symlinks, etc.

With fs-cd.so disappearing in QNX7, we looked at UDF, and saw that the docs said it didn't support symlinks, but in practice they seemed implemented and worked as expected.

Thus, you could augment this request to read:

(a) bring back ISO9960 in QNX7

and/or

(b) Document that QNX's implementation of UDF (read-only, etc.) supports symlinks.

Thanks!

post118690: Re: USB mass storage device and FS mount file matching

2018-03-21T15:02:41Z

I'm not sure if this is the right answer or not. I stumbled across this while trying to get USB devices to NOT automount. Look at /etc/mcd.mnt. This looks like the table that mcd uses to determine where to mount USB devices.

Raymond Balister

post118689: UDF "filesystem limitation" documentation

2018-03-21T14:56:14Z

Filesystem Limitations doc page - for at least QNX7 and QNX6.6 (*)

state that for filesystem UDF, that  Soft/Symbolic Links and Hard Links are not supported.

In fact, when exercising this feature, we find that QNX's UDF implementation does support hard and soft links.  Perhaps it was added some time ago, and they forgot to update the docs on it?

Can you update the documentation to either:

(a) reflect this support

(b) confirm that it is not supported, and state under what conditions it is not expected to work correctly.

Regards,

  Brian

(*)

http://www.qnx.com/developers/docs/7.0.0/#com.qnx.doc.neutrino.sys_arch/topic/fsys_Limitations.html

http://www.qnx.com/developers/docs/6.6.0.update/#com.qnx.doc.neutrino.sys_arch/topic/fsys_Limitations.html

post118688: [QNX 6.5.0] Security and STIGs - automounting CD-ROMS

2018-03-21T14:54:44Z

I'm working through applying STIGs to a QNX 6.5.0 system for a customer.  One of the requirements is to disable automounting of filesystems.  Disabling USB mass storage is another.  I've already been able to disable USB mass storage through the /etc/system/enum/devices/usb/mass config file.  Disabling the auto mounting of CD-ROMs seems to be a little more elusive.

The last item that I'll have to do is to remove the use of diskboot from the .ifs and use driver calls instead.  This is required due to diskboot giving the spacebar message.  Through that path you can boot to the file system (F5) without entering any credentials.  Big security hole.

Are there any suggestions on how to disable automounting of CD-ROMS.  I mention the .ifs part since it may be that I'll fix it by calling the drivers directly anyway.

Raymond Balister

post118682: imagefs checksum error

2018-03-18T13:22:10Z

smart-device-graphics.build

I added the following information in the file:

############################################################################

#Porting Qt

#Qt Libraries

[search=${QNX_TARGET}/external/Qt]

/Qt/bin/Demo=bin/Demo

/Qt/lib=lib

#Qt Dependent Libraries

[search=${QNX_TARGET}/armle-v7]

/Qt/lib/libpng16.so.0=usr/lib/libpng16.so.0

/Qt/lib/libicui18n.so.58=usr/lib/libicui18n.so.58

/Qt/lib/libicuuc.so.58=usr/lib/libicuuc.so.58

/Qt/lib/libicudata.so.58=usr/lib/libicudata.so.58

###########################################################################

###########################################################################

## END OF BUILD SCRIPT

###########################################################################

QNX-IFS file has 37M，there was an error "imagefs checksum error"

bsp\src\hardware\ipl\lib\image_2.c

#ifdef __ARM__

if (checksum_2(start + startup_hdr.startup_size, startup_hdr.stored_size - startup_hdr.startup_size) != 0) {

#else

if (checksum(start + startup_hdr.startup_size, startup_hdr.stored_size - startup_hdr.startup_size) != 0) {

#endif

ser_putstr("imagefs checksum error\n");

return (-1L);

}

post118580: Re: ext2 (for removable media)

2018-02-14T17:44:53Z

Too late for original poster, but for posterity for anyone else:

mkfs.ext2 -Onone -I128 ...

Will produce a QNX fs-ext2.so compatible filesystem.

Note that QNX7.x no longer ships fs-ext2.so

post117349: Repack non-bootable IFS problems

2017-01-19T08:34:33Z

Hi there.

I have a problem creating non bootable IFS image.

I've used dumpifs to dump the contents of non bootable ifs image.

Then i tried to pack it again using mkifs and parmeter file.

Here is part of it.

[prefix=]

[image=0x0]

[compress=2]

[dperms=0777 uid=0 gid=0]

[perms=1777]

[type=link] factories=usr/lib/factories

[data=copy code=uip] /usr/bin/chsum = d:\ifs-root2\usr\bin\chsum

/usr/lib/terminfo/terminfo.src = D:/ifs-root2/usr/lib/terminfo/terminfo.src

/usr/lib/terminfo/v/vt100 = D:/ifs-root2/usr/lib/terminfo/v/vt100

The image is created but when i use dumpifs to unpack it and compare the files bit by bit they are not the same :(

The regullar files (not executables or librarys) are not changed, terminfo.src and vt100 are 100% match.

Here is one file (chsum) to show the difference. chsum1 is is the file that i get when using dumpifs on the original image.

cksum2 is the file that i get when dumping the ifs that i've created based on the original one.

chsum1

https://mega.nz/#!9koywTRb!1s7Px_977I4Lt0-Z_uIH1rpU3YdCPGZnUy3ETmBC0DQ

chsum2

https://mega.nz/#!co5FiBZZ!MGlGPn8HN_YmJk6QTRNPAmHuVm4g5IyCBTVn5h4G0-Y

If i try to use [+raw] and set [perms=+x] then the files are not cahnged in content but the unit does not like them :(

Any ideas or suggestions are welcome.

Regards.

post117333: Mount compressed secondary ifs

2017-01-16T11:57:29Z

Hi all,

i would want to ask you how to remount secondary ifs as r/w.

When the the primary if boot its mounted in "/" but that is not visible in fstab and cant be remounted with mount -u.

Then the secondary ifs is mounted in "/" too and again not visible in fstab and can't be remounted.

This is harman-becker multimedia unit with nVidia Tegra running QNX.

I've tried to mount it by specifing the offset but that was not possible to or dont use the right parameters.

in /dev i see

fs0

fs0p0

fs0p1

fs0p2

fs0p3

ifs-root and secondary ifs as well as the IPL are on fs0p0.

If i try to rebuild that secondary ifs i get strange results,

i build the ifs successfully with mkifs, and then restore it.

If i compare the files one by one then i see that shared library and some of the executable files are not 100% match although i used the same file build the image. They are modified so the usage info is reversed.

If i specify [+raw] and use [perms=+x] then the files are 1:1 but they are not good for the unit and cant be started regardless of the fact that they are with 777 perms.

If you want more info i will provide.

Any help will be appreciated.

Kind regards.

post117204: nand flash driver

2016-11-24T07:54:59Z

Hello there,

Will BSP 6.4.0 in etfs file system is completely open source, I only see chipio.c, devio.c these two files, there are other files on the etfs file system?

Mr. Li

post116758: Re: mount EFS image

2016-09-08T11:47:55Z

Hi,

Did you found any solution for this. I am facing same issue. I have a .ifs image for my project however i want to create users when my os is running in Target board. I read on community that for writable files i have to create efs image. Now i have both image available but confused how to flash efs. I am using TFTP server mechanism to transfer image to my target board.

Please suggest me possible solution.

post116699: Does sopen support largefiles?

2016-08-29T15:29:07Z

I am developing under QNX 6.5.0. I would like to use both file sharing and large files. I have the compiler flags -D_FILE_OFFSET_BITS=64 and -D_LARGEFILE64_SOURCE. Does this give sopen large file support?

post116678: FFS3 question

2016-08-24T18:37:44Z

Hi everybody! I have a memory dump with FFS3. It's intresting how can i read that dump. I'm new to QNX (( I tried to work with QNX 6.5 on WMware and don't get any results. First of all i write my dump to usb flash without of any hope =) QNX didn't mount it. Next idea was to format usb flash as FFS3 and i failed again ( May be i should try to work with Momentics ? Or make a ram-disk? Any ideas ?

post115755: Re: NTFS formatted USB in QNX system (Automotive Radio) doesn't have write permission

2016-02-17T09:31:53Z

Thanks a lot Eric,

So currently our radio as Neutrino OS in it , to have a read/write permission work for NTFS, what can we do?

post115753: Re: NTFS formatted USB in QNX system (Automotive Radio) doesn't have write permission

2016-02-17T06:47:30Z

From the docs here:

http://www.qnx.com/developers/docs/am11/topic/com.qnx.doc.neutrino.sys_arch/topic/fsys_NTFS.html

---

Windows NT filesystem

The NT filesystem is used on Microsoft Windows NT and later.

The fs-nt.so shared object provides read-only access to NTFS disks on a QNX Neutrino system.

---

Regards,

Eric

post115752: NTFS formatted USB in QNX system (Automotive Radio) doesn't have write permission

2016-02-17T04:16:31Z

Hi,

Currently i am working for a infotainment radio which has qnx os in it.

When i insert USB with Fat32 format it has all the permissions(r,w,x) and when it is inserted with NTFS format it doesn't have the write permissions. can anyone please update on this. hardware is "am3517" (Texas Instruments)

Please provide your inputs on this. let me know if any further information required on this

devu-am3517-mg.so, io-usb

Thanks a lot in Advance

post115703: Re: FFS3 behaviour in case of power loss while writing

2016-02-09T17:01:26Z

Is there an answer to this?

I've encountered an event where a folder (where files get written randomly to) completely disappeared from the folder tree. The only thing I can see is that a flash file write access was interrupted by a power outage and screwed up the inodes or equivalent. Sucks that chkfsys is not available on FFS3.

post115360: Re: io-blk tuning guide?

2015-12-18T15:51:18Z

Attached is the output of fsysinfo when it gets into this state.

post115358: io-blk tuning guide?

2015-12-18T15:42:31Z

Is there a guide for tuning io-blk options for performance under various environments?  We have an embedded system that uses a 4GB Compact Flash card formatted with QNX 6 file system.  After the system runs, it creates a log file between 20KB and 100KB.  It keeps the last 30000 logs on the compact flash per customer requirements.  It needs to write the log in less than one second after the end of each run.  When running from powerup, the file write time is about .6 seconds.  If we access the system via the built in web server and read about 20 to 50 of these files, the write time jumps to 4 to 10 seconds for each file, even after we have closed the web session.

tracelogger is showing an inordinate amount of time in devb-eide eide_driver_thread, sharing the time with the idle thread.  Our thread that is performing the file write is mostly blocked in the Reply state.

I have tried various combinations of priorities and syncing with little improvement.  I have found that the io-blk cache setting has the most effect.  Currently, because of our system's memory requirements, the cache is set to 8MB.  When I boost this to an unacceptable 100MB, the problem mostly goes away.

So now I need to adjust the io-blk option to get me the smallest memory usage and still meet out requirements.  Unfortunately, the documentation does not give much detailed information about each option's effect and how they relate to the output of fsysinfo.  Is there any guidance on tuning io-blk?

post113857: Re: Booting from USB : Entire File System is READ ONLY

2015-05-25T10:44:39Z

Sounds like you booted qnx6 filesystem onf h/w that does not support "sync". Try this.option.

mount -tqnx6 -o sync=optional /dev/umass0t179 /

post113856: Booting from USB : Entire File System is READ ONLY

2015-05-25T09:41:46Z

Hi, I am booting a x86 system from USB. I don't have hard drive. My system works ok, i am able to telnet, qconn to the target.

But the entire file system is readonly. I am unable to create a directory or file.

Errors are: 1) Function not implemented, 2) Unable to find file

Please guide.

thanks, Manas

post113497: Re: Getting mcd to run for files found in a directory

2015-03-04T22:32:21Z

In my mcd config file I have the following:

-------------------------------

[/dev/umass*]

Callout     = PATH_MEDIA_PROCMGR

Argument    = /proc/mount

Priority    = 11,10

Start Rule  = MOUNT

Stop Rule   = UNMOUNT

[MOUNT]

Callout     = MOUNT_FSYS

Argument    = /etc/mcd.mnt

[UNMOUNT]

Callout     = UNMOUNT_FSYS

[/fs/usb*]

Callout     = PATH_MEDIA_PROCMGR

Argument    = /proc/mount

Priority    = 11,10

Start Rule  = FIRMWARE_IFS

[/var/spool/uploads/*]

Callout     = PATH_MEDIA_SCAN

Argument    = 1000

Start Rule  = FIRMWARE_IFS

[FIRMWARE_IFS]

Callout     = FNAME_PATTERN

Argument    = *.IFS,*.ifs

-------------------------------

There is a section that I am telling it if ifs found in that directory, call FIRMWARE_IFS. But why I cannot capture it at all?

Basically when I run

# cat /dev/mcd/FIRMWARE_IFS

and creating a sample ifs file:

echo "ers" > /var/spool/uploads/ds.ifs

FIRMWARE_IFS is not triggered!! Why?

post112602: Mounting QNX6 filesystem in Ubuntu - for both Read and Write

2014-12-10T22:40:18Z

Hi,

Is it possible to mount a qnx6 filesystem in ubuntu linux, for READ and WRITE?

I tried below cmd in ubuntu:

sudo mount -v -o rw -t qnx6 /dev/sdb1 /media/cf1

(also) -> sudo mount -o remount,rw /media/cf1/

Observed:

mount: warning: /media/cf1 seems to be mounted read-only.

/dev/sdb1 on /media/cf1 type qnx6 (ro)

Regards

Sheran

post112584: Re: QNX6 fs uses a lot of RAM?

2014-12-09T15:53:02Z

Doh, the missing comma was the problem. Thanks for the assistance.

post112579: Re: How to know the sector size of a disk programmatically

2014-12-09T13:39:47Z

I had a similar requirement to determine the amount of RAM, Disk size, and video modes. I wrote this test program to look for this information. The first part walks the memory descriptor of SYSPAGE and adds up the total RAM. Next I use statvfs64 to get the file system size. The third section is what you are interested in. It walks the hardware section of SYSPAGE looking specifically for filestore devices and printing info about them. The last section is using the BIOS INT10H to interrogate the VESA information.

post112578: Re: fs-qnx6 on a SSD

2014-12-09T13:21:23Z

I tried to get this same information some time ago. I found that many of the compact flash cards do meet the requirements but do not support the required commands, so QNX6 fs mounts them as read only unless you use the "sync=optional" argument. I did find that the Swissbit Industrial Compact Flash Cards do support the sync commands.

Note: I am in no way connected with Swissbit.

post112577: Re: QNX6 fs uses a lot of RAM?

2014-12-09T13:18:50Z

There needs to be a comma between the ramdisk and cache options.

Cheers.

Sent from my BlackBerry 10 smartphone on the Fido network.

  Original Message

From: Robert Murrell

Sent: Tuesday, December 9, 2014 8:14 AM

To: general-filesystems

Reply To: general-filesystems@community.qnx.com

Subject: QNX6 fs uses a lot of RAM?

At least I think its the culprit.  I'm developing under QNX 6.5.0.  Our system is an x86 PC-104 with 256 MB of RAM.  The devb-eide task is consuming 49MB of RAM.  The file system is on a 2GB Compact Flash.  I start the drivers as follows:

devb-eide quiet blk ramdisk=10m cache=2m qnx6 &

waitfor /dev/hd0

mount -t qnx6 -o sync=optional /dev/hd0t177 /

The ramdisk is occupying 10MB of this.  I've tried some of the memory saving tricks documented for io-blk but they did not seem to change this value.  I understand the nature of the the power fail safe system might use a lot of RAM, but is there any way I can control the RAM usage?

_______________________________________________

General

http://community.qnx.com/sf/go/post112574

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post112576: Re: QNX6 fs uses a lot of RAM?

2014-12-09T13:18:45Z

Use the "blk cache=" parameter to limit the amount of RAM utilized.

# devb-eide blk cache=2m

post112575: Re: QNX6 fs uses a lot of RAM?

2014-12-09T13:18:45Z

It's io-blk cache - it defaults to 15% of RAM IIRC. Check the docs for io-blk.

post112574: QNX6 fs uses a lot of RAM?

2014-12-09T13:14:54Z

At least I think its the culprit. I'm developing under QNX 6.5.0. Our system is an x86 PC-104 with 256 MB of RAM. The devb-eide task is consuming 49MB of RAM. The file system is on a 2GB Compact Flash. I start the drivers as follows:

devb-eide quiet blk ramdisk=10m cache=2m qnx6 &

waitfor /dev/hd0

mount -t qnx6 -o sync=optional /dev/hd0t177 /

The ramdisk is occupying 10MB of this. I've tried some of the memory saving tricks documented for io-blk but they did not seem to change this value. I understand the nature of the the power fail safe system might use a lot of RAM, but is there any way I can control the RAM usage?

post112281: Re: RE: Read QNX6 fs UUIDs

2014-11-06T00:36:47Z

That's great! Thank you very much. I'll just plug it right in to an mcd_content() hook, then. :)

-Will

post112280: RE: Read QNX6 fs UUIDs

2014-11-05T22:48:03Z

devctl(DCMD_FSYS_LABEL) will return a string containing a the UUID.

-----Original Message-----

From: Will Miles [mailto:community-noreply@qnx.com]

Sent: November-05-14 3:48 PM

To: general-filesystems

Subject: Read QNX6 fs UUIDs

Hi,

Is there any code or examples of how to read back the UUID from a QNX6 filesystem? I'm looking for a way to uniquely identify them early in our boot sequence without having to leave cookie files behind. I looked for utility that could do this, a function call, or documentation on the superblock format so I could do it myself, but I couldn't find anything obvious.

Thanks,

-Will

_______________________________________________

General

http://community.qnx.com/sf/go/post112279

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post112279: Read QNX6 fs UUIDs

2014-11-05T20:47:46Z

Hi,

Is there any code or examples of how to read back the UUID from a QNX6 filesystem? I'm looking for a way to uniquely identify them early in our boot sequence without having to leave cookie files behind. I looked for utility that could do this, a function call, or documentation on the superblock format so I could do it myself, but I couldn't find anything obvious.

Thanks,

-Will

post112154: QNX6fs and bad blocks

2014-10-24T23:17:55Z

The QNX4 file system had a facility to work around defects on a disk. You could run dcheck to mark the blocks as bad, the chkfsys to find out exactly how those bad blocks affected the file system, then zap etc. to repair.

Does the QNX6 file system have any such capabilities? It seems as though there is no recourse when hardware errors are encountered.

post111302: Error "unable to access /dev/hd0"

2014-08-05T03:40:36Z

System spec:-

QNX neutrino 6.5

p2041rdb freescale board

Westen Digital HDD 320 GB with 4 partitions

When the SATA driver is started by the following script , it notifies that the hardrive is detected

display_msg Starting SATA driver for SATA1 device...

devb-fslSata fslSata ioport=0xFFE220000,irq=97

waitfor /dev/hd0 10

but gives an error "unable to access /dev/hd0" and then after about 3-4 minutes it gives the notification - "Path =0- FSL SATA target=0 lun=0 Direct Access<0> - WDC WD3200BEKT-6 Rev:01.0"

.Only after this notification does the hd0 show up in /dev/ .

If i do fdisk /dev/hd0 show , it hangs and sometimes shows the 4 partitions and sometimes the list is blank with no details.How to fix this error??

Ken

post111252: Looking for MPC8349 BSP for USB features for 6.3.2

2014-07-29T16:56:18Z

Specifically, we require: devu-ehci-8349.so

post111226: Re: RE: Debug using USB Device instead of RJ45

2014-07-25T16:09:53Z

Thanks Andy, created in forum OSMeta.

post111224: RE: Debug using USB Device instead of RJ45

2014-07-25T15:53:48Z

> ioport=0x02184000, irq=75'

I see an extra space between "," and "irq": try removing the space.

If you could use the OSMeta discussion group I think it'll be a more applicable forum for this topic.

Cheers,

Andy

-----Original Message-----

From: Christopher Olarti [mailto:community-noreply@qnx.com]

Sent: Friday, July 25, 2014 11:51 AM

To: general-filesystems

Subject: Debug using USB Device instead of RJ45

I would like to use the USB Device as a USB to Ethernet port, so I can debug the board with usb instead of the RJ45 ethernet port.  So far I was able to get the USB Device Mass Storage working and tried to replicate the steps for the USB Device Ethernet (NCM), but I noticed there are differences; such as, the following:

1. In rules.lua when creating the Device_Stack = {

        cmd = 'io-usb-dcd -dusbncm-mx6sabrelite-ci ioport=0x02184000, irq=75';

path = '/dev/io-usb-dcd/io-usb';

        descriptors = {usbncm, usbser, umass};

   An error of

# Must specify irq with ioport optionIn function usbdc_attach_hub -- /builds/660_SDP/svn/hardware/support/publishers/usblauncher/dcd.c:247 device_root_hub != NULL -- assertion failed

2. In rules.lua there is no specific driver to invoke when creating device(), so instead tried using the following commands

start"mount -Tio-pkt -o mac=123456789abc, protocol=ncm devnp-usbdnet.so";

OR

driver"io-pkt-v4-hc -d usbdnet protocol=ncm, path=/dev/io-usb-dcd/io-usb, mac=123456789ab";

3. However, if I used the mass storage Device Stack of

cmd = 'io-usb-dcd -dusbumass-mx6sabrelite-ci ioport=0x02184000,irq=75'

and use the command for device()

driver"io-pkt-v4-hc -d usbdnet protocol=ncm, path=/dev/io-usb-dcd/io-usb, mac=123456789ab";

The program can run, but when looking in the Windows 7 Device Manager, I am seeing the device under "Other" instead of "Ports" or "Universal Serial Bus Controllers"

Please let me know if I am missing anything.

Thanks.

_______________________________________________

General

http://community.qnx.com/sf/go/post111223

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post111223: Debug using USB Device instead of RJ45

2014-07-25T15:50:46Z

I would like to use the USB Device as a USB to Ethernet port, so I can debug the board with usb instead of the RJ45 ethernet port.  So far I was able to get the USB Device Mass Storage working and tried to replicate the steps for the USB Device Ethernet (NCM), but I noticed there are differences; such as, the following:

1. In rules.lua when creating the Device_Stack = {

        cmd = 'io-usb-dcd -dusbncm-mx6sabrelite-ci ioport=0x02184000, irq=75';

path = '/dev/io-usb-dcd/io-usb';

        descriptors = {usbncm, usbser, umass};

   An error of

# Must specify irq with ioport optionIn function usbdc_attach_hub -- /builds/660_SDP/svn/hardware/support/publishers/usblauncher/dcd.c:247 device_root_hub != NULL -- assertion failed

2. In rules.lua there is no specific driver to invoke when creating device(), so instead tried using the following commands

start"mount -Tio-pkt -o mac=123456789abc, protocol=ncm devnp-usbdnet.so";

OR

driver"io-pkt-v4-hc -d usbdnet protocol=ncm, path=/dev/io-usb-dcd/io-usb, mac=123456789ab";

3. However, if I used the mass storage Device Stack of

cmd = 'io-usb-dcd -dusbumass-mx6sabrelite-ci ioport=0x02184000,irq=75'

and use the command for device()

driver"io-pkt-v4-hc -d usbdnet protocol=ncm, path=/dev/io-usb-dcd/io-usb, mac=123456789ab";

The program can run, but when looking in the Windows 7 Device Manager, I am seeing the device under "Other" instead of "Ports" or "Universal Serial Bus Controllers"

Please let me know if I am missing anything.

