aix79xx driver status

aix79xx driver status

[Strand Åge]

Hi,

I'm trying to debug an md RAID problem and was looking into aic79xx driver version, and now I'm confused.

The 2.6.13 kernel code lists the version for this driver as 1.3.11 (AIC79XX_DRIVER_VERSION), and it looks like the last update was back in 2003.

But I can find a 2.0.14 version of the driver at adaptec.com, dated 17.May 2005:
http://www.adaptec.com/worldwide/support/driversbycat.jsp?sess=no&language=English+US&cat=%2FOperating+System%2FLinux+Driver+Source+Code

Granted, it says the driver is "minimally tested", but I wonder why the aic79xx driver in the kernel hasn't been updated since 2003 when there are newer versions being developed by the vendor.

Is there a fork in the aic79xx driver tree? What was the reason? Is there any point in even trying the "minimally tested" driver from adaptec, or is that to be considered dangerous?

I'm having problems with an md RAID running off an Adaptec 39320A-R card (14 disks, one hot spare). I'm getting PKT busfree errors at random intervals which result in RAID disks going offline - so I'm keen to see if a driver update may help. 


Best regards,
Aage Strand

Re: aix79xx driver status

[Jens Laas]

(05.09.21 kl.11:44) Strand Åge skrev följande till linux-scsi@vger.kernel.org:

>
> Hi,
>
> I'm trying to debug an md RAID problem and was looking into aic79xx 
> driver version, and now I'm confused.

> I'm having problems with an md RAID running off an Adaptec 39320A-R card 
> (14 disks, one hot spare). I'm getting PKT busfree errors at random 
> intervals which result in RAID disks going offline - so I'm keen to see 
> if a driver update may help.

Ill just add that we have seen PKT busfree errors too (which take the card 
offline) with this driver.

(Adaptec ASC-39320D if I remember correctly).

Also confused about versions and how they relate.

Cheers,
Jens Låås

>
>
> Best regards,
> Aage Strand
> -
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

-----------------------------------------------------------------------
     'This mail automatically becomes portable when carried.'
-----------------------------------------------------------------------
     Jens Låås                              Email: jens.laas@data.slu.se
     Department of Computer Services, SLU   Phone: +46 18 67 35 15
     Vindbrovägen 1
     P.O. Box 7079
     S-750 07 Uppsala
     SWEDEN
-----------------------------------------------------------------------
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

RE: aix79xx driver status

[Strand Åge]

>-----Original Message-----
>From: Jens Laas [mailto:jens.laas@data.slu.se]
>Sent: Wed 9/21/2005 12:54 PM
>To: Strand Åge
>Cc: linux-scsi@vger.kernel.org
>Subject: Re: aix79xx driver status
> 
>(05.09.21 kl.11:44) Strand Åge skrev följande till linux-scsi@vger.kernel.org:
>
>>
>> Hi,
>>
>> I'm trying to debug an md RAID problem and was looking into aic79xx 
>> driver version, and now I'm confused.
>
>> I'm having problems with an md RAID running off an Adaptec 39320A-R card 
>> (14 disks, one hot spare). I'm getting PKT busfree errors at random 
>> intervals which result in RAID disks going offline - so I'm keen to see 
>> if a driver update may help.
>
>Ill just add that we have seen PKT busfree errors too (which take the card 
>offline) with this driver.
>
>(Adaptec ASC-39320D if I remember correctly).

Hm. "Good" to hear I'm not alone :-)
Have you tried replacing the controller with some other brand? 

I wonder if this is just an Adaptec related issue or if it relates to problems with the md driver as well.
Sometimes the PKT bug triggers another bug in the md driver which basically makes the md driver try to update the superblock 100 times (hard coded value) before failing. I'm running a 2.6.9 kernel, the md bug has been fixed in later kernels, according to Neil Brown - IIRC. 

>Also confused about versions and how they relate.

Yes, the version numbering is a bit confusing. On the linux-scsi mailing list I have seen messages about paches going into the aic7xxx/79xx tree as late as 2004, but there has been no update to the version numbers. It would be nice if someone could explain that one for me.

Anyhow, according to another message on this list there was a change in maintainer back i June 2004, where Justin T. Gibbs gave it up to Luben Tuikov, then working for Adaptec. I'll try to see if I can get in touch with him. 

Cheers,
aage
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: aix79xx driver status

[Christoph Hellwig]

On Wed, Sep 21, 2005 at 11:44:28AM +0200, Strand ?ge wrote:
> 
> Hi,
> 
> I'm trying to debug an md RAID problem and was looking into aic79xx driver version, and now I'm confused.
> 
> The 2.6.13 kernel code lists the version for this driver as 1.3.11 (AIC79XX_DRIVER_VERSION), and it looks like the last update was back in 2003.
> 
> But I can find a 2.0.14 version of the driver at adaptec.com, dated 17.May 2005:
> http://www.adaptec.com/worldwide/support/driversbycat.jsp?sess=no&language=English+US&cat=%2FOperating+System%2FLinux+Driver+Source+Code
> 
> Granted, it says the driver is "minimally tested", but I wonder why the aic79xx driver in the kernel hasn't been updated since 2003 when there are newer versions being developed by the vendor.
> 
> Is there a fork in the aic79xx driver tree? What was the reason? Is there any point in even trying the "minimally tested" driver from adaptec, or is that to be considered dangerous?

