Project Home
Project Home		 Documents
Documents		 Wiki
Wiki		 Discussion Forums
Discussions		 Project Information
Project Info
QNX Operating System/Discussion/OSMeta/pidin map on 6.4.x/List of Posts
Forum Topic - pidin map on 6.4.x: (15 Items)
   
Update
Warren Deitch(deleted)
04/13/2009 4:25 AM
post26734
pidin map on 6.4.x   
"pidin map" will only show a single "Mapped Phys Memory".  This occurs  on x86 and PPC (600).  I have a number of 
drivers that map multiple regions - usually a control region and a data region but only a single region is reported.  
 A nand file system with control registers at 0xe0005000 (4Kb) and data at 0x10000000 (256Mb).  But with 6.4.0/1 only 
one region is shown

6.4.0 and 6.4.1

# pidin -Pfs-yaffs map
     pid tid name               prio STATE            code  data        stack
  131095   1 sbin/fs-yaffs       10r RECEIVE          272K  652K  8192(516K)*
             Mapped Phys Memory @80100000 (e0005000)        256M  S
#
# pidin -Pfs-yaffs me
     pid tid name               prio STATE            code  data        stack
  131095   1 sbin/fs-yaffs       10r RECEIVE          272K  652K  8192(516K)*
            /dev/mem           @80100000 (       0)        256M
#


6.3.2
# pidin -Pfs-yaffs me
     pid tid name               prio STATE           code  data         stack
    8203   1 proc/boot/fs-yaffs  10r RECEIVE          260K  656K  8192(516K)*
            mem                @80100000             4096  4096
            mem                @80102000             256M
#

btw just tried the latest sin and it has the same problem.  a single mem region only
Chris Travis
04/13/2009 8:09 AM
post26743
RE: pidin map on 6.4.x   
This looks like a known issue PR66114, but it is unclear if this will be
fixed in 6.4.1.

Chris

-----Original Message-----
From: Warren Deitch [mailto:community-noreply@qnx.com] 
Sent: Monday, April 13, 2009 4:25 AM
To: osmeta-core_os
Subject: pidin map on 6.4.x

"pidin map" will only show a single "Mapped Phys Memory".  This occurs
on x86 and PPC (600).  I have a number of drivers that map multiple
regions - usually a control region and a data region but only a single
region is reported.  
 A nand file system with control registers at 0xe0005000 (4Kb) and data
at 0x10000000 (256Mb).  But with 6.4.0/1 only one region is shown

6.4.0 and 6.4.1

# pidin -Pfs-yaffs map
     pid tid name               prio STATE            code  data
stack
  131095   1 sbin/fs-yaffs       10r RECEIVE          272K  652K
8192(516K)*
             Mapped Phys Memory @80100000 (e0005000)        256M  S
#
# pidin -Pfs-yaffs me
     pid tid name               prio STATE            code  data
stack
  131095   1 sbin/fs-yaffs       10r RECEIVE          272K  652K
8192(516K)*
            /dev/mem           @80100000 (       0)        256M
#


6.3.2
# pidin -Pfs-yaffs me
     pid tid name               prio STATE           code  data
stack
    8203   1 proc/boot/fs-yaffs  10r RECEIVE          260K  656K
8192(516K)*
            mem                @80100000             4096  4096
            mem                @80102000             256M
#

btw just tried the latest sin and it has the same problem.  a single mem
region only

_______________________________________________
OSMeta
http://community.qnx.com/sf/go/post26734
Colin Burgess(deleted)
04/13/2009 9:25 AM
post26754
Re: pidin map on 6.4.x   
Aleksander posted a diff for CI here - http://community.qnx.com/sf/go/post25371 - it just needs someone to review
it.