Thanks.

post110922: fs-qnx6 on a SSD

2014-07-03T08:37:50Z

Hi,

We have a customer planning to use fs-qnx6 on his next gen x86 project.

He is aware of the issues regarding 'sync' on compact flash.

He is asking if SSD would be more likely to implement 'sync' correctly, and if so can we recommend any manufacturers/devices that we know will work?

Thanks,

Garry

post110721: RE: RE: usblauncher creation

2014-06-17T19:14:41Z

Is your build file setup so that an inserted SD card is mountable? Maybe for starters, it'll be easier for you to copy over the 'usblauncher' binary and /etc/usblauncher files to the SD card and mount it after the system is booted, instead of modifying the IFS.

What is the location of the qnx660 install?

If for example the location (on Linux) is /opt/qnx660

then you'll find the etc files here:

/opt/qnx660/target/qnx6/etc/usblauncher

The executable binary here

/opt/qnx660/target/qnx6/armle-v7/bin/usblauncher

If your SD card is DOS, you could copy the binary to /dev/shmem, then set it executable (chmod +x usblauncher). For development purposes you can also mount a DOS filesystem as executable to run Neutrino binaries from it.

Help working with BSPs:

http://www.qnx.com/developers/docs/660/index.jsp?topic=%2Fcom.qnx.doc.neutrino.building%2Ftopic%2Fbsp.html

Making the IFS with 'mkifs' via the command-line:

http://www.qnx.com/developers/docs/660/index.jsp?topic=%2Fcom.qnx.doc.neutrino.utilities%2Ftopic%2Fm%2Fmkifs.html

Note: for general questions about usblauncher, feel free to post to the OS Meta forum:

http://community.qnx.com/sf/discussion/do/listTopics/projects.core_os/discussion.metawiki

For BSP-specific questions you can post to the BSP forum:

http://community.qnx.com/sf/discussion/do/listTopics/projects.bsp/discussion.bsp

Cheers,

Andy

-----Original Message-----

From: Christopher Olarti [mailto:community-noreply@qnx.com]

Sent: Thursday, June 12, 2014 5:33 PM

To: general-filesystems

Subject: Re: RE: usblauncher creation

Thanks for the response. However, when looking for the .build file that I should update, I came across 3 other .build files. The following is the original .build file that I am using:

/prebuilt/armle-v7/boot/build/sabrelite.build

However, the following are other .build files that I found:

/src/hardware/startup/boards/imx6x/sabrelite/build.build

/src/hardware/support/appsmedia-sample/appsmedia-sample-common.build

/src/hardware/support/appsmedia-sample/appsmedia-sample-imx6.build

Also when looking in the build scripts themselves, I am not exactly sure where to include the usblauncher location from the Host... If the modification is in sabrelite.build, then would it be under .console.sh when setting the environment variables, or would it be with .bootstrap when identifying the PATHs and LD_LIBRARY_PATH.

Thanks.

_______________________________________________

General

http://community.qnx.com/sf/go/post110684

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post110684: Re: RE: usblauncher creation

2014-06-12T21:32:58Z

Thanks for the response. However, when looking for the .build file that I should update, I came across 3 other .build files. The following is the original .build file that I am using:

/prebuilt/armle-v7/boot/build/sabrelite.build

However, the following are other .build files that I found:

/src/hardware/startup/boards/imx6x/sabrelite/build.build

/src/hardware/support/appsmedia-sample/appsmedia-sample-common.build

/src/hardware/support/appsmedia-sample/appsmedia-sample-imx6.build

Also when looking in the build scripts themselves, I am not exactly sure where to include the usblauncher location from the Host... If the modification is in sabrelite.build, then would it be under .console.sh when setting the environment variables, or would it be with .bootstrap when identifying the PATHs and LD_LIBRARY_PATH.

Thanks.

post110683: RE: usblauncher creation

2014-06-12T20:57:57Z

To get the /etc/usblauncher directory: the files are on your host (e.g. D:\qnx660\target\qnx6\etc\usblauncher) and you'd need to modify the build file for the BSP so that they are included in the IFS, or otherwise mount a filesystem that has these files so they are available on the target.

You'll need to start pps, usblauncher with a config file with the Host_Stack and Device_Stack sections enabled (the /etc/usblauncher/rules.lua has them commented out, remove --[[ and --]] markers to do this) then refer to

http://www.qnx.com/developers/docs/660/topic/com.qnx.doc.dev_pub.ref_guide/topic/usblauncher_stack_modes.html

and

http://www.qnx.com/developers/docs/660/index.jsp?topic=%2Fcom.qnx.doc.dev_pub.ref_guide%2Ftopic%2Fusblauncher_config_usb_descriptors.html

for updating the USB descriptors.

On the target

# pps

# mkdir /pps/qnx

# usblauncher -vvv -c /etc/usblauncher/rules.lua

The usblauncher command-line options are described here:

http://www.qnx.com/developers/docs/660/index.jsp?topic=%2Fcom.qnx.doc.dev_pub.ref_guide%2Ftopic%2Fusblauncher_usage.html

Cheers,

Andy

-----Original Message-----

From: Christopher Olarti [mailto:community-noreply@qnx.com]

Sent: Thursday, June 12, 2014 2:17 PM

To: general-filesystems

Subject: usblauncher creation

What would I do to initialize the usblauncher so i can update the necessary usb descriptors when making the i.MX6 SabreLite a Client/Device. I am trying to initialize the feature during the bootscript and I am unable to see the /etc/usblauncher/ directory.

Specifications:

QNX6.6 with BSP

i.MX6 SabreLite

Thanks.

_______________________________________________

General

http://community.qnx.com/sf/go/post110678

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post110678: usblauncher creation

2014-06-12T18:17:07Z

What would I do to initialize the usblauncher so i can update the necessary usb descriptors when making the i.MX6 SabreLite a Client/Device. I am trying to initialize the feature during the bootscript and I am unable to see the /etc/usblauncher/ directory.

Specifications:

QNX6.6 with BSP

i.MX6 SabreLite

Thanks.

post110316: diskboot in 6.6 deprecated how to upgrade?

2014-05-14T08:53:57Z

Hello,

We are using a PC104 as a development platform(target).

Normally we created a OS image with mkifs and copy it to a bootable compact flash drive, copy the contents of the file system of our Virtual Neutrino machine to the flash drive and then boot from the hard drive using the diskboot command. (script files at bottom of message).

However in version 6.6 of QNX diskboot is deprecated. I have now successfully created a bootable flash drive with an OS image using the QNX 6.6 generic x86 BSP. However due to this I do not have a file system anymore (with home folder and other parts) in which I can store my driver binaries and other components which I require for my target.

The question is, without the diskboot function which is deprecated, how can I get a properly functioning file system again on my harddrive? At this point after booting I cannot create directories or do anything with it.

Regards,

Jan Jaap

In version 6.5 we used to:

1. create a bootable compact flash file. and add the loaders with dloader

2. Create a OS image using mkifs and put it on the flashdrive.

====== mkifs script file =====

#

#

# GENERATED FILE - QNX System Builder Export

#   Image Type:     ifs

#   Image Sub-type:

#

[compress=3]

[virtual=x86,bios] .bootstrap = {

    startup-bios

    [module=aps] PATH=/proc/boot:/bin:/usr/bin:/usr/sbin:/sbin LD_LIBRARY_PATH=/proc/boot:/lib:/usr/lib:/lib/dll procnto-instr -v

}

# Start-up script

[+script] .script = {

# Programs require the runtime linker (ldqnx.so) to be at a fixed location

procmgr_symlink ../../proc/boot/libc.so.3 /usr/lib/ldqnx.so.2

# Create a 10% debugging APS partition

sched_aps Debugging 10

# Start up some consoles

[pri=10o] PATH=/proc/boot:/bin LD_LIBRARY_PATH=/proc/boot diskboot -b1 -D0 -odevc-con,-n4

# These env variables inherited by all the programs which follow

SYSNAME=nto

TERM=qansi

# redefine PATH (optional)

PATH=/proc/boot:/bin:/usr/bin:/usr/sbin:/sbin

LD_LIBRARY_PATH=/proc/boot:/lib:/usr/lib:/lib/dll

display_msg Welcome to QNX Neutrino 6.4.0 on an PC compatible BIOS system

display_msg .boot build 25/06/2010

}

[dperms=0777 uid=0 gid=0]

# Components

[type=link] /tmp = /dev/shmem

[perms=0777]

libc.so

fpemu.so.2

[data=copy]

devc-con

pci-bios

seedres

/bin/sh = ksh

slogger

sloginfo

libhiddi.so

libusbdi.so

libcam.so

io-blk.so

cam-disk.so

fs-qnx4.so

devu-ehci.so

devu-ohci.so

devu-uhci.so

devh-usb.so

devh-ps2ser.so

devb-eide

diskboot

fesh

devc-con-hid

io-usb

io-hid

==========

3. Copy the contents of the virtual machine to the flash drive:

==== Our file copy script on the QNX Neutrino VM ======

#!/bin/sh

echo Mounting..

mount /dev/hd2t79 /mnt/cf

echo Copying...

cp -rc /usr/lib /mnt/cf/usr/lib

cp -rc /usr/bin /mnt/cf/usr/bin

cp -rc /usr/sbin /mnt/cf/usr/sbin

cp -rc /lib /mnt/cf/lib

cp -rc /bin /mnt/cf/bin

cp -rc /etc /mnt/cf/etc

cp -rc /sbin /mnt/cf/sbin

cp -rc /usr/include /mnt/cf/usr/include

cp -rc /usr/photon /mnt/cf/usr/photon

cp -rc /usr/share /mnt/cf/usr/share

cp -rc /usr/libexec /mnt/cf/usr/libexec

rm /mnt/cf/etc/system/setupisdone

#Special settings for terminal QNX target.

cp ./.diskroot /mnt/cf/.diskroot

touch /mnt/cf/etc/system/config/nophoton

cp -c ./rc.local /mnt/cf/etc/rc.d/rc.local

chmod +x /mnt/cf/etc/rc.d/rc.local

cp ./inetd.conf /mnt/cf/etc/inetd.conf

mkdir /mnt/cf/root

touch /mnt/cf/root/.profile

echo Unmounting...

umount /mnt/cf

echo Done...

=========

And insert the flash drive into the PC104 and boot, where we had a properly functioning file system.

post110017: Re: USB mass storage device and FS mount file matching

2014-04-24T04:35:10Z

Dear Andrew,

thank you for your answer,

I use 6.5 SDP so usblauncher isn´t possible to use.

Do you have any other idea how to match devb-umass devices and mounts filesystems in 6.5 SDP?

Thanx

meg

post110000: Re: USB mass storage device and FS mount file matching

2014-04-23T16:46:52Z

Which version of QNX SDP are you using?

SDP 6.6 includes usblauncher that may help you; the documentation is here:

http://www.qnx.com/developers/docs/660/index.jsp?topic=%2Fcom.qnx.doc.dev_pub.ref_guide%2Ftopic%2Fusblauncher.html

It launches drivers such as devb-umass, mounts filesystems, and links them together with PPS objects.

post109965: USB mass storage device and FS mount file matching

2014-04-18T12:25:39Z

Hello,

I´m looking for a way how to find coherence between USB mass storage device and mounted file in the file system.

I´m programming an application which will show the tabe of conected umass devices to the user with all posibble description and with its mounted files in filesystem.

Does exists some functions (or system commands) which I can use for direct maching?

I have find only one possible solution up to now - to majke a list running devb-umass processes, here I obtain a device vendor and by the order of processes in the list to enumerate the order number in mount file name. This probably will work by I prefere to find some more elegant way.

Thank you for help in advance

meg

post109694: Re: Issue with USB stick under 6.6

2014-03-31T14:33:54Z

On Mon, Mar 31, 2014 at 06:04:36AM -0400, Mario Charest wrote:

> > On Thu, Mar 27, 2014 at 02:57:39PM -0400, Mitchell Schoenbrun wrote:

>

> > There were a couple of packages that parsed the ls output on install

> > and got confused.  I don't remember which ones ATM.

>

> At the risk of sounding Igorish, so at the risk for breaking existing customer script in favor of compatibility with packages most of us are not able to build anymore, the behavior was changed ?

I'm trying to hack up a way to build most things with what's

released in 6.6.

>

> The change in behavior is not document in ls and I don`t recall reading about it anywhere.

This is being addressed in a release note.  On the other hand, the

root exception was never doc'd AFAICT so it now matches its previously

doc'd behaviour.

post109693: Re: Issue with USB stick under 6.6

2014-03-31T10:04:36Z

> On Thu, Mar 27, 2014 at 02:57:39PM -0400, Mitchell Schoenbrun wrote:

> There were a couple of packages that parsed the ls output on install

> and got confused. I don't remember which ones ATM.

At the risk of sounding Igorish, so at the risk for breaking existing customer script in favor of compatibility with packages most of us are not able to build anymore, the behavior was changed ?

The change in behavior is not document in ls and I don`t recall reading about it anywhere.

post109651: Re: Issue with USB stick under 6.6

2014-03-27T19:43:28Z

And yes, I'll mention this change in the release notes.

> On Thu, Mar 27, 2014 at 02:57:39PM -0400, Mitchell Schoenbrun wrote:

> > >

> > > No it was changed.  We were the only system that did

> > > the dots for root and it broke stuff.

> > >

> > Sean,

> >

> >     Now that I understand the change, it's not a problem.   But I'm very

> curious as what might have broke.   Anything obvious, or all subtle stuff?

>

> There were a couple of packages that parsed the ls output on install

> and got confused.  I don't remember which ones ATM.

post109650: Re: Issue with USB stick under 6.6

2014-03-27T19:03:21Z

On Thu, Mar 27, 2014 at 02:57:39PM -0400, Mitchell Schoenbrun wrote:

> >

> > No it was changed.  We were the only system that did

> > the dots for root and it broke stuff.

> >

> Sean,

>

>     Now that I understand the change, it's not a problem.   But I'm very curious as what might have broke.   Anything obvious, or all subtle stuff?

There were a couple of packages that parsed the ls output on install

and got confused.  I don't remember which ones ATM.

post109649: Re: Issue with USB stick under 6.6

2014-03-27T18:57:39Z

>

> No it was changed. We were the only system that did

> the dots for root and it broke stuff.

>

Sean,

Now that I understand the change, it's not a problem. But I'm very curious as what might have broke. Anything obvious, or all subtle stuff?

Mitchell

post109648: Re: Issue with USB stick under 6.6

2014-03-27T17:06:51Z

No it was changed. We were the only system that did

the dots for root and it broke stuff.

On Thu, Mar 27, 2014 at 01:04:03PM -0400, Colin Burgess wrote:

> It's likely that you had an alias in your 6.5 system to add the -a to ls

>

> On 2014-03-27 11:05 AM, Mitchell Schoenbrun wrote:

> > The difference appears to be "ls".

> > Under QNX 6.5 (and every QNX back to 4) when run as root, ls shows the DOT files. Under 6.6 it does not. Maybe this is how it is supposed to work?

> >

> >

> >

> > _______________________________________________

> >

> > General

> > http://community.qnx.com/sf/go/post109640

> > To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

>

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post109647

> To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post109647: Re: Issue with USB stick under 6.6

2014-03-27T17:04:10Z

It's likely that you had an alias in your 6.5 system to add the -a to ls

On 2014-03-27 11:05 AM, Mitchell Schoenbrun wrote:

> The difference appears to be "ls".

> Under QNX 6.5 (and every QNX back to 4) when run as root, ls shows the DOT files. Under 6.6 it does not. Maybe this is how it is supposed to work?

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post109640

> To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post109641: Re: RE: RE: Issue with USB stick under 6.6

2014-03-27T15:09:27Z

Yes. ls -a will show the '.' and '..' .

On Thu, Mar 27, 2014 at 11:05:30AM -0400, Mitchell Schoenbrun wrote:

> The difference appears to be "ls".

> Under QNX 6.5 (and every QNX back to 4) when run as root, ls shows the DOT files. Under 6.6 it does not. Maybe this is how it is supposed to work?

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post109640

> To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post109640: Re: RE: RE: Issue with USB stick under 6.6

2014-03-27T15:05:30Z

The difference appears to be "ls".

Under QNX 6.5 (and every QNX back to 4) when run as root, ls shows the DOT files. Under 6.6 it does not. Maybe this is how it is supposed to work?

post109639: RE: RE: Issue with USB stick under 6.6

2014-03-27T14:44:40Z

mkqnx6fs should create the /.boot directory automatically. Yes, the /.boot directory is special, and doing an "mkdir /.boot" isn't going to work. Do you still have the 6.5.0 version of mkqnx6fs to try, and compare (use the same OS, just run the older binary).

-----Original Message-----

From: Mitchell Schoenbrun [mailto:community-noreply@qnx.com]

Sent: March-27-14 10:41 AM

To: general-filesystems

Subject: Re: RE: Issue with USB stick under 6.6

Thanks for getting back to me so fast on this. I notice that unlike 6.5, mkqnx6fsys does not create the .boot directory. Will just doing a "mkdir .boot" solve that? It seemed like there was something special about the .boot directory that used to be created.

_______________________________________________

General

http://community.qnx.com/sf/go/post109638

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post109638: Re: RE: Issue with USB stick under 6.6

2014-03-27T14:41:09Z

Thanks for getting back to me so fast on this. I notice that unlike 6.5, mkqnx6fsys does not create the .boot directory. Will just doing a "mkdir .boot" solve that? It seemed like there was something special about the .boot directory that used to be created.

post109637: RE: Issue with USB stick under 6.6

2014-03-27T14:34:43Z

You need mount -t qnx6 -o sync=optional /dev/umass0t179 /fs/usb0

The USB device is not reporting that it supports the hardware sync command.

-----Original Message-----

From: Mitchell Schoenbrun [mailto:community-noreply@qnx.com]

Sent: March-27-14 10:33 AM

To: general-filesystems

Subject: Issue with USB stick under 6.6

In what may turn out to be a never ending attempt to create a bootable QNX 6.6 system using only QNX 6.6 I ran into this roadblock.

From a QNX 6.5 VM I do the following:

# devb-umass cam pnp disk name=umass

# ls /dev/umass0

/dev/umass0 /dev/umass0t179

# mkqnx6fsys -q /dev/umass0179

# mount -tqnx6 /dev/umass0t179 /fs/usb0

When I do exactly the same thing from the QNX 6.6 VM I get the following:

#mount -tqnx6 /dev/umass0t179

mount: Can't mount /fs/usb0 (type qnx6)

mount: Possible reason: Read-only file system

Anyone run into this before?

_______________________________________________

General

http://community.qnx.com/sf/go/post109636

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post109636: Issue with USB stick under 6.6

2014-03-27T14:32:58Z

In what may turn out to be a never ending attempt to create a bootable QNX 6.6 system using only QNX 6.6 I ran into this roadblock.

From a QNX 6.5 VM I do the following:

# devb-umass cam pnp disk name=umass

# ls /dev/umass0

/dev/umass0 /dev/umass0t179

# mkqnx6fsys -q /dev/umass0179

# mount -tqnx6 /dev/umass0t179 /fs/usb0

When I do exactly the same thing from the QNX 6.6 VM I get the following:

#mount -tqnx6 /dev/umass0t179

mount: Can't mount /fs/usb0 (type qnx6)

mount: Possible reason: Read-only file system

Anyone run into this before?

post109528: Writing to raw NOR flash

2014-03-20T10:32:01Z

Hi,

I have been given a task to see if its possible to write to raw NOR flash from c code.

I understand that NOR flash needs to be erased at block level before writing to it, however in this case the raw flash will already be erased and data will only ever be written to a location once. So I believe I don't need to erase beforehand.

I tried using system("dd...etc") which works, however, it takes too long for my needs. The time taken seems to depend on the block size and the address position in NOR. This may be to do with the implementation of 'dd'.

When writing to a file in NOR flash that has a fs mounted the write is quick.

Is there an API that allows me to write directly to raw NOR flash or manage the writing to raw NOR flash? Maybe similar to that which the fs uses?

Thanks

Simon

post109276: Re: RE: 6.6 Release note

2014-03-07T18:06:55Z

I figured as much - it's somewhat disappointing to us, as we used it to support a wide variety of boards and configurations here with our data acquisition software stack, but we'll figure out some way to roll it ourselves. I just wanted to point out that, like the TRIM support, the fact enum-devices was deprecated wasn't mentioned in the release notes. It mentions that enum-usb has been replaced by usblauncher, but there isn't any mention of the rest of the enum-* stack.

post109267: Re: RE: 6.6 Release note

2014-03-07T15:11:03Z

Enum-devices was deprecated because we no longer support self-hosted development systems. I guess the rationale was that if you're building a Neutrino target, you're going to set up the image to use the hardware you've chosen, instead of relying on device enumeration.

post109264: Re: RE: 6.6 Release note

2014-03-07T14:49:10Z

I now know what's wrong with the links, and I'm trying to figure out how to fix them.

post109254: Re: RE: 6.6 Release note

2014-03-07T01:06:48Z

Hi Steve,

I noticed that the documentation for enum-devices has disappeared from the utilities reference, but I don't recall seeing any mention of it being deprecated or otherwise removed in the release notes. The binaries are still present in the VM runtime ISO, however. Is this something else that got accidentally overlooked?

-Will

post109244: Re: RE: 6.6 Release note

2014-03-06T19:48:36Z

Async messaging is gone; async message queues (i.e., those managed by mq) are experimental.

I must have missed that reference to SH. I'll remove it.

post109243: RE: 6.6 Release note

2014-03-06T19:44:48Z

Not filesystem issue but:

asynchronous messaging (i.e., the asyncmsg_*() functions) are deprecated but futher down the document they are specified as experimental

========

Ld

Between the 2.10.1 version of the GNU linker in QNX Momentics 6.2.1 and the 2.12.1 version in QNX Momentics 6.3.0, a bug was fixed in the handling of relocation addends for SH targets. As a result of this fix, SH startup binaries (e.g. startup-systemh) that were created prior to QNX Momentics 6.3.0 won't work correctly if included in a boot image generated in QNX Momentics 6.3 or later.

Workaround: Rebuild the startup binary using the QNX Momentics Tool Suite 6.6. The resulting startup will work with 6.2.1 or later.

==== >>> Can`t build SH targets on 6.6 !

Most link in the hyperlink in the release note point to : http://www.qnx.com/developers/articles/index.html#new_os but that pages is all the old technical articles.

Cheers,

-----Message d'origine-----

De : Steve Reid [mailto:community-noreply@qnx.com]

Envoyé : Thursday, March 06, 2014 1:39 PM

À : general-filesystems

Objet : Re: 6.6 Release note

Good suggestion--I'll see what I can cook up. Thanks, Mario.

