[PATCH] aic94xx: convert to request_firmware interface

[PATCH] aic94xx: convert to request_firmware interface

[James Bottomley]

[-- Attachment #1: Type: text/plain, Size: 183 bytes --]

This patch is too big to inline, so I attached it compressed.

This, I think, concludes the work necessary to get this thing into
scsi-misc and begin working on it upstream.

James



[-- Attachment #2: tmp.diff.gz --]
[-- Type: application/x-gzip, Size: 28027 bytes --]

Re: [PATCH] aic94xx: convert to request_firmware interface

[James Bottomley]

On Sun, 2006-08-27 at 21:48 -0500, James Bottomley wrote:
> This patch is too big to inline, so I attached it compressed.
> 
> This, I think, concludes the work necessary to get this thing into
> scsi-misc and begin working on it upstream.

I forgot to say that the firmware file is here:

http://www.kernel.org/pub/linux/kernel/people/jejb/aic94xx-seq.fw

James

Re: [PATCH] aic94xx: convert to request_firmware interface

[Douglas Gilbert]

James Bottomley wrote:
> On Sun, 2006-08-27 at 21:48 -0500, James Bottomley wrote:
>> This patch is too big to inline, so I attached it compressed.
>>
>> This, I think, concludes the work necessary to get this thing into
>> scsi-misc and begin working on it upstream.
> 
> I forgot to say that the firmware file is here:
> 
> http://www.kernel.org/pub/linux/kernel/people/jejb/aic94xx-seq.fw

James,
I presume that is a version of the infamous "razor"
code. How do I know which version it is (before trying
it)?

Doug Gilbert

Re: [PATCH] aic94xx: convert to request_firmware interface

[James Bottomley]

On Mon, 2006-08-28 at 08:52 -0400, Douglas Gilbert wrote:
> I presume that is a version of the infamous "razor"
> code. How do I know which version it is (before trying
> it)?

Well ... same way you do any binary blob with version text in it ... use
strings.

James

Re: [PATCH] aic94xx: convert to request_firmware interface

[Luben Tuikov]

--- James Bottomley <James.Bottomley@SteelEye.com> wrote:

> Re: [PATCH] aic94xx: convert to request_firmware interface
> This patch is too big to inline, so I attached it compressed.

This will not work.

> This, I think, concludes the work necessary to get this thing into
> scsi-misc and begin working on it upstream.

No, this is simply metaphorically to put it, the "doctor
examining themselves and saying ``I'm healthy, everything is alright.''".

     Luben

Re: [PATCH] aic94xx: convert to request_firmware interface

[Arjan van de Ven]

On Mon, 2006-08-28 at 08:52 -0700, Luben Tuikov wrote:
> --- James Bottomley <James.Bottomley@SteelEye.com> wrote:
> 
> > Re: [PATCH] aic94xx: convert to request_firmware interface
> > This patch is too big to inline, so I attached it compressed.
> 
> This will not work.


Hi,

I think your mailer has started corrupting emails again, it cut off the
"because ..." part of the sentence.... (or did you try to avoid getting
hit by the spam traps?)

Greetings,
   Arjan van de Ven

Re: [PATCH] aic94xx: convert to request_firmware interface

[James Bottomley]

On Mon, 2006-08-28 at 08:52 -0700, Luben Tuikov wrote:
> --- James Bottomley <James.Bottomley@SteelEye.com> wrote:
> 
> > Re: [PATCH] aic94xx: convert to request_firmware interface
> > This patch is too big to inline, so I attached it compressed.
> 
> This will not work.

It's currently working in my x86 system.

I haven't managed to test on a BE system yet to make sure I got all the
swaps correct, though.

James

Re: [PATCH] aic94xx: convert to request_firmware interface

[Douglas Gilbert]

James Bottomley wrote:
> On Mon, 2006-08-28 at 08:52 -0400, Douglas Gilbert wrote:
>> I presume that is a version of the infamous "razor"
>> code. How do I know which version it is (before trying
>> it)?
> 
> Well ... same way you do any binary blob with version text in it ... use
> strings.

James,
What I'm angling at is a firmware file name like:
   aic94xx-seq_10c6.req

and the driver attempts to read:
   aic94xx-seq*.req

If there a multiple matches then it takes the one
with the most recent date (reporting this in syslog).


For the benefit of others: in my experience the
stability of my adaptec 48300 SAS HBA is somewhat
related to the firmware level it is running (both
the card BIOS and when the linux driver has
taken over).

Doug Gilbert

Re: [PATCH] aic94xx: convert to request_firmware interface

[James Bottomley]

On Mon, 2006-08-28 at 16:43 -0400, Douglas Gilbert wrote:
> What I'm angling at is a firmware file name like:
>    aic94xx-seq_10c6.req

I don't think that's a good idea ... the current razor firmware is
backwards compatible, so if I hard code a version into the file there'd
be no upgrade ability.

> and the driver attempts to read:
>    aic94xx-seq*.req

It can't do that from the kernel ... this is all conditioned on how the
user level firmware loader actually works

> If there a multiple matches then it takes the one
> with the most recent date (reporting this in syslog).

The way you do this is to have a generic name (aic94xx-seq.fw) and then
symbolically link it to the versioned file.

The firmware itself is versioned, so if the driver comes to require a
particular version (or later), it can be programmed to complain.

James

Re: [PATCH] aic94xx: convert to request_firmware interface

[Mike Anderson]

James Bottomley <James.Bottomley@SteelEye.com> wrote:
> On Mon, 2006-08-28 at 08:52 -0700, Luben Tuikov wrote:
> > --- James Bottomley <James.Bottomley@SteelEye.com> wrote:
> > 
> > > Re: [PATCH] aic94xx: convert to request_firmware interface
> > > This patch is too big to inline, so I attached it compressed.
> > 
> > This will not work.
> 
> It's currently working in my x86 system.
> 
> I haven't managed to test on a BE system yet to make sure I got all the
> swaps correct, though.

The firmware loading is working on my ppc64 system.

-andmike
--
Michael Anderson
andmike@us.ibm.com

Re: [PATCH] aic94xx: convert to request_firmware interface

[Douglas Gilbert]

James Bottomley wrote:
> On Mon, 2006-08-28 at 16:43 -0400, Douglas Gilbert wrote:
>> What I'm angling at is a firmware file name like:
>>    aic94xx-seq_10c6.req
> 
> I don't think that's a good idea ... the current razor firmware is
> backwards compatible, so if I hard code a version into the file there'd
> be no upgrade ability.
> 
>> and the driver attempts to read:
>>    aic94xx-seq*.req
> 
> It can't do that from the kernel ... this is all conditioned on how the
> user level firmware loader actually works
> 
>> If there a multiple matches then it takes the one
>> with the most recent date (reporting this in syslog).
> 
> The way you do this is to have a generic name (aic94xx-seq.fw) and then
> symbolically link it to the versioned file.
> 
> The firmware itself is versioned, so if the driver comes to require a
> particular version (or later), it can be programmed to complain.

Which may explain why firmware for a WIFI card that
I have has a versioned filename with a README in the
distribution tarball. The README says to add a symlink.
The README also tells me which directory to put the
firmware file and symlink in.

Doug Gilbert