Project Home
Project Home		 Documents
Documents		 Wiki
Wiki		 Discussion Forums
Discussions		 Project Information
Project Info
Filesystems/Discussion/General/2 x Spansion S29GL01GR 1Gbit devf-generic issues/List of Posts
Forum Topic - 2 x Spansion S29GL01GR 1Gbit devf-generic issues: (3 Items)
   
Update
Gervais Mulongoy
06/30/2010 2:28 PM
post58236
2 x Spansion S29GL01GR 1Gbit devf-generic issues   
Hello all:

Customer case where they have two 1G storage units sitting behind a custom FPGA which makes them appear to the system as
 a single 2GB device. (src: http://theloop.ott.qnx.com/departments/its/salesforce/sflookup/index.php?searchfield=
casenumber&id=00102881&lookup=+Lookup+Case+#case-description)


The customer reported seeing the following issues:

- an increase in CPU usage during a writes to flash - this happens when the partition spans both 1Gb storage units
- erratic behavior when device was 40-60% full
- amd_poll: 163 program error
(devf t3::f3s_aCFI_v2write:233) over poll waiting for write completion

I have advised them that the polling error is displayed when a 1-bit error status is toggled. Since it is only a single 
bit there isn't enough information to determine what the error is, so further testing is required (we will ignore this 
one for now).

I spoke with DaveS who suggested that they should create the largest partition possible (at least spanning both storage 
units) and write a pattern across the whole partition using the dd utility. I have also asked them to restart the driver
 with increased verbosity so we can verify that the driver is "seeing" the 2GB device correctly.

I don't know enough about devf-generic to analyze the verbose results so I could use a hand to confirm the logs indicate
 that they are using a 2GB device and that any errors can be accounted for. I also attach the one polling error they 
received during the pattern write (no indication where the writing was happening so I can't say if it was near or at the
 boundary between storage units - which should be invisible to the system afaik).

Thanks,

Gervais
Attachment: Text devf-logs.tgz 13.6 KB
David Sarrazin
06/30/2010 2:33 PM
post58237
RE: 2 x Spansion S29GL01GR 1Gbit devf-generic issues   
devf-generic is reporting that is sees two physical chips, spanning
0x800 128k sectors, and that the chips have a 16-bit data bus, and are
all wired to the same 16 data lines.

Did the pattern test work?

David

> -----Original Message-----
> From: Gervais Mulongoy [mailto:community-noreply@qnx.com] 
> Sent: June 30, 2010 2:29 PM
> To: general-filesystems
> Subject: 2 x Spansion S29GL01GR 1Gbit devf-generic issues
> 
> Hello all:
> 
> Customer case where they have two 1G storage units sitting 
> behind a custom FPGA which makes them appear to the system as 
> a single 2GB device. (src: 
> http://theloop.ott.qnx.com/departments/its/salesforce/sflookup
> /index.php?searchfield=casenumber&id=00102881&lookup=+Lookup+C
> ase+#case-description)
> 
> 
> The customer reported seeing the following issues:
> 
> - an increase in CPU usage during a writes to flash - this 
> happens when the partition spans both 1Gb storage units
> - erratic behavior when device was 40-60% full
> - amd_poll: 163 program error
> (devf t3::f3s_aCFI_v2write:233) over poll waiting for write completion
> 
> I have advised them that the polling error is displayed when 
> a 1-bit error status is toggled. Since it is only a single 
> bit there isn't enough information to determine what the 
> error is, so further testing is required (we will ignore this 
> one for now).
> 
> I spoke with DaveS who suggested that they should create the 
> largest partition possible (at least spanning both storage 
> units) and write a pattern across the whole partition using 
> the dd utility. I have also asked them to restart the driver 
> with increased verbosity so we can verify that the driver is 
> "seeing" the 2GB device correctly.
> 
> I don't know enough about devf-generic to analyze the verbose 
> results so I could use a hand to confirm the logs indicate 
> that they are using a 2GB device and that any errors can be 
> accounted for. I also attach the one polling error they 
> received during the pattern write (no indication where the 
> writing was happening so I can't say if it was near or at the 
> boundary between storage units - which should be invisible to 
> the system afaik).
> 
> Thanks,
> 
> Gervais
> 
> 
> 
> _______________________________________________
> 
> General
> http://community.qnx.com/sf/go/post58236
>
Gervais Mulongoy
06/30/2010 2:37 PM
post58239
Re: 2 x Spansion S29GL01GR 1Gbit devf-generic issues   
 They reported receiving just the one polling error during the dd write
- I don't know at which point of the dd it happened.

On 06/30/10 14:33, David Sarrazin wrote:
> devf-generic is reporting that is sees two physical chips, spanning
> 0x800 128k sectors, and that the chips have a 16-bit data bus, and are
> all wired to the same 16 data lines.
>
> Did the pattern test work?
>
> David
>
>> -----Original Message-----
>> From: Gervais Mulongoy [mailto:community-noreply@qnx.com] 
>> Sent: June 30, 2010 2:29 PM
>> To: general-filesystems
>> Subject: 2 x Spansion S29GL01GR 1Gbit devf-generic issues
>>
>> Hello all:
>>
>> Customer case where they have two 1G storage units sitting 
>> behind a custom FPGA which makes them appear to the system as 
>> a single 2GB device. (src: 
>> http://theloop.ott.qnx.com/departments/its/salesforce/sflookup
>> /index.php?searchfield=casenumber&id=00102881&lookup=+Lookup+C
>> ase+#case-description)
>>
>>
>> The customer reported seeing the following issues:
>>
>> - an increase in CPU usage during a writes to flash - this 
>> happens when the partition spans both 1Gb storage units
>> - erratic behavior when device was 40-60% full
>> - amd_poll: 163 program error
>> (devf t3::f3s_aCFI_v2write:233) over poll waiting for write completion
>>
>> I have advised them that the polling error is displayed when 
>> a 1-bit error status is toggled. Since it is only a single 
>> bit there isn't enough information to determine what the 
>> error is, so further testing is required (we will ignore this 
>> one for now).
>>
>> I spoke with DaveS who suggested that they should create the 
>> largest partition possible (at least spanning both storage 
>> units) and write a pattern across the whole partition using 
>> the dd utility. I have also asked them to restart the driver 
>> with increased verbosity so we can verify that the driver is 
>> "seeing" the 2GB device correctly.
>>
>> I don't know enough about devf-generic to analyze the verbose 
>> results so I could use a hand to confirm the logs indicate 
>> that they are using a 2GB device and that any errors can be 
>> accounted for. I also attach the one polling error they 
>> received during the pattern write (no indication where the 
>> writing was happening so I can't say if it was near or at the 
>> boundary between storage units - which should be invisible to 
>> the system afaik).
>>
>> Thanks,
>>
>> Gervais
>>
>>
>>
>> _______________________________________________
>>
>> General
>> http://community.qnx.com/sf/go/post58236
>>
>
>
> _______________________________________________
>
> General
> http://community.qnx.com/sf/go/post58237
>

Return


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