_______________________________________________

General

http://community.qnx.com/sf/go/post109238

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post109241: Re: 6.6 Release note

2014-03-06T19:02:04Z

Agreed! This is important.

post109238: Re: 6.6 Release note

2014-03-06T18:39:20Z

Good suggestion--I'll see what I can cook up. Thanks, Mario.

post109237: 6.6 Release note

2014-03-06T18:33:55Z

The Release notes aren`t taking about thew new TRIM support ( according to the output of sloginfo). That a very nice new feature worth talking about AFAIC

post107671: Unzip a Password Protected Zip File?

2013-12-19T21:33:47Z

Does anyone know if we have any libraries that we can use to unzip a password protected zip file?

I know we can use the uzip command when we telnet into the device, but I need to be able to unzip a password protect zip file with the FileManager App. Currently the FileManager app uses the libArchive library to extract zip files. However there are no APIs that support unzipping encrypted or password protected files.

post107556: Bootable IFS mounting IFS

2013-12-13T19:52:16Z

Most documentation I've read discusses how to do this with NOR FLASH, I'm wondering is it possible to do with NAND FLASH? Is underlying plumbing going to do the job or not? In other words would the following actually mount an IFS programmed at 0x600000 to /ifs2 (noting correct offset and size values coincide with the part)?

mount -tifs -o offset=0x600000,size=0x24000 /dev/mem /ifs2

As seen here:

http://www.qnx.com/developers/docs/qnxcar2/index.jsp?topic=%2Fcom.qnx.doc.neutrino.technotes%2Ftopic%2Fmultiple_Mounting.html

Thanks.

post106482: Re: RE: Is it possible to boot an EFS image?

2013-11-01T14:59:37Z

Cool, thank you.

post106481: RE: Is it possible to boot an EFS image?

2013-11-01T14:56:54Z

You're right that you'll have to use an IFS image in order to boot. The IFS can be pretty small (just procnto, libc, and devf), but it needs to be present.

-----Original Message-----

From: Jeremy Quandt [mailto:community-noreply@qnx.com]

Sent: November-01-13 10:55 AM

To: general-filesystems

Subject: Is it possible to boot an EFS image?

Is it possible to boot an EFS image? If my understanding is correct it is expected to boot an IFS and mount an EFS, but I wanted to confirm this thinking. Thanks.

_______________________________________________

General

http://community.qnx.com/sf/go/post106480

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post106480: Is it possible to boot an EFS image?

2013-11-01T14:54:58Z

Is it possible to boot an EFS image? If my understanding is correct it is expected to boot an IFS and mount an EFS, but I wanted to confirm this thinking. Thanks.

post105629: Re: Should I reduce my Task's priority before it writes to the hard drive?

2013-10-04T06:50:09Z

Now the question is what you exactly expect.

In general, setting any thread that does time consuming operations to a high priority is asking for trouble. As for writing to disk, your thread will block any lower priority threads when your thread has anything to do, which you should expect. In reality, this will be when copying data. What's really time consuming is transferring the data to and committing to disk but this does not use the CPU. So both your thread and the driver thread are blocked most of the time and lower priority threads in the system are able to run.

As always, data is king so please benchmark.

--

Mate

From: Tier Befreier

Sent: Friday, October 4, 2013 15:29 PM

To: general-filesystems

Reply To: general-filesystems@community.qnx.com

Subject: Should I reduce my Task's priority before it writes to the hard drive?

Hi,

Should I reduce my Task's priority before it writes to the hard drive?

Or is it handled like I would expect by the hard drive driver?

For example: a high priority task wants to write a huge file to the hard drive. Does it block other tasks while the operation is ongoing?

_______________________________________________

General

http://community.qnx.com/sf/go/post105628

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post105628: Should I reduce my Task's priority before it writes to the hard drive?

2013-10-04T06:29:09Z

Hi,

Should I reduce my Task's priority before it writes to the hard drive?

Or is it handled like I would expect by the hard drive driver?

For example: a high priority task wants to write a huge file to the hard drive. Does it block other tasks while the operation is ongoing?

post105510: Creating multiple partitions on NAND flash using ETFS driver

2013-09-30T22:35:48Z

How to create multiple partitions using fs-etfs-mx51? What are the options available?

post105393: RE: XML file Parser

2013-09-25T12:26:59Z

If you are doing C++ there is the rapid_xml library which I found to be very fast when parsing and does no allocation ( I think), it works within the buffer that contains the xml source. It`s all based on template

> -----Message d'origine-----

> De : Gervot Olivier [mailto:community-noreply@qnx.com]

> Envoyé : 25 septembre 2013 08:16

> À : general-filesystems

> Objet : Re: XML file Parser

>

> The XML to parse is not a file but a string provided by a socket. The frequency

> is not yet kwown.

>

> Due to the SIL 2 requirements, dynamics allocations are not allowed.

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post105390

> To cancel your subscription to this discussion, please e-mail general-

> filesystems-unsubscribe@community.qnx.com

post105390: Re: XML file Parser

2013-09-25T12:15:44Z

The XML to parse is not a file but a string provided by a socket. The frequency is not yet kwown.

Due to the SIL 2 requirements, dynamics allocations are not allowed.

post105348: Re: XML file Parser

2013-09-24T13:25:44Z

Hello, Chris Hobbs here. I work on our safety-certified products.

The bad news is that our XML parsing library doesn't form part of our certified product (although it can, of course, be used with the certified kernel and libc).

Would you be willing to provide some details of what you're trying to achieve? I assume that you are looking for SIL2 as defined for "on demand" mode (given that an XML library doesn't run continuously). Is this low-demand or high-demand?

post105345: XML file Parser

2013-09-24T12:37:50Z

I know that XML parsing is not really a "File system" topic but its the more approching

I would like to know it there is a library for QNX classified SIL2 (Safety Integrity Level 2) for parsing XML files?

Thanks

Olivier

post102954: ETFS and static wear-leveling corrupted file

2013-07-08T07:32:41Z

Hi,

we did a some mistake in our application. A module was continuously writing log data

onto our NAND flash (128MB) using etfs as filesystem.

After some time the system was forced to do a restart. The application was not starting.

I did some analysis and saw that ldd came up with an error that it can't load libssl.so.1?

What went wrong? Why it is not possible to load this file, no one touch it for any reason. Then I had a read through again on the wear-leveling ...

======================================================================

Static wear-leveling

Filesystems often consist of a large number of static files that are read but not written. These files will occupy flash blocks that have no reason to be erased. If the majority of the files in flash are static, this will cause the remaining blocks containing dynamic data to wear at a dramatically increased rate.

ETFS notices these under-worked static blocks and forces them into service by copying their data to an over-worked block. This solves two problems: It gives the over-worked block a rest, since it now contains static data, and it forces the under-worked static block into the dynamic pool of blocks.

======================================================================

And it looks like that exactly this happened. The file was relocated, but during this process something went wrong. I compared the binary data see the following hex dump.

[ORIGINAL]

=========================================================

Offset(h) 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F

00000000  7F 45 4C 46 01 02 01 00 00 00 00 00 00 00 00 00  .ELF............

00000010  00 03 00 14 00 00 00 01 00 00 E6 2C 00 00 00 34  ..........æ,...4

00000020  00 04 4A E4 00 00 80 00 00 34 00 20 00 05 00 28  ..Jä..€..4. ...(

00000030  00 1E 00 1B 00 00 00 01 00 00 00 00 00 00 00 00  ................

00000040  00 00 00 00 00 04 0E 68 00 04 0E 68 00 00 00 05  .......h...h....

00000050  00 00 10 00 00 00 00 01 00 04 0E 88 00 04 1E 88  ...........ˆ...ˆ

00000060  00 04 1E 88 00 00 37 24 00 00 50 D0 00 00 00 07  ...ˆ..7$..PÐ....

00000070  00 00 10 00 00 00 00 02 00 04 16 04 00 04 26 04  ..............&.

00000080  00 04 26 04 00 00 00 B8 00 00 00 B8 00 00 00 06  ..&....¸...¸....

00000090  00 00 00 04 00 00 00 04 00 04 45 AC 00 00 00 D4  ..........E¬...Ô

000000A0  00 00 00 D4 00 00 00 20 00 00 00 00 00 00 00 04  ...Ô... ........

000000B0  00 00 00 04 00 00 00 04 00 04 45 D0 00 04 0E 70  ..........EÐ...p

000000C0  00 04 0E 70 00 00 00 18 00 00 00 00 00 00 00 04  ...p............

000000D0  00 00 00 10 00 00 00 00 00 00 00 00 00 00 00 00  ................

000000E0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

000000F0  00 00 00 00 00 00 02 09 00 00 03 09 00 00 02 C4  ...............Ä

00000100  00 00 01 0F 00 00 02 64 00 00 02 9D 00 00 02 D5  .......d.......Õ

00000110  00 00 01 C3 00 00 02 2F 00 00 00 00 00 00 02 C5  ...Ã.../.......Å

00000120  00 00 02 E4 00 00 02 66 00 00 02 8F 00 00 02 9F  ...ä...f.......Ÿ

00000130  00 00 02 25 00 00 02 06 00 00 02 C0 00 00 02 B7  ...%.......À...·

00000140  00 00 01 40 00 00 02 F5 00 00 01 61 00 00 00 00  ...@...õ...a....

=========================================================

[CORRUPTED]

=========================================================

Offset(h) 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F

00000000  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000010  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000020  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000030  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000040  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000050  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000060  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000070  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000080  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000090  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

000000A0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

000000B0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

000000C0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

000000D0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

000000E0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

000000F0  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000100  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000110  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000120  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000130  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

00000140  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................

=========================================================

At the offset of 800h the data is OK again.

What went wrong? Is there anything to activate that an "error" durning relocation is detected?

Any help, suggestions is welcome and if I'm wrong with my theory also please tell me.

Kind Regards

Thomas

post102366: Shutdown sometimes paused at devb-eide

2013-06-18T08:05:58Z

I have modified /boot/qnxbasesmp.build and make a new IFS using mkifs qnxbasesmp.build /.boot/qnxbasesmp.ifs.

However the qnx box cannot reboot by shutdown, it will pause at shutdown daemons.

I noticed that this is caused by the failure of slaying devb-eide.

To speed up startup, I add a time before diskboot in build:

seedres

slogger -s128k

pci-bios -B

waitfor /dev/pci

devb-eide cam quiet

waitfor /dev/hd0

This will be happen occasionally, especially after remake IFS.

Please help me get it solved.

Thank you in advance.

Allen

post102104: Use devf-ram and efs file build with min_size

2013-06-10T07:53:54Z

Hi,

I can't figure out how to test my build with deaf-ram. Here is the build file :

[block_size=128k spare_blocks=1 min_size=18432k]

dummy.dat = {

This is a dummy data file.

}

mkefs build my_efs

slay devf-ram

cp my_efs /dev/shmem/fs0

devf-ram -s0,18432k,,,128k

I can't find the auto mounted folder.

If I don't set the min_size value, it works like a charm. Maybe I'm missing something...

Thanks for the help

post101853: Re: devf-ram does not automount as described

2013-05-29T21:39:08Z

I think you meant -s0,1m,,,128k

The third argument for this option is aoffset (array offset info flash), the fourth is asize (size of flash array equal to wsize), and the fifth is usize (physical erase sector size).  Works great!  Thanks for the help!

# cat /dos/bp_fs.img > /dev/shmem/fs1

# devf-ram -i1 -s0,1m,,,128k &

[1] 40983

#

[1] + Done                 devf-ram -i1 -s0,1m,,,128k

# ls

bin      dos      fs1p0    sbin     usr

dev      etc      proc     tmp

# cd /fs1p0

# ls

dummy.dat

# cat dummy.dat

This is a dummy data file.

#

post101851: Re: devf-ram does not automount as described

2013-05-29T21:26:55Z

Specify the sector size with -s0,1m,128k

Sent from my BlackBerry 10 smartphone.

From: Erick Roane

Sent: Wednesday, May 29, 2013 5:24 PM

To: general-filesystems

Reply To: general-filesystems@community.qnx.com

Subject: Re: devf-ram does not automount as described

Still not showing up.

# cat /dos/bp_fs.img > /dev/shmem/fs1

# devf-ram -i 1 &

[1] 57367

#

[1] + Done devf-ram -i 1

# ls

bin dev dos etc proc sbin tmp usr

#

Maybe I'm doing something wrong with my efs filesytem build. Here is my Makefile

all : my_efs_img

echo "Making QNX6 file image"

my_efs_img : build

mkefs -v build my_efs_img

clean :

echo "Cleaning QNX 6 file system image"

rm -f my_efs_img

And here is my build file:

[search=C:\QNX650\target\qnx6\armle-v7\bin;C:\QNX650\target\qnx6\armle-v7\sbin;C:\QNX650\target\qnx6\armle-v7\usr\bin;C:\QNX650\target\qnx6\armle-v7\usr\sbin;C:\QNX650\target\qnx6\armle-v7\usr\photon\bin;C:\QNX650\target\qnx6\armle-v7\usr\libexec;C:\QNX650\target\qnx6\armle-v7\lib\dll;C:\QNX650\target\qnx6\armle-v7\usr\photon\dll;C:\QNX650\target\qnx6\armle-v7\lib;C:\QNX650\target\qnx6\armle-v7\usr\lib;C:\QNX650\target\qnx6\armle-v7\usr\photon\lib;C:\QNX650\target\qnx6\armle-v7\etc;C:\QNX650\target\qnx6\etc;C:\QNX650\target\qnx6\usr\photon;C:\QNX650\target\qnx6\usr\photon\keyboard;C:\QNX650\target\qnx6\usr\lib;C:\QNX650\target\qnx6\usr\share;C:\QNX650\target\qnx6\armle-v7\boot\sys;C:\QNX650\target\qnx6\usr\lib\terminfo;C:\QNX650\target\qnx6\usr\lib\terminfo\a;C:\QNX650\target\qnx6\usr\lib\terminfo\q;C:\QNX650\target\qnx6\usr\lib\terminfo\t;C:\QNX650\target\qnx6\usr\lib\terminfo\v;C:\QNX650\target\qnx6\usr\lib\terminfo\x;C:\QNX650\target\qnx6\etc\system\config;C:\QNX650\target\qnx6\etc\ssh]

[block_size=128k spare_blocks=1 min_size=384k]

dummy.dat = {

This is a dummy data file.

}

_______________________________________________

General

http://community.qnx.com/sf/go/post101850

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post101850: Re: devf-ram does not automount as described

2013-05-29T21:24:23Z

Still not showing up.

# cat /dos/bp_fs.img > /dev/shmem/fs1

# devf-ram -i 1 &

[1] 57367

#

[1] + Done                 devf-ram -i 1

# ls

bin     dev     dos     etc     proc    sbin    tmp     usr

#

Maybe I'm doing something wrong with my efs filesytem build.  Here is my Makefile

all : my_efs_img

echo "Making QNX6 file image"

my_efs_img : build

mkefs -v build my_efs_img

clean :

echo "Cleaning QNX 6 file system image"

rm -f my_efs_img

And here is my build file:

[search=C:\QNX650\target\qnx6\armle-v7\bin;C:\QNX650\target\qnx6\armle-v7\sbin;C:\QNX650\target\qnx6\armle-v7\usr\bin;C:\QNX650\target\qnx6\armle-v7\usr\sbin;C:\QNX650\target\qnx6\armle-v7\usr\photon\bin;C:\QNX650\target\qnx6\armle-v7\usr\libexec;C:\QNX650\target\qnx6\armle-v7\lib\dll;C:\QNX650\target\qnx6\armle-v7\usr\photon\dll;C:\QNX650\target\qnx6\armle-v7\lib;C:\QNX650\target\qnx6\armle-v7\usr\lib;C:\QNX650\target\qnx6\armle-v7\usr\photon\lib;C:\QNX650\target\qnx6\armle-v7\etc;C:\QNX650\target\qnx6\etc;C:\QNX650\target\qnx6\usr\photon;C:\QNX650\target\qnx6\usr\photon\keyboard;C:\QNX650\target\qnx6\usr\lib;C:\QNX650\target\qnx6\usr\share;C:\QNX650\target\qnx6\armle-v7\boot\sys;C:\QNX650\target\qnx6\usr\lib\terminfo;C:\QNX650\target\qnx6\usr\lib\terminfo\a;C:\QNX650\target\qnx6\usr\lib\terminfo\q;C:\QNX650\target\qnx6\usr\lib\terminfo\t;C:\QNX650\target\qnx6\usr\lib\terminfo\v;C:\QNX650\target\qnx6\usr\lib\terminfo\x;C:\QNX650\target\qnx6\etc\system\config;C:\QNX650\target\qnx6\etc\ssh]

[block_size=128k spare_blocks=1 min_size=384k]

dummy.dat = {

This is a dummy data file.

}

post101849: Re: devf-ram does not automount as described

2013-05-29T21:06:33Z

Pass -i 1 to the driver when you start it.

Sent from my BlackBerry 10 smartphone.

From: Erick Roane

Sent: Wednesday, May 29, 2013 5:00 PM

To: general-filesystems

Reply To: general-filesystems@community.qnx.com

Subject: Re: devf-ram does not automount as described

Again, I cannot use fs0 because that is my boot ROM. But I tried this with fs1 and this is what I got:

# cat /dos/my_efs_img > /dev/shmem/fs1

# devf-ram &

[1] 61463

#

[1] + Done devf-ram

# ls

bin dev dos etc proc sbin tmp usr

# ls /dev

bpf hd0t11 ptyp0 shmem tap2 ttyp6

bpf0 hd0t179 ptyp1 slog tap3 ttyp7

can0 i2c1 ptyp2 snd text tun0

console i2c2 ptyp3 socket tty tun1

dbgmem i2c3 ptyp4 stderr ttyp0 tun2

fs0 io-usb ptyp5 stdin ttyp1 tun3

fs0p0 mem ptyp6 stdout ttyp2 tymem

fs1 null ptyp7 tap ttyp3 zero

fs1p0 pipe sem tap0 ttyp4

hd0 profiler ser2 tap1 ttyp5

#

Notice the /dev/fs1p0 appears in the system but it is not mounted. If I try to explicitly mount it:

# flashctl -p /dev/fs1p0 -m

flashctl: mounting partition failed

Because the simulated flash file system in RAM must always be erased before it is mounted when using flashctl.

_______________________________________________

General

http://community.qnx.com/sf/go/post101848

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post101848: Re: devf-ram does not automount as described

2013-05-29T21:00:48Z

Again, I cannot use fs0 because that is my boot ROM.  But I tried this with fs1 and this is what I got:

# cat /dos/my_efs_img > /dev/shmem/fs1

# devf-ram &

[1] 61463

#

[1] + Done                 devf-ram

# ls

bin     dev     dos     etc     proc    sbin    tmp     usr

# ls /dev

bpf          hd0t11       ptyp0        shmem        tap2         ttyp6

bpf0         hd0t179      ptyp1        slog         tap3         ttyp7

can0         i2c1         ptyp2        snd          text         tun0

console      i2c2         ptyp3        socket       tty          tun1

dbgmem       i2c3         ptyp4        stderr       ttyp0        tun2

fs0          io-usb       ptyp5        stdin        ttyp1        tun3

fs0p0        mem          ptyp6        stdout       ttyp2        tymem

fs1          null         ptyp7        tap          ttyp3        zero

fs1p0        pipe         sem          tap0         ttyp4

hd0          profiler     ser2         tap1         ttyp5

#

Notice the /dev/fs1p0 appears in the system but it is not mounted.  If I try to explicitly mount it:

# flashctl -p /dev/fs1p0 -m

flashctl: mounting partition failed

Because the simulated flash file system in RAM must always be erased before it is mounted when using flashctl.

post101847: Re: devf-ram does not automount as described

2013-05-29T20:49:01Z

Copy the efs to /dev/shmem/fs0, and then start devf-ram.

Sent from my BlackBerry 10 smartphone.

From: Erick Roane

Sent: Wednesday, May 29, 2013 4:42 PM

To: general-filesystems

Reply To: general-filesystems@community.qnx.com

Subject: Re: RE: RE: devf-ram does not automount as described

Agreed, but to my other question, what other way can I transfer a pre-built image to devf-ram? I actually want to use it until my real flash system is available.

_______________________________________________

General

http://community.qnx.com/sf/go/post101845

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post101845: Re: RE: RE: devf-ram does not automount as described

2013-05-29T20:42:39Z

Agreed, but to my other question, what other way can I transfer a pre-built image to devf-ram? I actually want to use it until my real flash system is available.

post101844: RE: RE: devf-ram does not automount as described

2013-05-29T20:32:35Z

Here's a link to the HTML version of that section:

http://www.qnx.com/developers/docs/6.5.0_sp1/topic/com.qnx.doc.neutrino_building/building_nto.html?cp=13_2_3_5_2#TransferringToFlash

I think this section uses devf-ram as a rather artificial example to illustrate how to transfer an image to flash using flash filesystem raw mountpoints. Perhaps we should make the example more realistic or point out that that step won't actually work with devf-ram.

Steve Reid (stever@qnx.com)

Technical Editor

QNX Software Systems

-----Original Message-----

From: Erick Roane [mailto:community-noreply@qnx.com]

Sent: Wednesday, May 29, 2013 3:36 PM

To: general-filesystems

Subject: Re: RE: devf-ram does not automount as described

Then why does http://support7.qnx.com/download/download/20951/building.pdf give the following guidance:

"To use the [devf-ram] ﬂash ﬁlesystem partition, you need to slay the driver and start it again so it can recognize the partitions and mount them. For instance, with devf-ram:

$ slay devf-ram

$ devf-ram &

From this point, you have a /fs0p1 mountpoint that’s in fact a directory and contains the ﬁles you speciﬁed with mkefs to create your ﬂash image."

You can substitude the "fs0p1" in the paragraph above with "fs1p0" for my Nitrogen6X system since /dev/fs0 is the on-board boot ROM. What other way can I mount the image using devf-ram?

_______________________________________________

General

http://community.qnx.com/sf/go/post101841

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post101841: Re: RE: devf-ram does not automount as described

2013-05-29T19:36:12Z

Then why does http://support7.qnx.com/download/download/20951/building.pdf give the following guidance:

"To use the [devf-ram] ﬂash ﬁlesystem partition, you need to slay the driver and start it again so it can recognize the partitions and mount them. For instance, with devf-ram:

$ slay devf-ram

$ devf-ram &

From this point, you have a /fs0p1 mountpoint that’s in fact a directory and contains the ﬁles you speciﬁed with mkefs to create your ﬂash image."

You can substitude the "fs0p1" in the paragraph above with "fs1p0" for my Nitrogen6X system since /dev/fs0 is the on-board boot ROM. What other way can I mount the image using devf-ram?

post101840: RE: devf-ram does not automount as described

2013-05-29T19:31:35Z

> -----Message d'origine-----

> De : Erick Roane [mailto:community-noreply@qnx.com]

> Envoyé : 29 mai 2013 15:24

> À : general-filesystems

> Objet : devf-ram does not automount as described

>

> Per http://support7.qnx.com/download/download/20951/building.pdf I

> should be able to do the following on the Nitrogen6X development system

> (using default BSP).  Assume "my_efs_img" is on the micro-SD card and I have

> mounted it accordingly  "mount -t dos /dev/hd0t11 /dos".