Colin

Chris Travis wrote:
> This looks like a known issue PR66114, but it is unclear if this will be
> fixed in 6.4.1.
> 
> Chris
> 
> -----Original Message-----
> From: Warren Deitch [mailto:community-noreply@qnx.com] 
> Sent: Monday, April 13, 2009 4:25 AM
> To: osmeta-core_os
> Subject: pidin map on 6.4.x
> 
> "pidin map" will only show a single "Mapped Phys Memory".  This occurs
> on x86 and PPC (600).  I have a number of drivers that map multiple
> regions - usually a control region and a data region but only a single
> region is reported.  
>  A nand file system with control registers at 0xe0005000 (4Kb) and data
> at 0x10000000 (256Mb).  But with 6.4.0/1 only one region is shown
> 
> 6.4.0 and 6.4.1
> 
> # pidin -Pfs-yaffs map
>      pid tid name               prio STATE            code  data
> stack
>   131095   1 sbin/fs-yaffs       10r RECEIVE          272K  652K
> 8192(516K)*
>              Mapped Phys Memory @80100000 (e0005000)        256M  S
> #
> # pidin -Pfs-yaffs me
>      pid tid name               prio STATE            code  data
> stack
>   131095   1 sbin/fs-yaffs       10r RECEIVE          272K  652K
> 8192(516K)*
>             /dev/mem           @80100000 (       0)        256M
> #
> 
> 
> 6.3.2
> # pidin -Pfs-yaffs me
>      pid tid name               prio STATE           code  data
> stack
>     8203   1 proc/boot/fs-yaffs  10r RECEIVE          260K  656K
> 8192(516K)*
>             mem                @80100000             4096  4096
>             mem                @80102000             256M
> #
> 
> btw just tried the latest sin and it has the same problem.  a single mem
> region only
> 
> _______________________________________________
> OSMeta
> http://community.qnx.com/sf/go/post26734
> 
> 
> _______________________________________________
> OSMeta
> http://community.qnx.com/sf/go/post26743
> 

-- 
cburgess@qnx.com
Warren Deitch(deleted)
04/14/2009 3:33 AM
post26856
Re: pidin map on 6.4.x   
Thank you Colin,

   The patch works for me on PPC.  Both pidin map and pidin me show all the physical memory details now.   I have not 
tried it on x86 yet.
Warren Deitch(deleted)
04/15/2009 8:45 AM
post27041
Re: pidin map on 6.4.x   
Sorry to say the patch has introduced other bugs.  pidin arg and pidin env no longer work.  Tried on 6.4.1 and 6.4.0. 
They both fault with 'couldn't read stack: Server fault on msg pass'
pidin arg displays only the pid then faults.
pidin env displays the pid and name then faults.

All the other pidin xxx worked ok.
Aleksandar Ristovski(deleted)
04/15/2009 9:15 AM
post27049
Re: pidin map on 6.4.x   
Warren Deitch wrote:
> Sorry to say the patch has introduced other bugs.  pidin arg and pidin env no longer work.  Tried on 6.4.1 and 6.4.0. 
They both fault with 'couldn't read stack: Server fault on msg pass'
> pidin arg displays only the pid then faults.
> pidin env displays the pid and name then faults.
> 
> All the other pidin xxx worked ok.
> 

Warren, thank you for your feedback.

The errors you saw were on a  ppc target, right? I was 
testing on x86 and all worked.


---
Aleksandar
Aleksandar Ristovski(deleted)
04/15/2009 9:17 AM
post27050
Re: pidin map on 6.4.x   
Aleksandar Ristovski wrote:
> Warren Deitch wrote:
>> Sorry to say the patch has introduced other bugs.  pidin arg and pidin 
>> env no longer work.  Tried on 6.4.1 and 6.4.0. They both fault with 
>> 'couldn't read stack: Server fault on msg pass'
>> pidin arg displays only the pid then faults.
>> pidin env displays the pid and name then faults.
>>
>> All the other pidin xxx worked ok.
>>
> 
> Warren, thank you for your feedback.
> 
> The errors you saw were on a  ppc target, right? I was testing on x86 
> and all worked.
  And now I tried on a ppc target, it worked for me.

Does it crash for you for *any* process, or some process in 
particular? I would like to reproduce it here.