Yes, there was a fork.  The kernel aic79xx got various fixes and got
integerated better with the scsi subsytem, using generic, fully debugged
code instead of it's own, etc..  Unfortunately there's also various fixes
in the adaptec releases which need to be brought back.  They are sometimes
non-trivial and until a short time ago non of the core SCSI developers
had aic79xx hardware.  The fixes will be merged back one day, similar
to the aic7xxx where this happened.

Re: aix79xx driver status

[Age Strand]

On Wed, 2005-09-21 at 14:29 +0100, Christoph Hellwig wrote:
> On Wed, Sep 21, 2005 at 11:44:28AM +0200, Strand ?ge wrote:
> > 
> > Hi,
> > 
> > I'm trying to debug an md RAID problem and was looking into aic79xx driver version, and now I'm confused.
> > 
> > The 2.6.13 kernel code lists the version for this driver as 1.3.11 (AIC79XX_DRIVER_VERSION), and it looks like the last update was back in 2003.
> > 
> > But I can find a 2.0.14 version of the driver at adaptec.com, dated 17.May 2005:
> > http://www.adaptec.com/worldwide/support/driversbycat.jsp?sess=no&language=English+US&cat=%2FOperating+System%2FLinux+Driver+Source+Code
> > 
> > Granted, it says the driver is "minimally tested", but I wonder why the aic79xx driver in the kernel hasn't been updated since 2003 when there are newer versions being developed by the vendor.
> > 
> > Is there a fork in the aic79xx driver tree? What was the reason? Is there any point in even trying the "minimally tested" driver from adaptec, or is that to be considered dangerous?
> 
> Yes, there was a fork.  The kernel aic79xx got various fixes and got
> integerated better with the scsi subsytem, using generic, fully debugged
> code instead of it's own, etc..  Unfortunately there's also various fixes
> in the adaptec releases which need to be brought back.  They are sometimes
> non-trivial and until a short time ago non of the core SCSI developers
> had aic79xx hardware.  The fixes will be merged back one day, similar
> to the aic7xxx where this happened.

As far as I understand you the aic7Yxx kernel driver tree is very much
alive and kicking.

So what's the deal with the version numbers on the aic79xx driver? It
has stayed put on 1.1.13 since 2003. Does that mean that there hasn't
been any new releases of this driver since then? 

I seem to recall seeing submitted patches to aic79xx on the list as late
as 2004. What's become of those? Is this just a case of patches going in
but version numbering not being updated?

Regarding lack of hardware for developers: Are you looking for hardware
donors (e.g. of 39320 cards), and would that aid development on the
aic79xx driver?


Best regards,
aage

Re: aix79xx driver status

[Xose Vazquez Perez]

Age Strand  wrote:

> As far as I understand you the aic7Yxx kernel driver tree is very much
> alive and kicking.

> So what's the deal with the version numbers on the aic79xx driver? It
> has stayed put on 1.1.13 since 2003. Does that mean that there hasn't
> been any new releases of this driver since then?

> I seem to recall seeing submitted patches to aic79xx on the list as late
> as 2004. What's become of those? Is this just a case of patches going in
> but version numbering not being updated?

aic7Yxx drivers have been maintained very badly in the _kernel_ .
Outside, Adaptec people has their own tree.

You can get the latest ADAPTEC patches from:
http://marc.theaimsgroup.com/?a=108731197000012&r=5&w=2
http://marc.theaimsgroup.com/?a=108731197000012&r=4&w=2
under the subject "patch XX/28] Sync up drivers/scsi/aic7xxx".

But maybe,  there are newer patches inside Adaptec,
ask it to Luben Tuikov <luben_tuikov@adaptec.com>


This is a _very critical_ driver and I hope someone take it back to life again.

Re: aix79xx driver status

[Xose Vazquez Perez]

Age Strand  wrote:

> As far as I understand you the aic7Yxx kernel driver tree is very much
> alive and kicking.

> So what's the deal with the version numbers on the aic79xx driver? It
> has stayed put on 1.1.13 since 2003. Does that mean that there hasn't
> been any new releases of this driver since then?

> I seem to recall seeing submitted patches to aic79xx on the list as late
> as 2004. What's become of those? Is this just a case of patches going in
> but version numbering not being updated?

aic7Yxx drivers have been maintained very badly in the _kernel_ .
Outside, Adaptec people has their own tree.

You can get the latest ADAPTEC patches from:
http://marc.theaimsgroup.com/?a=108731197000012&r=5&w=2
http://marc.theaimsgroup.com/?a=108731197000012&r=4&w=2
under the subject "patch XX/28] Sync up drivers/scsi/aic7xxx".

never applied because
:http://marc.theaimsgroup.com/?l=linux-scsi&m=109637681104990&w=2

But maybe,  there are newer patches inside Adaptec,
ask it to Luben Tuikov <luben_tuikov@adaptec.com>


This is a _very critical_ driver and I hope someone take it back to life again.

Re: aix79xx driver status

[Age Strand]

On Thu, 2005-09-22 at 11:54 +0200, Xose Vazquez Perez wrote:
> 
> This is a _very critical_ driver and I hope someone take it back to life again.

True. These cards are getting more and more mainstream so it would make
sense to bring it up to date, i.e. merge back patches from vendor, etc.

According to Luben Tuikov the maintainers of the kernel driver at the
moment are James Bottomley and Christoph Hellwig.

Christoph has already answered some of my questions about the driver.


aage