>

> # devf-ram &

> # flashctl -p /dev/fs0 -e

> # cat /dos/my_efs_img > /dev/fs1

> # slay devf-ram &

> # devf-ram &

> # cd /fs1p0

> ksh: cd: /fs1p0 - No such file or directory

Maybe I`m missing something but for auto-mount to work the device has to exist. When devf-ram is killed the content  of corresponding device, memory, is destroyed. There is nothing to auto-mount from when you restart devf-ram.

>

> The last line will not work because the automount did not work properly.  Is

> this because there is already a flash filesystem mounted on the Nitrogen6X

> board (fs0)?  I mean I should be able to mount multiple flash filesystems

> right?

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post101837

> To cancel your subscription to this discussion, please e-mail general-

> filesystems-unsubscribe@community.qnx.com

post101839: Re: devf-ram does not automount as described

2013-05-29T19:30:55Z

Sorry, that was a typo, no it is all on /dev/fs1 like so...

# devf-ram &

# flashctl -p /dev/fs1 -e

# cat /dos/my_efs_img > /dev/fs1

# slay devf-ram &

# devf-ram &

# cd /fs1p0

ksh: cd: /fs1p0 - No such file or directory

post101838: Re: devf-ram does not automount as described

2013-05-29T19:28:00Z

You are erasing /dev/fs0, but then writing the image to /dev/fs1. Is this what you intended?

Sent from my BlackBerry 10 smartphone.

From: Erick Roane

Sent: Wednesday, May 29, 2013 3:23 PM

To: general-filesystems

Reply To: general-filesystems@community.qnx.com

Subject: devf-ram does not automount as described

Per http://support7.qnx.com/download/download/20951/building.pdf I should be able to do the following on the Nitrogen6X development system (using default BSP). Assume "my_efs_img" is on the micro-SD card and I have mounted it accordingly "mount -t dos /dev/hd0t11 /dos".

# devf-ram &

# flashctl -p /dev/fs0 -e

# cat /dos/my_efs_img > /dev/fs1

# slay devf-ram &

# devf-ram &

# cd /fs1p0

ksh: cd: /fs1p0 - No such file or directory

The last line will not work because the automount did not work properly. Is this because there is already a flash filesystem mounted on the Nitrogen6X board (fs0)? I mean I should be able to mount multiple flash filesystems right?

_______________________________________________

General

http://community.qnx.com/sf/go/post101837

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post101837: devf-ram does not automount as described

2013-05-29T19:23:32Z

Per http://support7.qnx.com/download/download/20951/building.pdf I should be able to do the following on the Nitrogen6X development system (using default BSP). Assume "my_efs_img" is on the micro-SD card and I have mounted it accordingly "mount -t dos /dev/hd0t11 /dos".

# devf-ram &

# flashctl -p /dev/fs0 -e

# cat /dos/my_efs_img > /dev/fs1

# slay devf-ram &

# devf-ram &

# cd /fs1p0

ksh: cd: /fs1p0 - No such file or directory

The last line will not work because the automount did not work properly. Is this because there is already a flash filesystem mounted on the Nitrogen6X board (fs0)? I mean I should be able to mount multiple flash filesystems right?

post101322: Re: Mounting qnx4 fs with errors

2013-05-13T08:38:54Z

hello

i have some problem. in harman&decker cic for bmw x1.

i can mount this partitions, but in fdisk i see only extendet part.

how i can recreate 5 77 partition in fdisk ? (qnx6.5)

post101321: Re: Mounting qnx4 fs with errors

2013-05-13T08:37:57Z

> Ok, I finally found it out. I rebooted the whole system and was able to mount

> everything. May be there was a lock put on a file somewhere.

> Thanks,

> David.

>

> Le 31.05.2012 17:03, David Billard a écrit :

> > Ok, I found the "mount -t qnx6". Now I have a mounting problem. I can mount

> the

> > partitions:

> > /dev/hd0t77.1

> > /dev/hd0t77.2

> > /dev/hd0t77.3

> > /dev/hd0t77.4

> > /dev/hd0t77.5

> >

> > But not the /dev/hd0t77 (which is the biggest partition but the numbering

> scheme

> > suggests that this partition is subdivided in 1 to 5 sub partitions). Does

> it

> > means that there is an unpartitionned (and unformatted) space on the drive?

> >

> > Many thanks,

> >

> > David.

> > Le 31.05.2012 16:37, David Billard a écrit :

> >> Hi David,

> >>

> >> Sorry to bother you with my problem. How can I mount a qnx6 file system

> (I'm

> >> using the boot CD)?

> >>

> >> Many thanks,

> >>

> >> David.

> >>

> >> Le 15.05.2012 16:35, David Sarrazin a écrit :

> >>> What OS version are you running?  Partition type 77 is normally reserved

> for partitions of type QNX4.   Fdisk should say "QNX4" , simply because the

> type is 77.  You can also try chkqnx6fs, on the off chance that the disks are

> formatted with the QNX6 file system.

> >>>

> >>> David

> >>>

> >>>> -----Original Message-----

> >>>> From: David Billard [mailto:community-noreply@qnx.com]

> >>>> Sent: May-15-12 10:33 AM

> >>>> To: general-filesystems@community.qnx.com

> >>>> Subject: Re: Mounting qnx4 fs with errors

> >>>>

> >>>> Hi David,

> >>>>

> >>>> it says that my partition is not a QNX4 fs. A rapid view with fdisk says

> >>>> "unknown loader". Is there any way to detect which fs is on the

> partitions

> >>>> (I'm almost sure it's qnx4, but perhaps am i wrong).

> >>>>

> >>>> many thanks,

> >>>>

> >>>> David.

> >>>>

> >>>> Le 15.05.2012 16:21, David Sarrazin a écrit :

> >>>>> First thing to try is chkfsys, which is the QNX4 file system checker

> tool.  If

> >>>> you don't want to make any changes to the disk, the -f option will

> prevent

> >>>> the tool from modifying anything.

> >>>>>

> >>>>>> -----Original Message-----

> >>>>>> From: David Billard [mailto:community-noreply@qnx.com]

> >>>>>> Sent: May-15-12 10:35 AM

> >>>>>> To: general-filesystems

> >>>>>> Subject: Mounting qnx4 fs with errors

> >>>>>>

> >>>>>> Dear all,

> >>>>>>

> >>>>>> I'm new to QNX. I would like to recover files from a qnx4 filesystem

> >>>>>> on a harddrive. The hardrive has 6 partitions. I used the BOOT-CD and

> >>>>>> the photon file manager. If I do a df, it sees :

> >>>>>>

> >>>>>> /dev/hd0

> >>>>>> /dev/hd0t77

> >>>>>> /dev/hd0t77.1

> >>>>>> /dev/hd0t77.2

> >>>>>> /dev/hd0t77.3

> >>>>>> /dev/hd0t77.4

> >>>>>> /dev/hd0t77.5

> >>>>>>

> >>>>>> If I try to mount /dev/hd0t77 or any of its declination, it says that

> >>>>>> the fs is corrupted.

> >>>>>>

> >>>>>> To my knowledge, the disk is not damaged. Do you see any reason for

> >>>>>> this message or any solution for me?

> >>>>>>

> >>>>>> Is there a way to fix the fs?

> >>>>>>

> >>>>>> Many thanks,

> >>>>>>

> >>>>>> david.

> >>>>>>

> >>>>>>

> >>>>>>

> >>>>>>

> >>>>>> _______________________________________________

> >>>>>>

> >>>>>> General

> >>>>>> http://community.qnx.com/sf/go/post93087

> >>>>>> To cancel your subscription to this discussion, please e-mail

> >>>>>> general- filesystems-unsubscribe@community.qnx.com

> >>>>>

> >>>>>

> >>>>>

> >>>>>

> >>>>> _______________________________________________

> >>>>>

> >>>>> General

> >>>>> http://community.qnx.com/sf/go/post93088

> >>>>> To cancel your subscription to this discussion, please e-mail

> >>>>> general-filesystems-unsubscribe@community.qnx.com

> >>>>

> >>>> --

> >>>> Prof. David Billard                         | David.Billard@hesge.ch

> >>>> University of Applied Sciences in Geneva    | tel: +41 22 388 1885

> >>>> Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

> >>>>

> >>>>

> >>>>

> >>>>

> >>>>

> >>>> _______________________________________________

> >>>>

> >>>> General

> >>>> http://community.qnx.com/sf/go/post93090

> >>>> To cancel your subscription to this discussion, please e-mail general-

> >>>> filesystems-unsubscribe@community.qnx.com

> >>>

> >>>

> >>>

> >>>

> >>> _______________________________________________

> >>>

> >>> General

> >>> http://community.qnx.com/sf/go/post93091

> >>> To cancel your subscription to this discussion, please e-mail general-

> filesystems-unsubscribe@community.qnx.com

> >>

> >

>

> --

> Prof. David Billard                         | David.Billard@hesge.ch

> University of Applied Sciences in Geneva    | tel: +41 22 388 1885

> Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

>

post101310: Re: ext2 (for removable media)

2013-05-10T18:53:02Z

Now with what is presumably the correct command:

  # mount -text2 /dev/umass0t131 /fs/T1

  mount: Can't mount /fs/T1 (type ext2)

  mount: Possible reason: Corrupted file system detected

Did the ext2 format outpace fs-ext2.so?

post101308: ext2 (for removable media)

2013-05-10T17:36:12Z

I am trying to figure out the fastest way to get data off of our embedded system post-flight and into the hands of the scientists for data analysis. We expect to be acquiring ~40GB per flight, and it needs to get to three different systems running both Linux and Windows for data analysis. Having done the math, I figured the fastest method would be to use eSATA to write to an external disk, then move the disk to the first target system and repeat 3 times. Looking at the matrix of supported filesystems, it looks like ext2 is the only one that is both writable under QNX and readable under both Linux and Windows.

I've got the drive, I formatted it with ext2 on a CentOS 6.4 system, brought it over to QNX, see that the partition shows up (/dev/umass0t131), but:

  # mount -T ext2 /dev/umass0t131 /fs/T1

  mount: Can't mount /fs/T1 (type ext2)

  mount: Possible reason: Invalid argument

Any ideas what's going on?

Any suggestions?

post101216: Re: How to get storage and memory information?

2013-05-07T13:22:41Z

Thank you very much.

I really appreciate your help.

于 2013/5/7 21:18, Chris Travis 写道:

> Sorry, but this has Restricted Source Access, so if you are a customer or a partner of QNX Software Systems and need access to QNX proprietary source code in this project please contact your QSS Sales or Support representative.

>

> Unless somebody has some code they can share with you for storage info.

>

> -----Original Message-----

> From: Chris Travis

> Sent: May-07-13 9:07 AM

> To: 'general-filesystems@community.qnx.com'

> Subject: RE: How to get storage and memory information?

>

> Actually you wouldn't have access to this source link...this might be available on foundry27 for you...let me look.

>

> -----Original Message-----

> From: Chris Travis

> Sent: May-07-13 9:05 AM

> To: general-filesystems

> Subject: RE: How to get storage and memory information?

>

> You could have a look at the df source for disk space:

> http://svn.ott.qnx.com/view/product/trunk/utils/d/df/nto/

>

> Below is some C code for getting free and total memory.

>

> Chris

>

> #include <errno.h>

> #include <stdint.h>

> #include <unistd.h>

> #include <stdarg.h>

> #include <stdlib.h>

> #include <stdio.h>

> #include <string.h>

> #include <sys/stat.h>

> #include <sys/mman.h>

> #include <sys/syspage.h>

>

> uint64_t

> get_free_mem (void)

> {

> int ret;

> struct stat64 sbuf;

> ret = stat64("/proc", &sbuf);

> if (ret == -1) {

> printf("stat(/proc) failed: %s", strerror(errno));

> }

> return sbuf.st_size;

> }

>

> uint64_t

> get_total_mem (void)

> {

> char *str = SYSPAGE_ENTRY(strings)->data;

> struct asinfo_entry *as = SYSPAGE_ENTRY(asinfo);

> uint64_t total = 0;

> unsigned num;

>

> for (num = _syspage_ptr->asinfo.entry_size / sizeof(*as); num > 0; --num) {

> if(strcmp(&str[as->name], "ram") == 0) {

> total += as->end - as->start + 1;

> }

> ++as;

> }

> return total;

> }

>

> int main (int argc, char **argv)

> {

> uint64_t free_mem = get_free_mem();

> uint64_t total_mem = get_total_mem();

>

> printf("FreeMem/TotalMem: %llu/%llu",free_mem,total_mem ); }

>

> -----Original Message-----

> From: Allen Chen [mailto:community-noreply@qnx.com]

> Sent: May-07-13 3:29 AM

> To: general-filesystems

> Subject: How to get storage and memory information?

>

> I want to implement a program in C++ to get system storage information, such as total storage size, free storage size, total memory size and free memory size.

> How to do it?

>

> Allen Chen

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post101198

> To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post101215

> To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

>

post101215: RE: How to get storage and memory information?

2013-05-07T13:18:48Z

Sorry, but this has Restricted Source Access, so if you are a customer or a partner of QNX Software Systems and need access to QNX proprietary source code in this project please contact your QSS Sales or Support representative.

Unless somebody has some code they can share with you for storage info.

-----Original Message-----

From: Chris Travis

Sent: May-07-13 9:07 AM

To: 'general-filesystems@community.qnx.com'

Subject: RE: How to get storage and memory information?

Actually you wouldn't have access to this source link...this might be available on foundry27 for you...let me look.

-----Original Message-----

From: Chris Travis

Sent: May-07-13 9:05 AM

To: general-filesystems

Subject: RE: How to get storage and memory information?

You could have a look at the df source for disk space:

http://svn.ott.qnx.com/view/product/trunk/utils/d/df/nto/

Below is some C code for getting free and total memory.

Chris

#include <errno.h>

#include <stdint.h>

#include <unistd.h>

#include <stdarg.h>

#include <stdlib.h>

#include <stdio.h>

#include <string.h>

#include <sys/stat.h>

#include <sys/mman.h>

#include <sys/syspage.h>

uint64_t

get_free_mem (void)

{

 int ret;

 struct stat64 sbuf;

 ret = stat64("/proc", &sbuf);

 if (ret == -1) {

 printf("stat(/proc) failed: %s", strerror(errno));

 }

 return sbuf.st_size;

}

uint64_t

get_total_mem (void)

{

 char *str = SYSPAGE_ENTRY(strings)->data;

 struct asinfo_entry *as = SYSPAGE_ENTRY(asinfo);

 uint64_t total = 0;

 unsigned num;



 for (num = _syspage_ptr->asinfo.entry_size / sizeof(*as); num > 0; --num) {

 if(strcmp(&str[as->name], "ram") == 0) {

 total += as->end - as->start + 1;

 }

 ++as;

 }

 return total;

}

int main (int argc, char **argv)

{

 uint64_t free_mem = get_free_mem();

 uint64_t total_mem = get_total_mem();

 printf("FreeMem/TotalMem: %llu/%llu",free_mem,total_mem ); }

-----Original Message-----

From: Allen Chen [mailto:community-noreply@qnx.com]

Sent: May-07-13 3:29 AM

To: general-filesystems

Subject: How to get storage and memory information?

I want to implement a program in C++ to get system storage information, such as total storage size, free storage size, total memory size and free memory size.

How to do it?

Allen Chen

_______________________________________________

General

http://community.qnx.com/sf/go/post101198

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post101213: RE: How to get storage and memory information?

2013-05-07T13:07:01Z

Actually you wouldn't have access to this source link...this might be available on foundry27 for you...let me look.

-----Original Message-----

From: Chris Travis

Sent: May-07-13 9:05 AM

To: general-filesystems

Subject: RE: How to get storage and memory information?

You could have a look at the df source for disk space:

http://svn.ott.qnx.com/view/product/trunk/utils/d/df/nto/

Below is some C code for getting free and total memory.

Chris

#include <errno.h>

#include <stdint.h>

#include <unistd.h>

#include <stdarg.h>

#include <stdlib.h>

#include <stdio.h>

#include <string.h>

#include <sys/stat.h>

#include <sys/mman.h>

#include <sys/syspage.h>

uint64_t

get_free_mem (void)

{

 int ret;

 struct stat64 sbuf;

 ret = stat64("/proc", &sbuf);

 if (ret == -1) {

 printf("stat(/proc) failed: %s", strerror(errno));

 }

 return sbuf.st_size;

}

uint64_t

get_total_mem (void)

{

 char *str = SYSPAGE_ENTRY(strings)->data;

 struct asinfo_entry *as = SYSPAGE_ENTRY(asinfo);

 uint64_t total = 0;

 unsigned num;



 for (num = _syspage_ptr->asinfo.entry_size / sizeof(*as); num > 0; --num) {

 if(strcmp(&str[as->name], "ram") == 0) {

 total += as->end - as->start + 1;

 }

 ++as;

 }

 return total;

}

int main (int argc, char **argv)

{

 uint64_t free_mem = get_free_mem();

 uint64_t total_mem = get_total_mem();

 printf("FreeMem/TotalMem: %llu/%llu",free_mem,total_mem ); }

-----Original Message-----

From: Allen Chen [mailto:community-noreply@qnx.com]

Sent: May-07-13 3:29 AM

To: general-filesystems

Subject: How to get storage and memory information?

I want to implement a program in C++ to get system storage information, such as total storage size, free storage size, total memory size and free memory size.

How to do it?

Allen Chen

_______________________________________________

General

http://community.qnx.com/sf/go/post101198

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post101212: RE: How to get storage and memory information?

2013-05-07T13:05:34Z

You could have a look at the df source for disk space:

http://svn.ott.qnx.com/view/product/trunk/utils/d/df/nto/

Below is some C code for getting free and total memory.

Chris

#include <errno.h>

#include <stdint.h>

#include <unistd.h>

#include <stdarg.h>

#include <stdlib.h>

#include <stdio.h>

#include <string.h>

#include <sys/stat.h>

#include <sys/mman.h>

#include <sys/syspage.h>

uint64_t

get_free_mem (void)

{

 int ret;

 struct stat64 sbuf;

 ret = stat64("/proc", &sbuf);

 if (ret == -1) {

 printf("stat(/proc) failed: %s", strerror(errno));

 }

 return sbuf.st_size;

}

uint64_t

get_total_mem (void)

{

 char *str = SYSPAGE_ENTRY(strings)->data;

 struct asinfo_entry *as = SYSPAGE_ENTRY(asinfo);

 uint64_t total = 0;

 unsigned num;



 for (num = _syspage_ptr->asinfo.entry_size / sizeof(*as); num > 0; --num) {

 if(strcmp(&str[as->name], "ram") == 0) {

 total += as->end - as->start + 1;

 }

 ++as;

 }

 return total;

}

int main (int argc, char **argv)

{

 uint64_t free_mem = get_free_mem();

 uint64_t total_mem = get_total_mem();

 printf("FreeMem/TotalMem: %llu/%llu",free_mem,total_mem );

}

-----Original Message-----

From: Allen Chen [mailto:community-noreply@qnx.com]

Sent: May-07-13 3:29 AM

To: general-filesystems

Subject: How to get storage and memory information?

I want to implement a program in C++ to get system storage information, such as total storage size, free storage size, total memory size and free memory size.

How to do it?

Allen Chen

_______________________________________________

General

http://community.qnx.com/sf/go/post101198

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post101198: How to get storage and memory information?

2013-05-07T07:28:58Z

I want to implement a program in C++ to get system storage information, such as total storage size, free storage size, total memory size and free memory size.

How to do it?

Allen Chen

post100577: Re: NOR flash driver doesn't recognize partition

2013-04-15T13:30:21Z

Hello Daniel Lee,

I have the same problem. When I try to formate the flash the same error occur. Do you have any solution?

Thank you!

Kevin

post100109: can't identify SATA or EIDE device

2013-03-22T17:36:27Z

I am using a SILICON Systems 80 GB SSD. not being identified at boot up.. i get the following slogger errors. continously.. for days as the system was left in this state for a week. Reboot helps sometimes.. Will setting DRDY off help.. where can i get more info on the driver - device interactions ? using QNX6.

20130313_170922551_slog 3/13/2013 17:07 261 13 0 SLOG_ERROR eide_command: !DRDY: 0

20130313_170922551_slog 3/13/2013 17:07 261 13 0 SLOG_ERROR eide_dump_regs: error 0 status 0

20130313_170922551_slog 3/13/2013 17:07 261 13 0 SLOG_ERROR eide_command: !DRDY: 0

20130313_170922551_slog 3/13/2013 17:07 262 13 0 SLOG_ERROR eide_dump_regs: error 0 status 0

post100010: Re: RE: FDisk - Partitioning multiple filesystems of the same type

2013-03-19T16:50:45Z

If I use fdisk to create the partitions, the naming of them is not under my control. Fdisk decides that the second partition is hd0t179.1, only if I make the first and second partitions t179; however, documentation says that t177 is primary and t178 is secondary for QNX6, so why would it allow me to have two t179s?

post100009: RE: FDisk - Partitioning multiple filesystems of the same type

2013-03-19T16:45:54Z

With t07? are of the type QNX4 filesystem ( compatible with QNX6 ), and t17? are of type QNX6.  Check to doc about these filesystem to chose the one that best fits your need.

I think the 4 partitions limit is that of a standard primary boot loader, some fancier boot loader can extend that though.

I'm not sure if the QNX partition numbering is part of some standard.  In theory you could use any number it doesn't really matter. All that matter is one partition be set as boot partition and that is it. The naming of the partition is totally under your control.  However some tool like diskboot, used in the default startup sequence,  may have some specific requirement, don't know about that.

-----Message d'origine-----

De : Derek McIntyre [mailto:community-noreply@qnx.com]

Envoyé : Tuesday, March 19, 2013 11:38 AM

À : general-filesystems

Objet : FDisk - Partitioning multiple filesystems of the same type

I'm finding a few discrepancies in QNX's documentation of partitioning and proper setup of multiple partitions of the same file system type, and was hoping someone could clarify for me.

1.) Primary t0177 and secondary t0178, or t0177 and t0177.1: which is correct?

Several parts of the documentation, as well as other sites, mention that t077/t0177 are the primary partitions while t078/t0178 are secondary and t079/t0179 are tertiary for qnx4 and qnx6 respectively; however, fdisk will allow you to create multiple t0179 partitions and will label them as /dev/hd0t179 and /dev/hd0t179.1.  Is the first correct? And if so, what is the reason for the second, or is it a mistake?

2.) Is there a limit of 4 partitions of the same file system on the same hard drive?

The documentation mentions this rule, but there are only 3 file system numbers listed.  This leads me to believe there is a 4th t080/t180 partition file system type: is that true?

Any information would be appreciated.

Thanks!

_______________________________________________

General

http://community.qnx.com/sf/go/post100005

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post100005: FDisk - Partitioning multiple filesystems of the same type

2013-03-19T15:37:42Z

I'm finding a few discrepancies in QNX's documentation of partitioning and proper setup of multiple partitions of the same file system type, and was hoping someone could clarify for me.