Thanks,

Aleksandar
Warren Deitch(deleted)
04/15/2009 8:13 PM
post27160
Re: pidin map on 6.4.x   
"pidin arg"  and  "pidin -Pio-hid arg" both fail.    MPC 8349 (procnto-600-instr - last update svn ver 218052)
# pidin arg
     pid Arguments
       1 proc/boot/procnto-600-instr

couldn't read arg: Server fault on msg pass
       2

couldn't read arg: Server fault on msg pass
       3

couldn't read arg: Server fault on msg pass
       4

couldn't read arg: Server fault on msg pass
   16389

couldn't read arg: Server fault on msg pass
   16390

couldn't read arg: Server fault on msg pass
   20487

----------------------
More instances snipped
----------------------

couldn't read arg: Server fault on msg pass
  315423

couldn't read arg: Server fault on msg pass
  327712

couldn't read arg: Server fault on msg pass
  364570

couldn't read arg: Server fault on msg pass
  909323
# pidin -pio-hid arg
     pid Arguments

couldn't read arg: Server fault on msg pass
  155668
#
# pidin env
     pid name               Environment
       1 /procnto-600-instr

couldn't read stack: Server fault on msg pass
       2 proc/boot/slogger

couldn't read stack: Server fault on msg pass
       3 proc/boot/io-usb

couldn't read stack: Server fault on msg pass
       4 oc/boot/devb-umass

couldn't read stack: Server fault on msg pass
   16389 sbin/tinit

couldn't read stack: Server fault on msg pass
   16390 sbin/pipe

couldn't read stack: Server fault on msg pass
   20487 sbin/mqueue

couldn't read stack: Server fault on msg pass
   40969 usr/sbin/inetd

--------
snip
--------

couldn't read stack: Server fault on msg pass
  315423 hoton/bin/devi-hid

couldn't read stack: Server fault on msg pass
  327712 usr/photon/bin/pwm

couldn't read stack: Server fault on msg pass
  364570 bin/sh

couldn't read stack: Server fault on msg pass
  925707 bin/pidin
#
# pidin -pio-hid env
     pid name               Environment

couldn't read stack: Server fault on msg pass
  155668 sbin/io-hid
#
Colin Burgess(deleted)
04/15/2009 8:53 PM
post27163
RE: pidin map on 6.4.x   
I just committed a fix to services/system for this - it's nothing to do with pidin :-)

________________________________

From: Warren Deitch [mailto:community-noreply@qnx.com]
Sent: Wed 4/15/2009 8:13 PM
To: osmeta-core_os
Subject: Re: pidin map on 6.4.x



"pidin arg"  and  "pidin -Pio-hid arg" both fail.    MPC 8349 (procnto-600-instr - last update svn ver 218052)
# pidin arg
     pid Arguments
       1 proc/boot/procnto-600-instr

couldn't read arg: Server fault on msg pass
       2

couldn't read arg: Server fault on msg pass
       3

couldn't read arg: Server fault on msg pass
       4

couldn't read arg: Server fault on msg pass
   16389

couldn't read arg: Server fault on msg pass
   16390

couldn't read arg: Server fault on msg pass
   20487

----------------------
More instances snipped
----------------------

couldn't read arg: Server fault on msg pass
  315423

couldn't read arg: Server fault on msg pass
  327712

couldn't read arg: Server fault on msg pass
  364570

couldn't read arg: Server fault on msg pass
  909323
# pidin -pio-hid arg
     pid Arguments

couldn't read arg: Server fault on msg pass
  155668
#
# pidin env
     pid name               Environment
       1 /procnto-600-instr

couldn't read stack: Server fault on msg pass
       2 proc/boot/slogger

couldn't read stack: Server fault on msg pass
       3 proc/boot/io-usb

couldn't read stack: Server fault on msg pass
       4 oc/boot/devb-umass

couldn't read stack: Server fault on msg pass
   16389 sbin/tinit

couldn't read stack: Server fault on msg pass
   16390 sbin/pipe

couldn't read stack: Server fault on msg pass
   20487 sbin/mqueue

couldn't read stack: Server fault on msg pass
   40969 usr/sbin/inetd

--------
snip
--------

couldn't read stack: Server fault on msg pass
  315423 hoton/bin/devi-hid

couldn't read stack: Server fault on msg pass
  327712 usr/photon/bin/pwm

couldn't read stack: Server fault on msg pass
  364570 bin/sh

couldn't read stack: Server fault on msg pass
  925707 bin/pidin
#
# pidin -pio-hid env
     pid name               Environment

couldn't read stack: Server fault on msg pass
  155668 sbin/io-hid
#

_______________________________________________
OSMeta
http://community.qnx.com/sf/go/post27160
Attachment: Text winmail.dat 5.02 KB
Aleksandar Ristovski(deleted)
04/15/2009 10:07 PM
post27164
Re: pidin map on 6.4.x   
Colin Burgess wrote:
> I just committed a fix to services/system for this - it's nothing to do
> with pidin :-)

I was really confused since the message printed comes from 
code I haven't touched. Thanks Colin!