1.) Primary t0177 and secondary t0178, or t0177 and t0177.1: which is correct?

Several parts of the documentation, as well as other sites, mention that t077/t0177 are the primary partitions while t078/t0178 are secondary and t079/t0179 are tertiary for qnx4 and qnx6 respectively; however, fdisk will allow you to create multiple t0179 partitions and will label them as /dev/hd0t179 and /dev/hd0t179.1. Is the first correct? And if so, what is the reason for the second, or is it a mistake?

2.) Is there a limit of 4 partitions of the same file system on the same hard drive?

The documentation mentions this rule, but there are only 3 file system numbers listed. This leads me to believe there is a 4th t080/t180 partition file system type: is that true?

Any information would be appreciated.

Thanks!

post99755: Re: RE: Performing read operation on fd after mmap

2013-03-08T16:12:53Z

Hi David,

thanks for the input, so do you mean after mmap on a shm file and using the file fd directly or using the pointer returned by mmap doesnot give any performance benefit.

Then why do we need to have mmap ? can the shm file fd be directly used for transfer without using mmap ?

post99744: RE: Performing read operation on fd after mmap

2013-03-07T22:37:35Z

Using the fd directly will perform a file operation, and not modify the mmap memory buffers. When the memory buffers are unmapped, their contents will overwrite whatever changes you made to the file. For this reason, you should only access the file via the mmap memory buffers OR through the fd, do not mix them.

Your application will know better than the kernel when you are finished with a specific piece of buffer, so you application can make better decisions about when to flush the changes to disk. Other than this, I don't know if there is any performance penalty for using mmap versus reading the file directly. I would suggest though, if you are looking for performance, make sure you do NOT use fopen()/fread()/fwrite(), and do your file access in the largest possible buffer (32k or 64k read and write sizes offer good performance).

David

-----Original Message-----

From: Balaaji Tirouvengadam [mailto:community-noreply@qnx.com]

Sent: Thursday, March 07, 2013 1:59 PM

To: general-filesystems

Subject: Performing read operation on fd after mmap

Hi,

I am trying to find the performance difference between

1.) using the pointer returned by mmap() of shm_open() retruned fd

2.) shm_open(), call mmap(), ignore the pinter returned and just use the fd

for data read/write operations.

Does the both operation gives same performance results ? wehther the both does only memory operation instead of one memory and the other as file operation ?

_______________________________________________

General

http://community.qnx.com/sf/go/post99738

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post99738: Performing read operation on fd after mmap

2013-03-07T18:58:59Z

Hi,

I am trying to find the performance difference between

1.) using the pointer returned by mmap() of shm_open() retruned fd

2.) shm_open(), call mmap(), ignore the pinter returned and just use the fd

for data read/write operations.

Does the both operation gives same performance results ? wehther the both does only memory operation instead of one memory and the other as file operation ?

post98638: Re: RE: RE: RE: ifs with ftp cabapility?

2013-01-17T17:08:37Z

Adding that file solved my problem. Of course I added a shadow file for password which was created in a writable file system. Thanks for your help.

Regards

post98591: Re: RE: Effect of Duplicate inode numbers in resource managed files

2013-01-16T09:40:07Z

Thanks for the reply David, the vast majority of the access to the files is with read() and write() after and open() call. Am I correct in assuming that when a file is open()ed by name it will always open the correct file regardless of the inode number ?

post98580: RE: Effect of Duplicate inode numbers in resource managed files

2013-01-15T16:29:40Z

Graham,

If you are attempting to execute programs from this resmgr, then you'll need to ensure the inodes are unique, as the process manager uses the inode to identify open executable files. Similiarily, nfsd and fs-nfs3 rely on the inode number for caching reasons. It depends on what clients are accessing this resource manager, and what the client is doing with the inode number.

David

> -----Original Message-----

> From: Watson, Graham [mailto:community-noreply@qnx.com]

> Sent: January-15-13 11:10 AM

> To: general-filesystems@community.qnx.com

> Subject: RE: Effect of Duplicate inode numbers in resource managed files

>

> 

> 

> HI Mario

>

> Thanks for your reply, I understand was an inode is and have had a lot of

> experience with UNIX filesystems. My question in more to do with the

> implications of having duplicate inodes in systems out in the field and what

> possible problems might arise. The systems we run control very heavy and

> expensive equipment and I would not want anything going wrong and

> someone possible getting hurt. If you can give me anymore info on what

> possible side effects might occour I would be very grateful.

>

> Regards

>

> Graham Watson | Software Engineer

> Waukesha Magnetic Bearings

> Downlands Business Park, Lyons Way, Worthing West Sussex BN14 9LA

> United Kingdom

> Mobile: +44 (0)7762 541863 | Office: +44 (0)1903 275543

> gwatson@waukbearing.com www.waukbearing.com P Please consider the

> environment before you print.

>

> -----Original Message-----

> From: Mario Beaulieu [mailto:community-noreply@qnx.com]

> Sent: 15 January 2013 14:22

> To: general-filesystems

> Subject: Re: Effect of Duplicate inode numbers in resource managed files

>

> The inode is a filesystem data structure that represents a file (its location on

> disk, owner, permissions, etc).

> If 2 "files" on the same partition have the same inode, they are actually 2

> names for the same file.

>

> You can create many names for the same file with the command "ln".

> When you do "ls -l", the first value after the permissions is the number of

> files pointing to that inode, or should I say, the number of filenames

> corresponding to that file.

> Below, I have created file "myfile" and the link number is 1.

> After creating a second name for that same file with command "ln", we see

> that the number of links is 2 for each of those:

>

> # echo "Hello" > myfile

> # ls -l

> total 1

> -rw-rw-r-- 1 root nto 6 Jan 15 09:07 myfile

> # ln myfile AnotherLinkToMyFyle

> # ls -l

> total 2

> -rw-rw-r-- 2 root nto 6 Jan 15 09:07 AnotherLinkToMyFyle

> -rw-rw-r-- 2 root nto 6 Jan 15 09:07 myfile

> # echo "Hello again" > otherFile

> # ls -l

> total 3

> -rw-rw-r-- 2 root nto 6 Jan 15 09:07 AnotherLinkToMyFyle

> -rw-rw-r-- 2 root nto 6 Jan 15 09:07 myfile

> -rw-rw-r-- 1 root nto 12 Jan 15 09:12 otherFile

> #

>

> See the inode number like the page number in a book. In addition to the

> inode, you have to check the partition. Inodes on different partitions are

> completely unrelated, like page numbers on 2 different books.

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post98567

> To cancel your subscription to this discussion, please e-mail general-

> filesystems-unsubscribe@community.qnx.com

> Waukesha Bearings Limited is a limited company registered

> in England and Wales. Registered number: 4203526. Registered office:

> 53-55 The Broadway, Joel Street, Northwood, Middlesex, HA6 1NZ 

>

> PLEASE NOTE! This message (and any associated

> files) is intended only for the use of the individual or entity to which it is

> addressed and may contain information that is confidential, subject to

> copyright or constitutes a trade secret. If you are not the intended recipient

> you are hereby notified that any dissemination, copying or distribution of this

> message, or files associated with this message, is strictly prohibited. If you

> have received this message in error, please notify us immediately by replying

> to the message and deleting it from your computer. Messages sent to and

> from us may be monitored.

>

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post98577

> To cancel your subscription to this discussion, please e-mail general-

> filesystems-unsubscribe@community.qnx.com

post98577: RE: Effect of Duplicate inode numbers in resource managed files

2013-01-15T16:12:38Z

HI Mario

Thanks for your reply, I understand was an inode is and have had a lot of experience with UNIX filesystems. My question in more to do with the implications of having duplicate inodes in systems out in the field and what possible problems might arise. The systems we run control very heavy and expensive equipment and I would not want anything going wrong and someone possible getting hurt. If you can give me anymore info on what possible side effects might occour I would be very grateful.

Regards

Graham Watson | Software Engineer

Waukesha Magnetic Bearings

Downlands Business Park, Lyons Way, Worthing

West Sussex BN14 9LA United Kingdom

Mobile: +44 (0)7762 541863 | Office: +44 (0)1903 275543

gwatson@waukbearing.com

www.waukbearing.com

 Please consider the environment before you print.

-----Original Message-----

From: Mario Beaulieu [mailto:community-noreply@qnx.com]

Sent: 15 January 2013 14:22

To: general-filesystems

Subject: Re: Effect of Duplicate inode numbers in resource managed files

The inode is a filesystem data structure that represents a file (its location on disk, owner, permissions, etc).

If 2 "files" on the same partition have the same inode, they are actually 2 names for the same file.

You can create many names for the same file with the command "ln".

When you do "ls -l", the first value after the permissions is the number of files pointing to that inode, or should I say, the number of filenames corresponding to that file.

Below, I have created file "myfile" and the link number is 1.

After creating a second name for that same file with command "ln", we see that the number of links is 2 for each of those:

# echo "Hello" > myfile

# ls -l

total 1

-rw-rw-r-- 1 root nto 6 Jan 15 09:07 myfile

# ln myfile AnotherLinkToMyFyle

# ls -l

total 2

-rw-rw-r-- 2 root nto 6 Jan 15 09:07 AnotherLinkToMyFyle

-rw-rw-r-- 2 root nto 6 Jan 15 09:07 myfile

# echo "Hello again" > otherFile

# ls -l

total 3

-rw-rw-r-- 2 root nto 6 Jan 15 09:07 AnotherLinkToMyFyle

-rw-rw-r-- 2 root nto 6 Jan 15 09:07 myfile

-rw-rw-r-- 1 root nto 12 Jan 15 09:12 otherFile

#

See the inode number like the page number in a book. In addition to the inode, you have to check the partition. Inodes on different partitions are completely unrelated, like page numbers on 2 different books.

_______________________________________________

General

http://community.qnx.com/sf/go/post98567

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

Waukesha Bearings Limited is a limited company registered in England and Wales. 

Registered number: 4203526. Registered office: 53-55 The Broadway, Joel Street, Northwood, Middlesex, HA6 1NZ 

PLEASE NOTE! This message (and any associated files) is intended only for the use of the individual or entity to which it is addressed and may contain information that is confidential,

subject to copyright or constitutes a trade secret. If you are not the intended recipient you are hereby notified that any dissemination, copying or distribution of this message,

or files associated with this message, is strictly prohibited. If you have received this message in error, please notify us immediately by replying to the message and deleting it from

your computer. Messages sent to and from us may be monitored.

post98567: Re: Effect of Duplicate inode numbers in resource managed files

2013-01-15T14:22:17Z

The inode is a filesystem data structure that represents a file (its location on disk, owner, permissions, etc).

If 2 "files" on the same partition have the same inode, they are actually 2 names for the same file.

You can create many names for the same file with the command "ln".

When you do "ls -l", the first value after the permissions is the number of files pointing to that inode, or should I say, the number of filenames corresponding to that file.

Below, I have created file "myfile" and the link number is 1.

After creating a second name for that same file with command "ln", we see that the number of links is 2 for each of those:

# echo "Hello" > myfile

# ls -l

total 1

-rw-rw-r--   1 root      nto               6 Jan 15 09:07 myfile

# ln myfile AnotherLinkToMyFyle

# ls -l

total 2

-rw-rw-r--   2 root      nto               6 Jan 15 09:07 AnotherLinkToMyFyle

-rw-rw-r--   2 root      nto               6 Jan 15 09:07 myfile

# echo "Hello again" > otherFile

# ls -l

total 3

-rw-rw-r--   2 root      nto               6 Jan 15 09:07 AnotherLinkToMyFyle

-rw-rw-r--   2 root      nto               6 Jan 15 09:07 myfile

-rw-rw-r--   1 root      nto              12 Jan 15 09:12 otherFile

#

See the inode number like the page number in a book. In addition to the inode, you have to check the partition. Inodes on different partitions are completely unrelated, like page numbers on 2 different books.

post98563: Effect of Duplicate inode numbers in resource managed files

2013-01-15T09:35:51Z

Hi,

We have a system running on QNX 6.4.1 which makes use of filesystem resource managers. During an investigation it was discovered that several files in different directories have the same inode number. The inode number is assigned by the resource manager and is not left with a zero value to have one assigned by the O/S. Can anyone give me any information they might have that details the possible implications of having files in the same filesystem with non unique inode numbers. The files themselves are accessed in a variety of ways,

From PHP, from O/S commands such as cat, more etc. and from O/S APIs open() and fopen().

The easy solution as you might point out is to make sure the inode numbers are non-unique, but we have systems in the field and urgently need to know if there are any known issues.

Thanks for your help

post98550: Re: RE: RE: RE: ifs with ftp cabapility?

2013-01-14T18:49:25Z

Thanks guys. I thought that file just only for restriction and ignored it. I will try it tomorrow and inform you.

Regards.

post98549: RE: RE: RE: ifs with ftp cabapility?

2013-01-14T16:45:51Z

Good call on the /etc/ftpusers, I had forgotten about that file. Also, ignore the mount instructions that is for a device with a writable filesystem mounted as /base ;-)

-----Original Message-----

From: Mario Beaulieu [mailto:community-noreply@qnx.com]

Sent: January-14-13 10:17 AM

To: general-filesystems

Subject: Re: RE: RE: ifs with ftp cabapility?

To use FTP, you will need a file:

/etc/ftpusers

If this file is absent, all ftp accesses will be denied.

In order to write in /etc (and in a few other directories), you need to remount the filesystem as writable.

To do that:

mount -uw /base

... now the system is writable

... do your stuff

... and restore it to read-only when you are done mount -ur /base

Note that you cannot overwrite or modify a file. To do such a thing, you will need to rename or delete the original, then create the "modified" version.

_______________________________________________

General

http://community.qnx.com/sf/go/post98546

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post98546: Re: RE: RE: ifs with ftp cabapility?

2013-01-14T15:16:42Z

To use FTP, you will need a file:

/etc/ftpusers

If this file is absent, all ftp accesses will be denied.

In order to write in /etc (and in a few other directories), you need to remount the filesystem as writable.

To do that:

mount -uw /base

... now the system is writable

... do your stuff

... and restore it to read-only when you are done

mount -ur /base

Note that you cannot overwrite or modify a file. To do such a thing, you will need to rename or delete the original, then create the "modified" version.

post98538: Re: RE: RE: ifs with ftp cabapility?

2013-01-13T22:56:48Z

A few options. Once you telnet into the target you can run ftp "from" the target to get the file to the target(in/dev/shmem), or if you really need to ftp into the target then you can put writable files into /dev/shmem ; just create a link in the build file from /etc/shadow to /dev/shmem/shadow. You could also use devf-ram to create a ram filesystem...that you could write to.

Devb suggestion wasn't mine.

Sent from my BlackBerry 10 smartphone.

From: enes aykurt

Sent: Sunday, January 13, 2013 4:57 PM

To: general-filesystems

Reply To: general-filesystems@community.qnx.com

Subject: Re: RE: RE: ifs with ftp cabapility?

Hi Chris,

I have already done the telnet, my problem is about ftp. For ftp as you know I need also the shadow file because ftp user must have a password. However I could not figure out how to form and locate an encrypted file in ifs.I tried to form a shadow file by using a QNX disk system and used it on my ifs however that didnt work.

Tomorrow I will try the devb driver modification that you have suggested. Can you please explain me the affect of this option ?

regards

_______________________________________________

General

http://community.qnx.com/sf/go/post98536

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post98536: Re: RE: RE: ifs with ftp cabapility?

2013-01-13T21:59:57Z

Hi Chris,

I have already done the telnet, my problem is about ftp. For ftp as you know I need also the shadow file because ftp user must have a password. However I could not figure out how to form and locate an encrypted file in ifs.I tried to form a shadow file by using a QNX disk system and used it on my ifs however that didnt work.

Tomorrow I will try the devb driver modification that you have suggested. Can you please explain me the affect of this option ?

regards

post98535: RE: RE: ifs with ftp cabapility?

2013-01-13T15:25:08Z

That will not help as you are not using a devb driver.

You need to include all of the files necessary in the ifs including /etc/passwd

Here is a link from the Support Knowledge Base that shows a build image with the necessary files that allows you to telnet into it:

http://www.qnx.com/support/knowledgebase.html?id=50130000000SzKu

Chris

-----Original Message-----

From: enes aykurt [mailto:community-noreply@qnx.com]

Sent: January-13-13 9:40 AM

To: general-filesystems

Subject: Re: RE: ifs with ftp cabapility?

can you please send me the documentation about that ?

_______________________________________________

General

http://community.qnx.com/sf/go/post98534

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post98534: Re: RE: ifs with ftp cabapility?

2013-01-13T14:39:30Z

can you please send me the documentation about that ?

post98531: RE: ifs with ftp cabapility?

2013-01-12T18:03:42Z

Try passing -osync=optional to devb-*

-----Message d'origine-----

De : enes aykurt [mailto:community-noreply@qnx.com]

Envoyé : Saturday, January 12, 2013 10:58 PM

À : general-filesystems

Objet : ifs with ftp cabapility?

Hi all,

I have got an intel atom processor card and I am booting it by using an SSD. I created an ifs image and booted the card with it. However after boot I could not write/modify the etc, usr, bin, lib files on target. I think that is because of the image's read only property. But I need some services which will use those files dynamically.

For example I need ftp service and as far as I know to use ftp a user should have a password. However to change the password, passwd utility using/modfiying the files under /etc. How can I manage to make those files writable? Where am I doing wrong? If someone help me I ll appretiate.

Thanks and regards

_______________________________________________

General

http://community.qnx.com/sf/go/post98528

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post98528: ifs with ftp cabapility?

2013-01-12T09:57:35Z

Hi all,

I have got an intel atom processor card and I am booting it by using an SSD. I created an ifs image and booted the card with it. However after boot I could not write/modify the etc, usr, bin, lib files on target. I think that is because of the image's read only property. But I need some services which will use those files dynamically.

For example I need ftp service and as far as I know to use ftp a user should have a password. However to change the password, passwd utility using/modfiying the files under /etc. How can I manage to make those files writable? Where am I doing wrong? If someone help me I ll appretiate.

Thanks and regards

post98368: Re: Help with mketfs buildfile

2013-01-04T05:41:50Z

You seem to be using a OS Image Filesystem (IFS) build file with the mketfs utility which creates an Embedded Transaction FileSystem (ETFS). They are not the same things and the difference and usage is documented, please have a look:

http://www.qnx.com/developers/docs/6.5.0_sp1/topic/com.qnx.doc.neutrino_building/building_nto.html?cp=13_2_3

http://www.qnx.com/developers/docs/6.5.0_sp1/topic/com.qnx.doc.neutrino_utilities/m/mkifs.html?cp=13_13_15_18

http://www.qnx.com/developers/docs/6.5.0_sp1/topic/com.qnx.doc.neutrino_utilities/m/mketfs.html?cp=13_13_15_15

[virtual] and [+script] are not supported in an mketfs build file.

post98331: Help with mketfs buildfile

2012-12-28T19:57:34Z

Hi, I'm trying to run mketfs to create the filesystem with the following buildfile:

[virtual=x86,bios +compress] .bootstrap = {

    startup-bios

    PATH=/proc/boot:/bin procnto

}

[+script] .script = {

    devc-con -e -n5 &

    reopen /dev/con1

    devf-i365sl -r -b3 -m2 -u2 -t4 &

    waitfor /fs0p0

    [+session] TERM=qansi PATH=/proc/boot:/bin esh &

}

[type=link] /tmp=/dev/shmem

[type=link] /bin=/fs0p0/bin

[type=link] /etc=/fs0p0/etc

libc.so

[type=link] /usr/lib/ldqnx.so.2=/proc/boot/libc.so

libsocket.so

[data=copy]

devf-i365sl

devc-con

esh

----------------------------------------------

When I run the mketfs command I get an error with the [virtual] attribute. Is there a replace for this attribute? Is it needed with mketfs?

Thanks in advance.

I'm running QNX 6.4.1

post98111: Re: NOR flash driver doesn't recognize partition

2012-12-14T08:28:46Z

> Hello,

>

> I am writing a FFS3 NOR flash driver for an i.MX31 based board. The device is

> a Spansion S71WS128PC0HF3 burst-mode NOR Flash (16MB, 16 bit bus width)

> connected to CS5 of the WEIM. I can boot the system from this flash and I

> created an EFS and burnt it into the flash via JTAG debugger. I wrote a small

> test program that simply maps in the base address and scans for the 'QSSL'

> signature of the partition - it finds it at the correct address. However, when

>  I run my flash driver it doesn't recognise it.

>

> The auto-detect of the device works ((devf  t1::f3s_flash_probe:248) chip

> total = 1, bus_width = 2, interleave = 1)

> Erasing via flashctl works.

> flashctl -p/dev/fs0 -o8M -l8M -e -v

> Erasing device /dev/fs0

> ................................................................

> (devf  t2::f3s_table_find:66) fs0p0 raw U: 80

> #

> Formatting fails, though:

> # flashctl -p/dev/fs0p0 -o8M -l8M -f -v

> Formatting device /dev/fs0p0

> amd_poll: 163 program error

> (devf  t2::amd_v2wordwrite:113) over poll waiting for write completion

> (devf  t2::f3s_ext_write:197) fs0p0 bad H[00] P[40] # 000010

> DCMD_F3S_FORMAT failed (errno 5)

> flashctl: format failed

>

> I tried various write() functions of the MTD library but always the same.

> Maybe the writing works but reading doesn't? I supply NULL for the v2read

> callout, so it should use memcpy(), which works, as seen from my test program.

>

>

> Any ideas greatly appriciated. Thank you.

>

> -Albrecht

hello , Gunter Geigemueller

      I met the same problem just the same as you.I can erase my flash use flashctl but can't format it, how did you fix it? Thank you very much.

Formatting device /dev/fs0p0

amd_poll: 163 program error

(devf  t2::amd_v2wordwrite:113) over poll waiting for write completion

(devf  t2::f3s_ext_write:197) fs0p0 bad H[00] P[40] # 000010

DCMD_F3S_FORMAT failed (errno 5)

flashctl: format failed

post97709: Removable EIDE Drive?

2012-12-01T17:59:31Z

This isn't strictly a filesystem question, but this seems like the best place to ask:

We have the following configuration:

  Secondary Master: Compact Flash (boot device, shows up as /dev/hd0)

  Primary Master: Fixed Hard Drive (shows up as /dev/hd1)

  Primary Slave: Removable Hard Drive (shows up as /dev/hd2)

All drives are supported by devb-eide.

The question is how to best manage the device removal and replacement. Under QNX4 we would umount /dev/hd2, then remove the drive. On replacement, we would mount -e /dev/hd2, then mount the partitions (no automount).

Does devb-eide support a removable drive? I know that /dev/hd2 did not appear after the drive was inserted the first time. It showed up after a reboot.

Worst case would be to shutdown the system whenever the drive needs to be inserted or removed. Other options might be shutting down the driver and restarting it as necessary, but I'm guessing I would have to first make sure I have separate driver invocations for primary and secondary, then umount both /dev/hd1 and /dev/hd2 before stopping the driver.

Is there any facility for asking devb-eide to rescan the interfaces?

Any other suggestions?

post97484: make (e=5): Access is denied error during compilation

2012-11-27T13:25:21Z

I am getting this error during compilation of my code in a XP Pc. I have only NX installed for gcc.

process_begin: CreateProcess(C:\Projects.Local\Embedded\SB\Core\Main\Tools\DiagUtil.com, ../../../Core/Main/Tools/DiagUtil.com /DiagCodesFile:DiagDef.h /Module:PPDA /UseClass:Pack /ImpXls:../../../Core/Main/Config_Files/CommonDiagnostics.xls /ImpXls:C:/Projects.Local/Embedded/SB/IOPack/Main/PPDA/Config_Files/PPDA_Diagnostics.xls /ExpDiagCodes:C:/Projects.Local/Embedded/SB/IOPack/Main/PPDA /ExpDiagMsgs:C:/Projects.Local/Embedded/SB/IOPack/Main/PPDA, ...) failed.

make (e=5): Access is denied.

make[1]: *** [diagsrc] Error 5

Please suggest.

post96803: Re: the doc about cam_configure

2012-11-02T15:48:43Z

> When I read the bsp code , I saw cam_configure() in the devb directory. This

> function is defined in libcam.so. Where I can get its doc or description? I

> just could not find any infomation in the user guide.

>

> BR

> Jerry

Has anyone ver replied to this post.

I could use the same info.

RJG

post96537: Re: RE: Maximum limit for the number of directories and directory can contain?

2012-10-23T06:23:53Z

From my own experience I would recommend checking number of free inodes on the disk, as it can be limiting the number of possible files on the disk. If necessary, create the filesystem with appropriate number of inodes. You can see these number with "chkqnx6fs -s" command

Regards, PKY

> The qnx6fs does not have any specific limit to the number of files or

> directories that can be present on the disk. Of course disk space and

> performance issues should be considered. File and directory look ups have an

> O(N) cost.

>

> -----Original Message-----

> From: Mark Bangert [mailto:community-noreply@qnx.com]

> Sent: October-18-12 5:40 AM

> To: general-filesystems

> Subject: Re: Maximum limit for the number of directories and directory can

> contain?

>

> QNX6

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post96424

> To cancel your subscription to this discussion, please e-mail general-

> filesystems-unsubscribe@community.qnx.com

post96531: RE: Maximum limit for the number of directories and directory can contain?

2012-10-22T22:57:26Z

The qnx6fs does not have any specific limit to the number of files or directories that can be present on the disk. Of course disk space and performance issues should be considered. File and directory look ups have an O(N) cost.

-----Original Message-----

From: Mark Bangert [mailto:community-noreply@qnx.com]

Sent: October-18-12 5:40 AM

To: general-filesystems

Subject: Re: Maximum limit for the number of directories and directory can contain?

QNX6

_______________________________________________

General

http://community.qnx.com/sf/go/post96424

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post96424: Re: Maximum limit for the number of directories and directory can contain?

2012-10-18T12:40:11Z

QNX6

post96422: Re: Maximum limit for the number of directories and directory can contain?

2012-10-18T12:01:47Z

What filesystem type?

post96397: Creating SCSI standard Target Device

2012-10-17T16:44:15Z

We recently got USB 3.0 device driver up and running on QNX for our board. I wanted to access a EMMC area available on the board and make it available as a storage area in the Host. In short I want my device to behave as a mass storage device. Is there and QNX SCSI related driver already available or do I have to write this driver from scratch. I recently saw a discussion on a libcam.so. Was wondering if I can use that for my operation. The EMMC area i have available is mounted and visible in \dev.

post96394: Maximum limit for the number of directories and directory can contain?

2012-10-17T15:23:00Z

Does anyone know if or what the maximum number of directories that can be created is. Our system creates a time stamped log file file in a new directory each time the system starts. Is there a limit on how many times this can be done??

post95713: Re: Mounting QNX6 new filesystems from other operating system

2012-09-20T14:26:18Z

Yes, i want that.

I think there are already exist something like that, but only for old qnx4-filesystems, not for qnx6-filesystems ? Like: http://qnxfs.narod.ru/

post95711: Re: Mounting QNX6 new filesystems from other operating system

2012-09-20T13:59:53Z

You have a QNX6 FS on your disk, you booted Linux and you want Linux to see your files?

A good question for a Linux forum. For almost sure, the answer is going to be "no."

From: Esa Heikkinen <community-noreply@qnx.com<mailto:community-noreply@qnx.com>>

Reply-To: <general-filesystems@community.qnx.com<mailto:general-filesystems@community.qnx.com>>

Date: Thu, 20 Sep 2012 07:27:05 -0400

To: general-filesystems <general-filesystems@community.qnx.com<mailto:general-filesystems@community.qnx.com>>

Subject: Mounting QNX6 new filesystems from other operating system

Is it possible to mount new QNX6 filesystems (type 177, 178 and 179) from other operating systems like Linux ?

_______________________________________________

General

http://community.qnx.com/sf/go/post95697

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com<mailto:general-filesystems-unsubscribe@community.qnx.com>

post95697: Mounting QNX6 new filesystems from other operating system

2012-09-20T11:27:05Z

Is it possible to mount new QNX6 filesystems (type 177, 178 and 179) from other operating systems like Linux ?

post95523: Re: RAM disk from NAND space

2012-09-12T13:01:16Z

I think I have found what I am looking for:

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_technotes/reloadable_ifs.html

post95508: RAM disk from NAND space

2012-09-11T17:55:53Z

Hello,

I'm currently working with a system that uses a NAND resident ETFS filesystem.  Since there are a large amount of executibles run here I was thinking that maybe performance could be increased by loading all the executibles/libraries in memory.  I'm thinking this look something like a contiguous image stored in NAND (outside of ETFS) that gets copied to RAM, mounted and processes are spawned.  Similar in nature to the boot image (or in linux initrd) but loaded after the kernel has begun.

Does this line of thinking make sense to people more experienced with QNX?

I was looking at devb-ram but this does not appear to be able to restore from an image.  Is this something anyone has played with?

Alternatively is there a way to have the ETFS system cached in RAM with write-back access?  Although for this to show benefit I would imagine that the entire partition would need to be pre-cached?

Thanks,

Devan

post95147: Re: RE: RE: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-27T14:21:12Z

The only reason I found this was as a result of a search for the string "Press F1-F4 to select drive or select partition 1,2,3? 1". This link kind of outlines everything.

http://www.qnx.com/developers/docs/6.3.2/neutrino/user_guide/starting.html

Again, thanks for the help.

Mark

post95144: RE: RE: RE: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-27T14:08:44Z

Excellent news! My pleasure...

Can you send me a link to the documentation you found?

Chris

-----Original Message-----

From: Mark Bangert [mailto:community-noreply@qnx.com]

Sent: August-27-12 10:07 AM

To: general-filesystems

Subject: Re: RE: RE: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

Hi Chris,

Success, thought you would like to know. The final piece of the puzzle was "dloader". I didn't realize that two loaders were required. It now makes sense. I found a piece of documentation over the weekend that solved the problem.

Thank you ever so much for all of your help.

Best Regards,

Mark

_______________________________________________

General

http://community.qnx.com/sf/go/post95143

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post95143: Re: RE: RE: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-27T14:06:51Z

Hi Chris,

Success, thought you would like to know. The final piece of the puzzle was "dloader". I didn't realize that two loaders were required. It now makes sense. I found a piece of documentation over the weekend that solved the problem.

Thank you ever so much for all of your help.

Best Regards,

Mark

post95131: Re: RE: RE: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-24T20:21:04Z

Chris,

I tried the mkqnxfs -B /dev/hd1T177 command and still no fire. If there is someone else that you think might be able to help that would be great. I really appreciate your help, you've been great.

Have a good weekend.

Mark

post95126: RE: RE: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-24T17:07:27Z

I'm not sure at this point, but I would think that the boot loader would have been written with the "mkqnx6fs /dev/hd1t177" command...but try rewriting the boot loader with:

"mkqnx6fs -B /dev/hd1t177"

If this doesn't work, then I'll get one of the file system guys to jump in.

-----Original Message-----

From: Mark Bangert [mailto:community-noreply@qnx.com]

Sent: August-24-12 11:08 AM

To: general-filesystems

Cc: DWarnock@Medrobotics.com

Subject: Re: RE: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

Chris,

I'm so close. I re-did the "mkqnx6fs /dev/hd1t177" command just in case screwed up something. I then moved a copy of my bootable ifs image that's on hd0 to the /fs/hd1-qnx6/.boot directory. I know that this is a working image. I remove hd0 and make hd1 my new hd0 and boot. I get the qnx "v1.2b Boot loader:" message but that's it. What I have noticed is that there is a previous message that I get with the working hd0 "Press F1-F4 to select drive or select Partition:", I don't get this message with my hd.

Also another bit of information. If I have both drives in the system and when it boots I can select drive 2 and it will boot from drive 2. So it seems to me that I'm still missing something in the preperation of my drive. Something is still missing.

Regards,

Mark

_______________________________________________

General

http://community.qnx.com/sf/go/post95117

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post95117: Re: RE: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-24T15:08:11Z

Chris,

I'm so close. I re-did the "mkqnx6fs /dev/hd1t177" command just in case screwed up something. I then moved a copy of my bootable ifs image that's on hd0 to the /fs/hd1-qnx6/.boot directory. I know that this is a working image. I remove hd0 and make hd1 my new hd0 and boot. I get the qnx "v1.2b Boot loader:" message but that's it. What I have noticed is that there is a previous message that I get with the working hd0 "Press F1-F4 to select drive or select Partition:", I don't get this message with my hd.

Also another bit of information. If I have both drives in the system and when it boots I can select drive 2 and it will boot from drive 2. So it seems to me that I'm still missing something in the preperation of my drive. Something is still missing.

Regards,

Mark

post95109: RE: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-24T05:50:59Z

/fs/hd1-qnx6/.boot/ is a directory, which is where you need to put you're ifs.  You may have written over your /fs/hd1-qnx6/.boot/ directory with your cp command ;-)  If so, you would have to rerun the mkqnx6fs command to recreate the qnx6 filesystem.

When you remove the hd0 drive then reboot; the loader(on the secondary disk which is now the primary) should show your ifs as a selection to boot from.

Once booted, as long as you have the io-blk driver(and required libs) running in your ifs, then the hard-drive should show up as

/dev/hd0 and hopefully be mounted as /

Now this filesystem will be empty unless you've copied files over to the /fs/hd1-qnx6/ previous to removing the initial disk.

If you get as far as selecting you're image and it doesn't boot properly, post your myqnx.build file.

Chris

-----Original Message-----

From: Mark Bangert [mailto:community-noreply@qnx.com]

Sent: August-23-12 5:36 PM

To: general-filesystems

Subject: Re: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

Almost there.  After performing the above, and rebooting I get a meessage on the screen "QNX v1.2b Boot Loader" and a blinking cursor.  I think the issue is not being able to copy my ifs from one drive to another.  This is basic but I'm have trouble.  I believe the command should be

    cp myqnx.ifs /fs/hd1-qnx6/.boot and the ifs goes from hd0 to hd1.

regards,

Mark

_______________________________________________

General

http://community.qnx.com/sf/go/post95103

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post95103: Re: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-23T21:35:37Z

Almost there. After performing the above, and rebooting I get a meessage on the screen "QNX v1.2b Boot Loader" and a blinking cursor. I think the issue is not being able to copy my ifs from one drive to another. This is basic but I'm have trouble. I believe the command should be

cp myqnx.ifs /fs/hd1-qnx6/.boot and the ifs goes from hd0 to hd1.

regards,

Mark

post95093: RE: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-23T19:15:33Z

So this is good, you have a bootable partition on your disk...ignore the fdisk /dev/hd1t77 wrt. fdisk.

So now you need to run "mkqnx6fs /dev/hd1t77" to create the qnx6 filesystem on your partition.

Then reboot/restart the io-blk driver, and put your disk partition should show up as /fs/hd1-qnx6 and put your ifs file in /fs/hd1-qnx6/.boot/

Then power off and remove your /dev/hd0 disk, upon reboot the loader should show you the ifs that is in your /.boot/

-----Original Message-----

From: Mark Bangert [mailto:community-noreply@qnx.com]

Sent: August-23-12 2:42 PM

To: general-filesystems

Subject: Re: RE: Booting a Minimal QNX6.5 from a HardDrive

If you do an fdisk /dev/hd1 show you get:

        _____OS_____     Start      End     ______Number_____    Size    Boot

        name    type    Cylinder  Cylinder  Cylinders  Blocks

    1.  QNX6   (177)          0     19456     19457   312576642  152625 MB  *

    2.  ______ (___)    _______   _______   _______   _________  _____

    3.  ______ (___)    _______   _______   _______   _________  _____

    4.  ______ (___)    _______   _______   _______   _________  _____

If you do fdisk  /dev/hd1t177 show you get:

        _____OS_____     Start      End     ______Number_____    Size    Boot

        name    type    Cylinder  Cylinder  Cylinders  Blocks

    1.  nonQNX (114)     242288   255782     13495   216793029  105855 MB  *

    2.  __dos_ (_6_)    204664   205699      1036    16641012   8125 MB  *

    3.  ______ (___)    _______   _______   _______   _________  _____

    4.  ______ (___)    _______   _______   _______   _________  _____

_______________________________________________

General

http://community.qnx.com/sf/go/post95089

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post95089: Re: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-23T18:41:39Z

If you do an fdisk /dev/hd1 show you get:

        _____OS_____     Start      End     ______Number_____    Size    Boot

        name    type    Cylinder  Cylinder  Cylinders  Blocks

    1.  QNX6   (177)          0     19456     19457   312576642  152625 MB  *

    2.  ______ (___)    _______   _______   _______   _________  _____

    3.  ______ (___)    _______   _______   _______   _________  _____

    4.  ______ (___)    _______   _______   _______   _________  _____

If you do fdisk  /dev/hd1t177 show you get:

        _____OS_____     Start      End     ______Number_____    Size    Boot

        name    type    Cylinder  Cylinder  Cylinders  Blocks

    1.  nonQNX (114)     242288   255782     13495   216793029  105855 MB  *

    2.  __dos_ (_6_)    204664   205699      1036    16641012   8125 MB  *

    3.  ______ (___)    _______   _______   _______   _________  _____

    4.  ______ (___)    _______   _______   _______   _________  _____

post95087: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-23T18:26:34Z

"show" just prints the disk partition table...without it you can manipulate it.

/dev/hd1 is the raw partition, whereas /dev/hd1t77 is the QNX6 partition(don't do anything with this one)

What does the output of "fdisk /dev/hd1 show" say?

-----Original Message-----

From: Mark Bangert [mailto:community-noreply@qnx.com]

Sent: August-23-12 1:13 PM

To: general-filesystems

Subject: Re: Booting a Minimal QNX6.5 from a HardDrive

what is the difference between these two commands; fdisk /dev/hd1 show and fdisk /dev/hd1t177

I get two very different results, which one is correct?

_______________________________________________

General

http://community.qnx.com/sf/go/post95083

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post95083: Re: Booting a Minimal QNX6.5 from a HardDrive

2012-08-23T17:12:45Z

what is the difference between these two commands;

fdisk /dev/hd1 show

and

fdisk /dev/hd1t177

I get two very different results, which one is correct?

post95069: Re: Booting a Minimal QNX6.5 from a HardDrive

2012-08-23T13:12:17Z

It appears you did...

dinit -h /dev/hd0

or

mkqnx6fs /dev/hd0

These would destroy the partition table as your post indicates.

Use fdisk first and create the type of partition you want ... 77 or 177 (qnx6fs)

Then specify the partition device, i.e. /dev/hd0t177 rather than /dev/hd0.

post95068: Re: Booting a Minimal QNX6.5 from a HardDrive

2012-08-23T12:58:26Z

Thank for the response.

No there is no message of any kind.

My first problem is I think that I'm not get the disk formatted correctly because the partition table look very strange.

_____OS_____ Start End ______Number_____ Size Boot

name type Cylinder Cylinder Cylinders Blocks

1. nonQNX 205 87476 87507 32 506704 247 MB *

2. nonQNX 88 0 4294967295 0 0 0 MB *

3. ______ (___) _______ _______ _______ _________ _____

4. ______ (___) _______ _______ _______ _________ _____

post95067: Re: Booting a Minimal QNX6.5 from a HardDrive

2012-08-23T12:51:28Z

Is there an error message when it fails to boot?

post95059: RE: Booting a Minimal QNX6.5 from a HardDrive

2012-08-23T02:45:32Z

These resources should help you:

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_technotes/qnx6bootloader.html

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_utilities/d/dinit.html

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_utilities/f/fs-qnx6.so.html

-----Original Message-----

From: Mark Bangert [mailto:community-noreply@qnx.com]

Sent: August-22-12 5:43 PM

To: general-filesystems

Subject: Booting a Minimal QNX6.5 from a HardDrive

The project that I'm working on requires a minimal QNX 6.5 system witch runs on an x86 platform. Currently the .ifs image we are using can be booted from a Thumb drive and works great. I now have to move this image to a Hard Drive.

I can fdisk and mount the drive and I can copy files to it, but when I make the hard drive the main boot device it won"t boot.

I not sure what boot loader I'm suppose to be using and I'm also not sure if the drive is formated correctly. I'm sure someone has done this before, and I'm looking for some direction.

Regards,

Mark

_______________________________________________

General

http://community.qnx.com/sf/go/post95049

To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

post95049: Booting a Minimal QNX6.5 from a HardDrive

2012-08-22T21:42:49Z

The project that I'm working on requires a minimal QNX 6.5 system witch runs on an x86 platform. Currently the .ifs image we are using can be booted from a Thumb drive and works great. I now have to move this image to a Hard Drive.

I can fdisk and mount the drive and I can copy files to it, but when I make the hard drive the main boot device it won"t boot.

I not sure what boot loader I'm suppose to be using and I'm also not sure if the drive is formated correctly. I'm sure someone has done this before, and I'm looking for some direction.

Regards,

Mark

post94721: Re: SD card driver - mount problems

2012-08-09T16:06:18Z

Just look at an existing implementation. For example, I used the Atmel AT91SAM BSP. Then, I used the existing Linux code for my board to see how to implement the low-level code and glued all together. Wasn't fun, but got there eventually.

> Hello todd,

> I'm also porting SD card filesystem on QNX. I write low level driver

> for SDHC card and test read() write() etc. ok, and I want to kown how to hook

> on the sim system? Is there some document about that? Thank you.

post94706: Re: SD card driver - mount problems

2012-08-09T06:46:00Z

Hello todd,

I'm also porting SD card filesystem on QNX. I write low level driver for SDHC card and test read() write() etc. ok, and I want to kown how to hook on the sim system? Is there some document about that? Thank you.

post94544: Re: RE: RE: Upgrade QNX 6.2 bootable CF to bootable QNX 6.5 without formatting

2012-07-31T21:59:16Z

Is it possible to get a dloader source code ?

post94542: RE: RE: Upgrade QNX 6.2 bootable CF to bootable QNX 6.5 without formatting

2012-07-31T21:45:47Z

Then you will need a dloader program that has been re-compiled to run under 6.2.0. Those errors are because the libc.so.3 will not be used until after you build it into a new IFS, which will require putting a 6.5.0 bootimage onto the device. The LD_PRELOAD environment variable is an attempt to make the load still try the new libc.so.3, even though the 6.2.0 libc.so.3 is already loaded from /proc/boot.

> -----Original Message-----

> From: janusz ruszel [mailto:community-noreply@qnx.com]

> Sent: July-31-12 5:43 PM

> To: general-filesystems

> Subject: Re: RE: Upgrade QNX 6.2 bootable CF to bootable QNX 6.5 without

> formatting

>

> I did copy QNX 6.5 c library together with dloader from QNX 6.5. Yes, it works

> under QNX 6.5. The point is to upgrade the loader to 6.5 version when the 6.2

> is still running.

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post94541

> To cancel your subscription to this discussion, please e-mail general-

> filesystems-unsubscribe@community.qnx.com

post94540: Re: QNX errors from hard disk driver.

2012-07-31T21:43:56Z

I guess that board supports Intel Advanced Programmable Interrupt Controllers (APIC). Did you start startup-apic program ? If so, we have had similar issue with another brand name board as well.

post94541: Re: RE: Upgrade QNX 6.2 bootable CF to bootable QNX 6.5 without formatting

2012-07-31T21:42:47Z

I did copy QNX 6.5 c library together with dloader from QNX 6.5. Yes, it works under QNX 6.5. The point is to upgrade the loader to 6.5 version when the 6.2 is still running.

post94539: RE: Upgrade QNX 6.2 bootable CF to bootable QNX 6.5 without formatting

2012-07-31T21:40:00Z

Janusz,

Those symbols require the new 6.5.0 libc.so.3, so you will need to make a new QNX boot image with 6.5.0 libraries before you can run the utility. You might have success if you set LD_PRELOAD=<path of 6.5 libc.so.3>.

David

> -----Original Message-----

> From: janusz ruszel [mailto:community-noreply@qnx.com]

> Sent: July-31-12 5:36 PM

> To: general-filesystems

> Subject: Upgrade QNX 6.2 bootable CF to bootable QNX 6.5 without

> formatting

>

> How to update boot loader on the CF card formatted and bootable under

> QNX 6.2 to QNX 6.5. I need to update boot loader only since 6.2 bootloader

> does not work with 6.5 .boot image. The flash card can not be re-formatted. I

> was attempting to run 6.5 dloader under QNX 6.2 with added libc.so.3 library

> and new pc2 file but I am getting not resolved symbols error :

>

> unknown symbol: __deregister_frame_info

> unknown symbol: _Jv_RegisterClasses

> unknown symbol: __register_frame_info

> unknown symbol: __deregister_frame_info

> unknown symbol: __register_frame_info

>

> Any helpful solution or suggestion are really appreciate.

> Janusz

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post94538

> To cancel your subscription to this discussion, please e-mail general-

> filesystems-unsubscribe@community.qnx.com

post94538: Upgrade QNX 6.2 bootable CF to bootable QNX 6.5 without formatting

2012-07-31T21:35:49Z

How to update boot loader on the CF card formatted and bootable under QNX 6.2 to QNX 6.5. I need to update boot loader only since 6.2 bootloader does not work with 6.5 .boot image. The flash card can not be re-formatted. I was attempting to run 6.5 dloader under QNX 6.2 with added libc.so.3 library and new pc2 file but I am getting not resolved symbols error :

unknown symbol: __deregister_frame_info

unknown symbol: _Jv_RegisterClasses

unknown symbol: __register_frame_info

unknown symbol: __deregister_frame_info

unknown symbol: __register_frame_info

Any helpful solution or suggestion are really appreciate.

Janusz

post93660: Bootloader for QNX4 file system

2012-06-14T12:46:59Z

I have to make a new QNX 6.2.1 bootable image with QNX4 fie system on a CF-card.

Can I initialize the CF-card (including bootloader) using QNX 6.3.2?