And thanks Warren for your feedback.

---
Aleksandar
Warren Deitch(deleted)
04/16/2009 9:59 AM
post27224
Re: pidin map on 6.4.x   
Tried it all again with the latest coreos update and the problem has gone.   All pidin xxx commands worked under procnto
-600-instr (updated today) .  Thanks to all of you!!!   :-)
I have a minor gripe with pidin irq - the columns do not line up if the irq reference or value  exceeds 9.   Given the 
amount of whitespace in front of the numbers, they could - pidin fds does.    

  131089   2 sbin/devf-generic
  131090   1 sbin/io-pkt-v4-hc
  131090   2 sbin/io-pkt-v4-hc
                6       0x80000000    0 T-- @0x480bd960:0x4813fafc
                7              0xb    0 T-- @0xfe385604:0x48172040
                8             0x21    0 T-- @0xfe391bbc:0x48180740
                9             0x22    0 T-- @0xfe391b74:0x48180740
                10             0x24    0 T-- @0xfe391bbc:0x4817eac0
                11             0x25    0 T-- @0xfe391b74:0x4817eac0
  131090   3 sbin/io-pkt-v4-hc
  131090   4 sbin/io-pkt-v4-hc
  131090   5 sbin/io-pkt-v4-hc
  131094   1 sbin/io-display
  131094   2 sbin/io-display
Colin Burgess(deleted)
04/16/2009 10:02 AM
post27228
Re: pidin map on 6.4.x   
Picky, picky! :-)

Warren Deitch wrote:
> Tried it all again with the latest coreos update and the problem has gone.   All pidin xxx commands worked under 
procnto-600-instr (updated today) .  Thanks to all of you!!!   :-)
> I have a minor gripe with pidin irq - the columns do not line up if the irq reference or value  exceeds 9.   Given the
 amount of whitespace in front of the numbers, they could - pidin fds does.    
> 
>   131089   2 sbin/devf-generic
>   131090   1 sbin/io-pkt-v4-hc
>   131090   2 sbin/io-pkt-v4-hc
>                 6       0x80000000    0 T-- @0x480bd960:0x4813fafc
>                 7              0xb    0 T-- @0xfe385604:0x48172040
>                 8             0x21    0 T-- @0xfe391bbc:0x48180740
>                 9             0x22    0 T-- @0xfe391b74:0x48180740
>                 10             0x24    0 T-- @0xfe391bbc:0x4817eac0
>                 11             0x25    0 T-- @0xfe391b74:0x4817eac0
>   131090   3 sbin/io-pkt-v4-hc
>   131090   4 sbin/io-pkt-v4-hc
>   131090   5 sbin/io-pkt-v4-hc
>   131094   1 sbin/io-display
>   131094   2 sbin/io-display
> 
> 
> _______________________________________________
> OSMeta
> http://community.qnx.com/sf/go/post27224
> 

-- 
cburgess@qnx.com
Ashwin Patwekar
07/08/2009 8:29 AM
post33373
Re: pidin map on 6.4.x   
Not related to the question posted. But by the way is fs-yaffs supported under 6.4.x ? Is the source under SVN?

Thanks,
Ashwin
Colin Burgess(deleted)
07/08/2009 9:11 AM
post33384
Re: pidin map on 6.4.x   
fs-yaffs is not a QNX binary - it must be Warren's.

Ashwin Patwekar wrote:
> Not related to the question posted. But by the way is fs-yaffs supported under 6.4.x ? Is the source under SVN?
> 
> Thanks,
> Ashwin
> 
> _______________________________________________
> OSMeta
> http://community.qnx.com/sf/go/post33373
> 

-- 
cburgess@qnx.com
Warren Deitch(deleted)
07/08/2009 7:45 PM
post33441
Re: pidin map on 6.4.x   
The people who provided the hardware also provided Rob Krten's yaffs package (from Parse Software Devices).   It has 
serious problems executing scripts from the filesystem but I have not recently looked for a QNX nand filesystem that 
supports the Micron MT29F2G08. 
That is still on my todo list.

Return


  A subsidiary of BlackBerry             All content ©2004-2013, QNX Software Systems