Using QNX 6.3.2 for this will simplyfy the process, using a USB - CF Card Reader preparing the CF-card.

I guess that my question boils down to:

Can I use the same bootloader for QNX 6.3.2 and QNX 6.2.1?

Mikael

post93619: etfs driver - testing

2012-06-12T17:26:21Z

I have ported Micron MT29F4G08 driver from other BSPs to MPC8313 device.

This seems to be working and I have "/dev/etfs1", "/dev/etfs2" "fs/etfs" etc

and have wrote a small appllication which created a file on fs/etfs, wrote

some data to this and read it back successfully.

I have however noticed in target file system navigator that I cannot browse

into folder fs and its size is showing as zero. It is possible to browse

folder fs from the console and also possible to copy folder fs from target

file system navigator to my desktop ending with the following structure:

fs/etfs/.badblks

fs/etfs/.counts

fs/etfs/.filetable

fs/etfs/.reserved

fs/etfs/<files_I_created>

Should I be able to access folder fs from target file system navigator, if

so any idea what could be going wrong for me?

Also any advice on the approach to use for validating my driver would be

appreciated?

Thanks.

post93461: Re: mount EFS image

2012-06-01T16:49:12Z

Hi Oleg,

As I can see, you should use flashctl utility for the mounting purpose.

For example:

>> flashctl -p *.efs -n /mnt/flash -m

post93453: Re: pathmgr_symlink() returns EPERM

2012-06-01T13:59:47Z

Thanks, David. I'll give this a try.

post93438: Re: pathmgr_symlink() returns EPERM

2012-06-01T01:41:15Z

pathmgr_symlink() is more like resmgr_attach() than symlink(). It requires that you have root privileges.

post93432: pathmgr_symlink() returns EPERM

2012-05-31T21:46:53Z

Hi there,

I am trying to use pathmgr_symlink() to create a symlink, but I am getting errno=EPERM (operation not permitted), while symlink() works. Can anyone please help explain why I am getting EPERM?

Regards,

post93420: Re: Mounting qnx4 fs with errors

2012-05-31T15:37:31Z

Ok, I finally found it out. I rebooted the whole system and was able to mount

everything. May be there was a lock put on a file somewhere.

Thanks,

David.

Le 31.05.2012 17:03, David Billard a écrit :

> Ok, I found the "mount -t qnx6". Now I have a mounting problem. I can mount the

> partitions:

> /dev/hd0t77.1

> /dev/hd0t77.2

> /dev/hd0t77.3

> /dev/hd0t77.4

> /dev/hd0t77.5

>

> But not the /dev/hd0t77 (which is the biggest partition but the numbering scheme

> suggests that this partition is subdivided in 1 to 5 sub partitions). Does it

> means that there is an unpartitionned (and unformatted) space on the drive?

>

> Many thanks,

>

> David.

> Le 31.05.2012 16:37, David Billard a écrit :

>> Hi David,

>>

>> Sorry to bother you with my problem. How can I mount a qnx6 file system (I'm

>> using the boot CD)?

>>

>> Many thanks,

>>

>> David.

>>

>> Le 15.05.2012 16:35, David Sarrazin a écrit :

>>> What OS version are you running?  Partition type 77 is normally reserved for partitions of type QNX4.   Fdisk should say "QNX4" , simply because the type is 77.  You can also try chkqnx6fs, on the off chance that the disks are formatted with the QNX6 file system.

>>>

>>> David

>>>

>>>> -----Original Message-----

>>>> From: David Billard [mailto:community-noreply@qnx.com]

>>>> Sent: May-15-12 10:33 AM

>>>> To: general-filesystems@community.qnx.com

>>>> Subject: Re: Mounting qnx4 fs with errors

>>>>

>>>> Hi David,

>>>>

>>>> it says that my partition is not a QNX4 fs. A rapid view with fdisk says

>>>> "unknown loader". Is there any way to detect which fs is on the partitions

>>>> (I'm almost sure it's qnx4, but perhaps am i wrong).

>>>>

>>>> many thanks,

>>>>

>>>> David.

>>>>

>>>> Le 15.05.2012 16:21, David Sarrazin a écrit :

>>>>> First thing to try is chkfsys, which is the QNX4 file system checker tool.  If

>>>> you don't want to make any changes to the disk, the -f option will prevent

>>>> the tool from modifying anything.

>>>>>

>>>>>> -----Original Message-----

>>>>>> From: David Billard [mailto:community-noreply@qnx.com]

>>>>>> Sent: May-15-12 10:35 AM

>>>>>> To: general-filesystems

>>>>>> Subject: Mounting qnx4 fs with errors

>>>>>>

>>>>>> Dear all,

>>>>>>

>>>>>> I'm new to QNX. I would like to recover files from a qnx4 filesystem

>>>>>> on a harddrive. The hardrive has 6 partitions. I used the BOOT-CD and

>>>>>> the photon file manager. If I do a df, it sees :

>>>>>>

>>>>>> /dev/hd0

>>>>>> /dev/hd0t77

>>>>>> /dev/hd0t77.1

>>>>>> /dev/hd0t77.2

>>>>>> /dev/hd0t77.3

>>>>>> /dev/hd0t77.4

>>>>>> /dev/hd0t77.5

>>>>>>

>>>>>> If I try to mount /dev/hd0t77 or any of its declination, it says that

>>>>>> the fs is corrupted.

>>>>>>

>>>>>> To my knowledge, the disk is not damaged. Do you see any reason for

>>>>>> this message or any solution for me?

>>>>>>

>>>>>> Is there a way to fix the fs?

>>>>>>

>>>>>> Many thanks,

>>>>>>

>>>>>> david.

>>>>>>

>>>>>>

>>>>>>

>>>>>>

>>>>>> _______________________________________________

>>>>>>

>>>>>> General

>>>>>> http://community.qnx.com/sf/go/post93087

>>>>>> To cancel your subscription to this discussion, please e-mail

>>>>>> general- filesystems-unsubscribe@community.qnx.com

>>>>>

>>>>>

>>>>>

>>>>>

>>>>> _______________________________________________

>>>>>

>>>>> General

>>>>> http://community.qnx.com/sf/go/post93088

>>>>> To cancel your subscription to this discussion, please e-mail

>>>>> general-filesystems-unsubscribe@community.qnx.com

>>>>

>>>> --

>>>> Prof. David Billard                         | David.Billard@hesge.ch

>>>> University of Applied Sciences in Geneva    | tel: +41 22 388 1885

>>>> Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

>>>>

>>>>

>>>>

>>>>

>>>>

>>>> _______________________________________________

>>>>

>>>> General

>>>> http://community.qnx.com/sf/go/post93090

>>>> To cancel your subscription to this discussion, please e-mail general-

>>>> filesystems-unsubscribe@community.qnx.com

>>>

>>>

>>>

>>>

>>> _______________________________________________

>>>

>>> General

>>> http://community.qnx.com/sf/go/post93091

>>> To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

>>

>

--

Prof. David Billard                         | David.Billard@hesge.ch

University of Applied Sciences in Geneva    | tel: +41 22 388 1885

Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

post93414: Re: Mounting qnx4 fs with errors

2012-05-31T15:06:41Z

Ok, I found the "mount -t qnx6". Now I have a mounting problem. I can mount the

partitions:

/dev/hd0t77.1

/dev/hd0t77.2

/dev/hd0t77.3

/dev/hd0t77.4

/dev/hd0t77.5

But not the /dev/hd0t77 (which is the biggest partition but the numbering scheme

suggests that this partition is subdivided in 1 to 5 sub partitions). Does it

means that there is an unpartitionned (and unformatted) space on the drive?

Many thanks,

David.

Le 31.05.2012 16:37, David Billard a écrit :

> Hi David,

>

> Sorry to bother you with my problem. How can I mount a qnx6 file system (I'm

> using the boot CD)?

>

> Many thanks,

>

> David.

>

> Le 15.05.2012 16:35, David Sarrazin a écrit :

>> What OS version are you running?  Partition type 77 is normally reserved for partitions of type QNX4.   Fdisk should say "QNX4" , simply because the type is 77.  You can also try chkqnx6fs, on the off chance that the disks are formatted with the QNX6 file system.

>>

>> David

>>

>>> -----Original Message-----

>>> From: David Billard [mailto:community-noreply@qnx.com]

>>> Sent: May-15-12 10:33 AM

>>> To: general-filesystems@community.qnx.com

>>> Subject: Re: Mounting qnx4 fs with errors

>>>

>>> Hi David,

>>>

>>> it says that my partition is not a QNX4 fs. A rapid view with fdisk says

>>> "unknown loader". Is there any way to detect which fs is on the partitions

>>> (I'm almost sure it's qnx4, but perhaps am i wrong).

>>>

>>> many thanks,

>>>

>>> David.

>>>

>>> Le 15.05.2012 16:21, David Sarrazin a écrit :

>>>> First thing to try is chkfsys, which is the QNX4 file system checker tool.  If

>>> you don't want to make any changes to the disk, the -f option will prevent

>>> the tool from modifying anything.

>>>>

>>>>> -----Original Message-----

>>>>> From: David Billard [mailto:community-noreply@qnx.com]

>>>>> Sent: May-15-12 10:35 AM

>>>>> To: general-filesystems

>>>>> Subject: Mounting qnx4 fs with errors

>>>>>

>>>>> Dear all,

>>>>>

>>>>> I'm new to QNX. I would like to recover files from a qnx4 filesystem

>>>>> on a harddrive. The hardrive has 6 partitions. I used the BOOT-CD and

>>>>> the photon file manager. If I do a df, it sees :

>>>>>

>>>>> /dev/hd0

>>>>> /dev/hd0t77

>>>>> /dev/hd0t77.1

>>>>> /dev/hd0t77.2

>>>>> /dev/hd0t77.3

>>>>> /dev/hd0t77.4

>>>>> /dev/hd0t77.5

>>>>>

>>>>> If I try to mount /dev/hd0t77 or any of its declination, it says that

>>>>> the fs is corrupted.

>>>>>

>>>>> To my knowledge, the disk is not damaged. Do you see any reason for

>>>>> this message or any solution for me?

>>>>>

>>>>> Is there a way to fix the fs?

>>>>>

>>>>> Many thanks,

>>>>>

>>>>> david.

>>>>>

>>>>>

>>>>>

>>>>>

>>>>> _______________________________________________

>>>>>

>>>>> General

>>>>> http://community.qnx.com/sf/go/post93087

>>>>> To cancel your subscription to this discussion, please e-mail

>>>>> general- filesystems-unsubscribe@community.qnx.com

>>>>

>>>>

>>>>

>>>>

>>>> _______________________________________________

>>>>

>>>> General

>>>> http://community.qnx.com/sf/go/post93088

>>>> To cancel your subscription to this discussion, please e-mail

>>>> general-filesystems-unsubscribe@community.qnx.com

>>>

>>> --

>>> Prof. David Billard                         | David.Billard@hesge.ch

>>> University of Applied Sciences in Geneva    | tel: +41 22 388 1885

>>> Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

>>>

>>>

>>>

>>>

>>>

>>> _______________________________________________

>>>

>>> General

>>> http://community.qnx.com/sf/go/post93090

>>> To cancel your subscription to this discussion, please e-mail general-

>>> filesystems-unsubscribe@community.qnx.com

>>

>>

>>

>>

>> _______________________________________________

>>

>> General

>> http://community.qnx.com/sf/go/post93091

>> To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

>

--

Prof. David Billard                         | David.Billard@hesge.ch

University of Applied Sciences in Geneva    | tel: +41 22 388 1885

Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

post93407: Re: Mounting qnx4 fs with errors

2012-05-31T14:44:17Z

Hi David,

Sorry to bother you with my problem. How can I mount a qnx6 file system (I'm

using the boot CD)?

Many thanks,

David.

Le 15.05.2012 16:35, David Sarrazin a écrit :

> What OS version are you running?  Partition type 77 is normally reserved for partitions of type QNX4.   Fdisk should say "QNX4" , simply because the type is 77.  You can also try chkqnx6fs, on the off chance that the disks are formatted with the QNX6 file system.

>

> David

>

>> -----Original Message-----

>> From: David Billard [mailto:community-noreply@qnx.com]

>> Sent: May-15-12 10:33 AM

>> To: general-filesystems@community.qnx.com

>> Subject: Re: Mounting qnx4 fs with errors

>>

>> Hi David,

>>

>> it says that my partition is not a QNX4 fs. A rapid view with fdisk says

>> "unknown loader". Is there any way to detect which fs is on the partitions

>> (I'm almost sure it's qnx4, but perhaps am i wrong).

>>

>> many thanks,

>>

>> David.

>>

>> Le 15.05.2012 16:21, David Sarrazin a écrit :

>>> First thing to try is chkfsys, which is the QNX4 file system checker tool.  If

>> you don't want to make any changes to the disk, the -f option will prevent

>> the tool from modifying anything.

>>>

>>>> -----Original Message-----

>>>> From: David Billard [mailto:community-noreply@qnx.com]

>>>> Sent: May-15-12 10:35 AM

>>>> To: general-filesystems

>>>> Subject: Mounting qnx4 fs with errors

>>>>

>>>> Dear all,

>>>>

>>>> I'm new to QNX. I would like to recover files from a qnx4 filesystem

>>>> on a harddrive. The hardrive has 6 partitions. I used the BOOT-CD and

>>>> the photon file manager. If I do a df, it sees :

>>>>

>>>> /dev/hd0

>>>> /dev/hd0t77

>>>> /dev/hd0t77.1

>>>> /dev/hd0t77.2

>>>> /dev/hd0t77.3

>>>> /dev/hd0t77.4

>>>> /dev/hd0t77.5

>>>>

>>>> If I try to mount /dev/hd0t77 or any of its declination, it says that

>>>> the fs is corrupted.

>>>>

>>>> To my knowledge, the disk is not damaged. Do you see any reason for

>>>> this message or any solution for me?

>>>>

>>>> Is there a way to fix the fs?

>>>>

>>>> Many thanks,

>>>>

>>>> david.

>>>>

>>>>

>>>>

>>>>

>>>> _______________________________________________

>>>>

>>>> General

>>>> http://community.qnx.com/sf/go/post93087

>>>> To cancel your subscription to this discussion, please e-mail

>>>> general- filesystems-unsubscribe@community.qnx.com

>>>

>>>

>>>

>>>

>>> _______________________________________________

>>>

>>> General

>>> http://community.qnx.com/sf/go/post93088

>>> To cancel your subscription to this discussion, please e-mail

>>> general-filesystems-unsubscribe@community.qnx.com

>>

>> --

>> Prof. David Billard                         | David.Billard@hesge.ch

>> University of Applied Sciences in Geneva    | tel: +41 22 388 1885

>> Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

>>

>>

>>

>>

>>

>> _______________________________________________

>>

>> General

>> http://community.qnx.com/sf/go/post93090

>> To cancel your subscription to this discussion, please e-mail general-

>> filesystems-unsubscribe@community.qnx.com

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post93091

> To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

--

Prof. David Billard                         | David.Billard@hesge.ch

University of Applied Sciences in Geneva    | tel: +41 22 388 1885

Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

post93103: Re: Mounting qnx4 fs with errors

2012-05-15T15:59:17Z

I'm using the boot-cd "qnxsdp-6.4.1-x86-200905201802-nto.iso".

chkqnx6fs returns nothing. That is it does something but then the prompt appears

and nothing is displayed at the screen.

Does it means it is really qnx6? In that case, do I need a special driver for

qnx6 fs?

David.

Le 15.05.2012 16:35, David Sarrazin a écrit :

> What OS version are you running?  Partition type 77 is normally reserved for partitions of type QNX4.   Fdisk should say "QNX4" , simply because the type is 77.  You can also try chkqnx6fs, on the off chance that the disks are formatted with the QNX6 file system.

>

> David

>

>> -----Original Message-----

>> From: David Billard [mailto:community-noreply@qnx.com]

>> Sent: May-15-12 10:33 AM

>> To: general-filesystems@community.qnx.com

>> Subject: Re: Mounting qnx4 fs with errors

>>

>> Hi David,

>>

>> it says that my partition is not a QNX4 fs. A rapid view with fdisk says

>> "unknown loader". Is there any way to detect which fs is on the partitions

>> (I'm almost sure it's qnx4, but perhaps am i wrong).

>>

>> many thanks,

>>

>> David.

>>

>> Le 15.05.2012 16:21, David Sarrazin a écrit :

>>> First thing to try is chkfsys, which is the QNX4 file system checker tool.  If

>> you don't want to make any changes to the disk, the -f option will prevent

>> the tool from modifying anything.

>>>

>>>> -----Original Message-----

>>>> From: David Billard [mailto:community-noreply@qnx.com]

>>>> Sent: May-15-12 10:35 AM

>>>> To: general-filesystems

>>>> Subject: Mounting qnx4 fs with errors

>>>>

>>>> Dear all,

>>>>

>>>> I'm new to QNX. I would like to recover files from a qnx4 filesystem

>>>> on a harddrive. The hardrive has 6 partitions. I used the BOOT-CD and

>>>> the photon file manager. If I do a df, it sees :

>>>>

>>>> /dev/hd0

>>>> /dev/hd0t77

>>>> /dev/hd0t77.1

>>>> /dev/hd0t77.2

>>>> /dev/hd0t77.3

>>>> /dev/hd0t77.4

>>>> /dev/hd0t77.5

>>>>

>>>> If I try to mount /dev/hd0t77 or any of its declination, it says that

>>>> the fs is corrupted.

>>>>

>>>> To my knowledge, the disk is not damaged. Do you see any reason for

>>>> this message or any solution for me?

>>>>

>>>> Is there a way to fix the fs?

>>>>

>>>> Many thanks,

>>>>

>>>> david.

>>>>

>>>>

>>>>

>>>>

>>>> _______________________________________________

>>>>

>>>> General

>>>> http://community.qnx.com/sf/go/post93087

>>>> To cancel your subscription to this discussion, please e-mail

>>>> general- filesystems-unsubscribe@community.qnx.com

>>>

>>>

>>>

>>>

>>> _______________________________________________

>>>

>>> General

>>> http://community.qnx.com/sf/go/post93088

>>> To cancel your subscription to this discussion, please e-mail

>>> general-filesystems-unsubscribe@community.qnx.com

>>

>> --

>> Prof. David Billard                         | David.Billard@hesge.ch

>> University of Applied Sciences in Geneva    | tel: +41 22 388 1885

>> Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

>>

>>

>>

>>

>>

>> _______________________________________________

>>

>> General

>> http://community.qnx.com/sf/go/post93090

>> To cancel your subscription to this discussion, please e-mail general-

>> filesystems-unsubscribe@community.qnx.com

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post93091

> To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

--

Prof. David Billard                         | David.Billard@hesge.ch

University of Applied Sciences in Geneva    | tel: +41 22 388 1885

Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

post93088: RE: Mounting qnx4 fs with errors

2012-05-15T14:40:55Z

First thing to try is chkfsys, which is the QNX4 file system checker tool. If you don't want to make any changes to the disk, the -f option will prevent the tool from modifying anything.

> -----Original Message-----

> From: David Billard [mailto:community-noreply@qnx.com]

> Sent: May-15-12 10:35 AM

> To: general-filesystems

> Subject: Mounting qnx4 fs with errors

>

> Dear all,

>

> I'm new to QNX. I would like to recover files from a qnx4 filesystem on a

> harddrive. The hardrive has 6 partitions. I used the BOOT-CD and the photon

> file manager. If I do a df, it sees :

>

> /dev/hd0

> /dev/hd0t77

> /dev/hd0t77.1

> /dev/hd0t77.2

> /dev/hd0t77.3

> /dev/hd0t77.4

> /dev/hd0t77.5

>

> If I try to mount /dev/hd0t77 or any of its declination, it says that the fs is

> corrupted.

>

> To my knowledge, the disk is not damaged. Do you see any reason for this

> message or any solution for me?

>

> Is there a way to fix the fs?

>

> Many thanks,

>

> david.

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post93087

> To cancel your subscription to this discussion, please e-mail general-

> filesystems-unsubscribe@community.qnx.com

post93091: RE: Mounting qnx4 fs with errors

2012-05-15T14:40:32Z

What OS version are you running?  Partition type 77 is normally reserved for partitions of type QNX4.   Fdisk should say "QNX4" , simply because the type is 77.  You can also try chkqnx6fs, on the off chance that the disks are formatted with the QNX6 file system.

David

> -----Original Message-----

> From: David Billard [mailto:community-noreply@qnx.com]

> Sent: May-15-12 10:33 AM

> To: general-filesystems@community.qnx.com

> Subject: Re: Mounting qnx4 fs with errors

>

> Hi David,

>

> it says that my partition is not a QNX4 fs. A rapid view with fdisk says

> "unknown loader". Is there any way to detect which fs is on the partitions

> (I'm almost sure it's qnx4, but perhaps am i wrong).

>

> many thanks,

>

> David.

>

> Le 15.05.2012 16:21, David Sarrazin a écrit :

> > First thing to try is chkfsys, which is the QNX4 file system checker tool.  If

> you don't want to make any changes to the disk, the -f option will prevent

> the tool from modifying anything.

> >

> >> -----Original Message-----

> >> From: David Billard [mailto:community-noreply@qnx.com]

> >> Sent: May-15-12 10:35 AM

> >> To: general-filesystems

> >> Subject: Mounting qnx4 fs with errors

> >>

> >> Dear all,

> >>

> >> I'm new to QNX. I would like to recover files from a qnx4 filesystem

> >> on a harddrive. The hardrive has 6 partitions. I used the BOOT-CD and

> >> the photon file manager. If I do a df, it sees :

> >>

> >> /dev/hd0

> >> /dev/hd0t77

> >> /dev/hd0t77.1

> >> /dev/hd0t77.2

> >> /dev/hd0t77.3

> >> /dev/hd0t77.4

> >> /dev/hd0t77.5

> >>

> >> If I try to mount /dev/hd0t77 or any of its declination, it says that

> >> the fs is corrupted.

> >>

> >> To my knowledge, the disk is not damaged. Do you see any reason for

> >> this message or any solution for me?

> >>

> >> Is there a way to fix the fs?

> >>

> >> Many thanks,

> >>

> >> david.

> >>

> >>

> >>

> >>

> >> _______________________________________________

> >>

> >> General

> >> http://community.qnx.com/sf/go/post93087

> >> To cancel your subscription to this discussion, please e-mail

> >> general- filesystems-unsubscribe@community.qnx.com

> >

> >

> >

> >

> > _______________________________________________

> >

> > General

> > http://community.qnx.com/sf/go/post93088

> > To cancel your subscription to this discussion, please e-mail

> > general-filesystems-unsubscribe@community.qnx.com

>

> --

> Prof. David Billard                         | David.Billard@hesge.ch

> University of Applied Sciences in Geneva    | tel: +41 22 388 1885

> Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

>

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post93090

> To cancel your subscription to this discussion, please e-mail general-

> filesystems-unsubscribe@community.qnx.com

post93090: Re: Mounting qnx4 fs with errors

2012-05-15T14:35:17Z

Hi David,

it says that my partition is not a QNX4 fs. A rapid view with fdisk says

"unknown loader". Is there any way to detect which fs is on the partitions (I'm

almost sure it's qnx4, but perhaps am i wrong).

many thanks,

David.

Le 15.05.2012 16:21, David Sarrazin a écrit :

> First thing to try is chkfsys, which is the QNX4 file system checker tool.  If you don't want to make any changes to the disk, the -f option will prevent the tool from modifying anything.

>

>> -----Original Message-----

>> From: David Billard [mailto:community-noreply@qnx.com]

>> Sent: May-15-12 10:35 AM

>> To: general-filesystems

>> Subject: Mounting qnx4 fs with errors

>>

>> Dear all,

>>

>> I'm new to QNX. I would like to recover files from a qnx4 filesystem on a

>> harddrive. The hardrive has 6 partitions. I used the BOOT-CD and the photon

>> file manager. If I do a df, it sees :

>>

>> /dev/hd0

>> /dev/hd0t77

>> /dev/hd0t77.1

>> /dev/hd0t77.2

>> /dev/hd0t77.3

>> /dev/hd0t77.4

>> /dev/hd0t77.5

>>

>> If I try to mount /dev/hd0t77 or any of its declination, it says that the fs is

>> corrupted.

>>

>> To my knowledge, the disk is not damaged. Do you see any reason for this

>> message or any solution for me?

>>

>> Is there a way to fix the fs?

>>

>> Many thanks,

>>

>> david.

>>

>>

>>

>>

>> _______________________________________________

>>

>> General

>> http://community.qnx.com/sf/go/post93087

>> To cancel your subscription to this discussion, please e-mail general-

>> filesystems-unsubscribe@community.qnx.com

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post93088

> To cancel your subscription to this discussion, please e-mail general-filesystems-unsubscribe@community.qnx.com

--

Prof. David Billard                         | David.Billard@hesge.ch

University of Applied Sciences in Geneva    | tel: +41 22 388 1885

Bât. F, 7 route de Drize - CH-1227 Carouge  | fax: +41 22 388 1701

post93087: Mounting qnx4 fs with errors

2012-05-15T14:34:55Z

Dear all,

I'm new to QNX. I would like to recover files from a qnx4 filesystem on a harddrive. The hardrive has 6 partitions. I used the BOOT-CD and the photon file manager. If I do a df, it sees :

/dev/hd0

/dev/hd0t77

/dev/hd0t77.1

/dev/hd0t77.2

/dev/hd0t77.3

/dev/hd0t77.4

/dev/hd0t77.5

If I try to mount /dev/hd0t77 or any of its declination, it says that the fs is corrupted.

To my knowledge, the disk is not damaged. Do you see any reason for this message or any solution for me?

Is there a way to fix the fs?

Many thanks,

david.

post92854: How to get smart information of disks behind RAID controller programmatically using QNX6.5.0

2012-05-01T02:40:08Z

Smart information for disks behind RAID controller using QNX6.5.0

Hi all,

Anybody familiar with getting smart information using QNX6.5.0. I am a newbie in this area. I want to get SMART

information from disks behind a RAID controller. I have seen a project named smartmontools in sourceforge.net. But it is

not supported for QNX in the case where the disks are behind RAID controller.

Anyone please help me.

Thanks a million,

Honey S

post92853: How to know the sector size of a disk programmatically

2012-05-01T02:36:43Z

Hi all,

I am a newbie in QNX. I am using QNX 6.5.0. I need to write a C++ program to find the sector size of a disk. I couldn't find any command/function for that. Please help me with a solution.

Thanks in advance,

Honey S

post92674: Query regarding reset cause and its equivalent mapping.

2012-04-20T15:10:06Z

Hi folks,

There was some reset observed on the qnx device and it seems to be hardware reset.

So what I see from system_info.txt file is that

First case:            Reset Cause =  0x00000500

Second case:      Reset Cause =  0x00000508

May I know what does this reset cause mean and which component is causing the reset.

Where do I find the mapping for all reset causes.

Thanks n Regards,

--------------------------

Satish Kumar Thakolam

Software Defect Analyst - Radio Product Software - Germany - R&D

Research In Motion Deutschland GmbH

Office:    +49 234 43 66 24 20

[cid:image001.gif@01CD1F18.61195A20]

post92621: RE: RE: C API similar to waitfor

2012-04-18T19:14:38Z

No, I haven't written anything about waitfor() yet.

Steve Reid (stever@qnx.com)

Technical Editor

QNX Software Systems

-----Original Message-----

From: Rohit Nair [mailto:community-noreply@qnx.com]

Sent: Wednesday, April 18, 2012 2:02 PM

To: general-filesystems

Subject: Re: RE: C API similar to waitfor

>PR 77631 is a laundry list of undocumented functions. You'll be happy to learn that waitfor() is on the list.

Is this a document that is available on this website ? Can you link to it ?

_______________________________________________

General

http://community.qnx.com/sf/go/post92619

post92619: Re: RE: C API similar to waitfor

2012-04-18T18:02:05Z

>PR 77631 is a laundry list of undocumented functions. You'll be happy to learn that waitfor() is on the list.

Is this a document that is available on this website ? Can you link to it ?

post92610: RE: C API similar to waitfor

2012-04-18T15:02:12Z

PR 77631 is a laundry list of undocumented functions. You'll be happy to learn that waitfor() is on the list.

Steve Reid (stever@qnx.com)

Technical Editor

QNX Software Systems

-----Original Message-----

From: Chris Travis [mailto:community-noreply@qnx.com]

Sent: Wednesday, April 18, 2012 10:24 AM

To: general-filesystems

Subject: RE: C API similar to waitfor

Sure, but according to our documentation and the "actual?" lib/c/qnx/waitfor.c it uses stat() :-) http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_utilities/w/waitfor.html

plus aside from the definition in libgen.c, waitfor() is not documented. :-(

post92609: Re: C API similar to waitfor

2012-04-18T14:36:02Z

If waitfor() doesn't use poll, it is preferred.

Which library includes waitfor()?

Thanks for quick responses.

-Chun

post92608: Re: C API similar to waitfor

2012-04-18T14:33:06Z

waitfor() should be used so that when / if it's

changed to not poll in a loop, everyone will

benefit.

-seanb

On Wed, Apr 18, 2012 at 10:23:49AM -0400, Chris Travis wrote:

> Sure, but according to our documentation and the "actual?" lib/c/qnx/waitfor.c it uses stat() :-)

> http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_utilities/w/waitfor.html

> plus aside from the definition in libgen.c, waitfor() is not documented. :-(

>

> -----Original Message-----

> From: Colin Burgess [mailto:community-noreply@qnx.com]

> Sent: April-18-12 10:11 AM

> To: general-filesystems

> Subject: Re: C API similar to waitfor

>

> Actually it uses waitfor() - it's in libgen.h

>

> :-)

>

> On 12-04-18 10:05 AM, Chris Travis wrote:

> > waitfor uses stat()

> >

> > Here is an example:

> >

> > int waitfor( char *path, unsigned wait_msec, unsigned wait_period_msec

> > ) {

> > struct stat sbuf;

> > unsigned total = 0, ret;

> >

> > do {

> > delay( wait_period_msec );

> > total += wait_period_msec;

> > } while ( (ret = stat( path,&sbuf )) == -1&& total< wait_msec );

> > if ( total>= wait_msec )

> > errno = ETIMEDOUT;

> > return ret;

> > }

> >

> > Chris

> >

> >

> > -----Original Message-----

> > From: Chun He [mailto:community-noreply@qnx.com]

> > Sent: April-18-12 2:39 AM

> > To: general-filesystems

> > Subject: C API similar to waitfor

> >

> > need to wait till a path (or a device node) to exists or appear in C program.

> > Is there a C API equal to waitfor command?

> >

> > best regards,

> > Chun

> >

> >

> >

> > _______________________________________________

> >

> > General

> > http://community.qnx.com/sf/go/post92596

> >

> >

> >

> >

> > _______________________________________________

> >

> > General

> > http://community.qnx.com/sf/go/post92605

> >

>

> --

> cburgess@qnx.com

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post92606

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post92607

>

post92607: RE: C API similar to waitfor

2012-04-18T14:23:48Z

Sure, but according to our documentation and the "actual?" lib/c/qnx/waitfor.c it uses stat() :-)

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_utilities/w/waitfor.html

plus aside from the definition in libgen.c, waitfor() is not documented. :-(

-----Original Message-----

From: Colin Burgess [mailto:community-noreply@qnx.com]

Sent: April-18-12 10:11 AM

To: general-filesystems

Subject: Re: C API similar to waitfor

Actually it uses waitfor() - it's in libgen.h

:-)

On 12-04-18 10:05 AM, Chris Travis wrote:

> waitfor uses stat()

>

> Here is an example:

>

> int waitfor( char *path, unsigned wait_msec, unsigned wait_period_msec

> ) {

> struct stat sbuf;

> unsigned total = 0, ret;

>

> do {

> delay( wait_period_msec );

> total += wait_period_msec;

> } while ( (ret = stat( path,&sbuf )) == -1&& total< wait_msec );

> if ( total>= wait_msec )

> errno = ETIMEDOUT;

> return ret;

> }

>

> Chris

>

>

> -----Original Message-----

> From: Chun He [mailto:community-noreply@qnx.com]

> Sent: April-18-12 2:39 AM

> To: general-filesystems

> Subject: C API similar to waitfor

>

> need to wait till a path (or a device node) to exists or appear in C program.

> Is there a C API equal to waitfor command?

>

> best regards,

> Chun

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post92596

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post92605

>

--

cburgess@qnx.com

_______________________________________________

General

http://community.qnx.com/sf/go/post92606

post92606: Re: C API similar to waitfor

2012-04-18T14:10:31Z

Actually it uses waitfor() - it's in libgen.h

:-)

On 12-04-18 10:05 AM, Chris Travis wrote:

> waitfor uses stat()

>

> Here is an example:

>

> int waitfor( char *path, unsigned wait_msec, unsigned wait_period_msec )

> {

> struct stat sbuf;

> unsigned total = 0, ret;

>

> do {

> delay( wait_period_msec );

> total += wait_period_msec;

> } while ( (ret = stat( path,&sbuf )) == -1&& total< wait_msec );

> if ( total>= wait_msec )

> errno = ETIMEDOUT;

> return ret;

> }

>

> Chris

>

>

> -----Original Message-----

> From: Chun He [mailto:community-noreply@qnx.com]

> Sent: April-18-12 2:39 AM

> To: general-filesystems

> Subject: C API similar to waitfor

>

> need to wait till a path (or a device node) to exists or appear in C program.

> Is there a C API equal to waitfor command?

>

> best regards,

> Chun

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post92596

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post92605

>

--

cburgess@qnx.com

post92605: RE: C API similar to waitfor

2012-04-18T14:04:59Z

waitfor uses stat()

Here is an example:

int waitfor( char *path, unsigned wait_msec, unsigned wait_period_msec )

{

struct stat sbuf;

unsigned total = 0, ret;

do {

delay( wait_period_msec );

total += wait_period_msec;

} while ( (ret = stat( path, &sbuf )) == -1 && total < wait_msec );

if ( total >= wait_msec )

errno = ETIMEDOUT;

return ret;

}

Chris

-----Original Message-----

From: Chun He [mailto:community-noreply@qnx.com]

Sent: April-18-12 2:39 AM

To: general-filesystems

Subject: C API similar to waitfor

need to wait till a path (or a device node) to exists or appear in C program.

Is there a C API equal to waitfor command?

best regards,

Chun

_______________________________________________

General

http://community.qnx.com/sf/go/post92596

post92596: C API similar to waitfor

2012-04-18T06:39:22Z

need to wait till a path (or a device node) to exists or appear in C program.

Is there a C API equal to waitfor command?

best regards,

Chun

post92341: Re: QNX 6 power fail-safe file system and compact flash cards

2012-03-29T04:32:54Z

Dennis, where are you getting this from? The qnx6 file system is particlarily good for removable media, owing to its snapshot traansaction behavior.

qnx4 file system should NOT be used if you have any requirement of file system consistency in the presence of media removes or sudden power off.

----- Original Message -----

From: Dennis Kellly [mailto:community-noreply@qnx.com]

Sent: Wednesday, March 28, 2012 10:31 AM
To: general-filesystems <post92323@community.qnx.com>

Subject: Re: QNX 6 power fail-safe file system and compact flash cards

You may find that qnx4fs is more reliable than qnx6fs with CF media. This suggestion is a result of a customer beta field test. Their failure rate was greatly reduced when they switched to qnx4fs... but you need to do your own testing!

_______________________________________________

General

http://community.qnx.com/sf/go/post92323

post92339: Re: QNX 6 power fail-safe file system and compact flash cards

2012-03-29T03:20:23Z

>>> I'm also looking at fsync() and sync() and their notes about committing data.

These should be effective immediately following fwrite() or write(). fwrite() can be more "at risk" than write() since there is an intermediate buffer. Use fdatasync() to force a flush of the intermediate buffer.

You can also execute utility "sync" to force the disk io manager to flush all inodes, etc. But use caution as this is a relatively long operation, possibly impacting system timing.

post92324: Re: QNX 6 power fail-safe file system and compact flash cards

2012-03-28T15:05:09Z

Great. We moved from the QNX 4 file system because it had the annoying habit of creating partial files on power fail. The consensus from QNX support and this forum was to use the QNX 6 file system. So for the benefit of anyone searching the forums for solutions to similar problems:

DON'T USE THE QNX 6 FILESYSTEM UNLESS YOU HAVE ROTATING HARD DISKS.

So, is there any guidance on making the QNX 4 filesystem more tolerant to power failure? Before I changed over, I was running chkfsys on powerup to clean broken files but this took time and didn't catch all the partial files. I'm also looking at fsync() and sync() and their notes about committing data.

Any other suggestions?

post92323: Re: QNX 6 power fail-safe file system and compact flash cards

2012-03-28T14:31:58Z

You may find that qnx4fs is more reliable than qnx6fs with CF media. This suggestion is a result of a customer beta field test. Their failure rate was greatly reduced when they switched to qnx4fs... but you need to do your own testing!

post92319: QNX 6 power fail-safe file system and compact flash cards

2012-03-28T12:30:18Z

We are using the QNX 6 file system on a compact flash card. The card we are using does not meet the criteria for power fail safe recovery so we are using the "sync=optional" to get it to mount read/write. We had an unfortunate incident where a unit was powered off while writing a data file, and the boot image became truncated.

We need to make the system power fail safe. We could tolerate a corrupted data file but not the boot image. This kind of failure can't be fixed in the field. I've tried several compact flash cards to find one that supports the file system, but this is a trial-and-error procedure.

1. Does anyone know of ANY compact flash card that fully supports the QNX 6 Power Fail Safe file system?

2. When will the QNX 6 file system be designed to support non-rotating media?

post92154: Resource manager (toggle it's existence in /dev)

2012-03-16T19:48:57Z

Is there a way to remove a named entry from /dev without actually detaching the resource manager?

I would like to, depending on some trigger, make it appear like the resource manager doesn't exist. Unfortunately, the resource manager will lose its root privilege after attaching so I can't do a detach/attach all the time.

Thanks,

Steve

post92147: Query :Regarding QNX

2012-03-16T10:14:45Z

Sir,

Thanks for the earlier reply regarding bootable CD. Now things are

done, but I am still stuck up with the problem.

I am able to connect host and Target system, proof is in QNX target

navigator I see QNX local host connected, I see a number of process, I am

attaching a screen shot of it.

I am not able to proceed further, if I want to access device on PCI bus how

to proceed.

Regards,

Amit negi

post92146: shell * fails on my resmgr directory

2012-03-16T07:00:02Z

I have a resmgr at attached to /dev/bucase QNX 6.4.1 (x86)

it has the next structure

# ls /dev/bucase/*

/dev/bucase/file1

/dev/bucase/test:

f     ff

So when I use next sequence of calls after start of resmgr

# echo /dev/bucase/*

/dev/bucase/file1 /dev/bucase/test

# echo /dev/bucase/test/*

echo /dev/bucase/test/*

/dev/bucase/test/f

So second call to echo fails, cause I have two entries in test directory

But ls /dev/bucase/test always work properly

# ls /dev/bucase/test

f     ff

Attaching minimum code with bug

Problem doesn't occur when I comment string that frees reply buffer to MsgReply call.

int handle_readdir(resmgr_context_t *ctp, io_read_t *msg, iofunc_ocb_t* ocb, struct myentry* entry){

char* reply = (char*)malloc(msg->i.nbytes);

        int rsize;

       //Cycle at directory entries

       rsize = (char*)de - reply;

if (-1 == MsgReply(ctp->rcvid, rsize, realrep, rsize)){

perror("!!!!!readdir MsgReply");

}

free(reply); //A comment of this string helps to avoid the problem but I cant clarify it (reply must be copied at the moment of its freeing, isnt it?)

return _RESMGR_NOREPLY;

}

post91966: RE: creating bootable cd of qnxsdp-6.5.0-201007091524-dvd

2012-03-08T04:15:18Z

Hi ,

Thanks for the reply, things are done some issue with my pc

Regards,

Amit negi

-----Original Message-----

From: Dennis Kellly [mailto:community-noreply@qnx.com]

Sent: Wednesday, March 07, 2012 5:59 PM

To: general-filesystems

Subject: Re: creating bootable cd of qnxsdp-6.5.0-201007091524-dvd

Download the CD ISO ... rather than the DVD image?

http://www.qnx.com/download/feature.html?programid=21182

_______________________________________________

General

http://community.qnx.com/sf/go/post91936

post91936: Re: creating bootable cd of qnxsdp-6.5.0-201007091524-dvd

2012-03-07T12:28:34Z

Download the CD ISO ... rather than the DVD image?

http://www.qnx.com/download/feature.html?programid=21182

post91929: creating bootable cd of qnxsdp-6.5.0-201007091524-dvd

2012-03-07T06:25:27Z

please can anyone tell me the procedure to make a boot-able CD for QNX target

post91803: Re: Mount of qnx6 file system

2012-02-27T12:21:21Z

Try adding "qnx6 sync=optional" to the end of your devb-xxx line. Apparently, your CF card responds truthfully that is does not support the "flush to disk" command required for qnx6fs.

post91802: Mount of qnx6 file system

2012-02-27T12:05:13Z

I installed QNX 6.5.0 on a CF with the QNX6 file system 179.

I'm only able to mount this CF as "read only" in a QNX 6.5.0 Virtual Machine on my PC.

Why is this? I would like to mount and copy some files to the image. I've done this in the past on QNX file system 79.

Is there any solution/trick to do this?

Mikael

post91629: Re: fs-qnx6: is full file system recovery possible?

2012-02-16T06:28:41Z

chkqnx6fs could correct some type of corruptions, but most case(like this one), it is not possible.

Could you provide the use info?

You can get it by

use -i `which -s fs-qnx6.so`

Thanks

yuxi

post91568: RE: RE: How to transfer a file from USB or Playbook?

2012-02-14T15:55:29Z

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_utilities/d/devb-umass.html

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_user_guide/fsystems.html

-----Original Message-----

From: HuiHsien Wu [mailto:community-noreply@qnx.com]

Sent: February-14-12 10:38 AM

To: general-filesystems

Subject: Re: RE: How to transfer a file from USB or Playbook?

For usb stick. Not USB line.

I just want to put a file to usb stick, and then plug-in the usb stick to QNX neutrino OS PC.

Finally, copy the file from usb stick to other folder.

Sorry, original description isn't clear.

Thanks your attention.

_______________________________________________

General

http://community.qnx.com/sf/go/post91566

post91566: Re: RE: How to transfer a file from USB or Playbook?

2012-02-14T15:37:43Z

For usb stick. Not USB line.

I just want to put a file to usb stick, and then plug-in the usb stick to QNX neutrino OS PC.

Finally, copy the file from usb stick to other folder.

Sorry, original description isn't clear.

Thanks your attention.

post91565: Re: RE: How to transfer a file from USB or Playbook?

2012-02-14T15:34:00Z

You can do it to playbook - but not to an ordinary QNX host. For playbook, you use the blackberry application "Blackberry Device Manager" and a usb cable.

Playbook file transfers are easier though using "explorer \\w.x.y.z" - i.e. wireless. :)

post91564: RE: How to transfer a file from USB or Playbook?

2012-02-14T15:28:03Z

> -----Message d'origine-----

> De : HuiHsien Wu [mailto:community-noreply@qnx.com]

> Envoyé : 14 février 2012 09:50

> À : general-filesystems

> Objet : How to transfer a file from USB or Playbook?

>

> Problem:

> I want to transfer a file from a Windows PC to a QNX neutrino PC by a USB.

> Could not find a way to do this action.

>

That is not supported, even by Windows I think. I don`t think you can have to "master" devices.

> Thank your support,

>

>

>

>

>

>

> _______________________________________________

>

> General

> http://community.qnx.com/sf/go/post91557

>

post91562: RE: Mount a zip file as a filesystem

2012-02-14T15:00:54Z

Rob Krten's book, The QNX Cookbook, includes code for a tar filesystem that you might find useful.

Steve Reid (stever@qnx.com)

Technical Editor

QNX Software Systems

-----Original Message-----

From: Andrew Sherk [mailto:community-noreply@qnx.com]

Sent: Tuesday, February 14, 2012 9:50 AM

To: general-filesystems

Subject: Re: Mount a zip file as a filesystem

The deflate utility and inflator resource manager are documented here:

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_utilities/d/deflate.html

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_utilities/i/inflator.html

-asherk

> You are describing a "virtual filesystem". QNX provides two: inflator (see "

> system architecture" guide) and pkgserver (used as an install tool).

>

> AFIK none exists for zip archives - though you could write one!

>

_______________________________________________

General

http://community.qnx.com/sf/go/post91558

post91558: Re: Mount a zip file as a filesystem

2012-02-14T14:50:13Z

The deflate utility and inflator resource manager are documented here:

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_utilities/d/deflate.html

http://www.qnx.com/developers/docs/6.5.0/index.jsp?topic=/com.qnx.doc.neutrino_utilities/i/inflator.html

-asherk

> You are describing a "virtual filesystem". QNX provides two: inflator (see "

> system architecture" guide) and pkgserver (used as an install tool).

>

> AFIK none exists for zip archives - though you could write one!

>

post91557: How to transfer a file from USB or Playbook?

2012-02-14T14:50:02Z

Problem:

I want to transfer a file from a Windows PC to a QNX neutrino PC by a USB.

Could not find a way to do this action.

Thank your support,

post91524: Re: Mount a zip file as a filesystem

2012-02-13T14:58:20Z

You are describing a "virtual filesystem". QNX provides two: inflator (see "system architecture" guide) and pkgserver (used as an install tool).

AFIK none exists for zip archives - though you could write one!

post91522: Mount a zip file as a filesystem

2012-02-13T14:42:05Z

Is there any way to mount a zip file as a read-only file system? Rather than copy and unzip, it would be really handy in my application to take a zip file on a USB drive and mount it as a file system. Unzipping would be done on-the-fly as files are read. Is there any way this can be done?