bnx2: broken firware loading

bnx2: broken firware loading

[Frank Reppin]

Hellos,

... firmware loading seems to be somewhat broken now with/since

https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/drivers/net/ethernet/broadcom/bnx2.c?id=3e1be7ad2d38c6bd6aeef96df9bd0a7822f4e51c

which gives me - on 4.8.6:

bnx2 0000:05:00.0: Direct firmware load for bnx2/bnx2-mips-06-6.2.3.fw 
failed with error -2

Reverting this commit does fix it (tested on 4.8.6 too) - but it's 
ofcourse not what you've initially intended.

TIA,
Frank Reppin

-- 
43rd Law of Computing:
         Anything that can go wr
fortune: Segmentation violation -- Core dumped

Re: bnx2: broken firware loading

[Baoquan He]

Yes, Paul has reported this. Sorry about this!

Will work on it right away.

Thanks
Baoquan

On 11/08/16 at 11:50pm, Frank Reppin wrote:
> Hellos,
> 
> ... firmware loading seems to be somewhat broken now with/since
> 
> https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/drivers/net/ethernet/broadcom/bnx2.c?id=3e1be7ad2d38c6bd6aeef96df9bd0a7822f4e51c
> 
> which gives me - on 4.8.6:
> 
> bnx2 0000:05:00.0: Direct firmware load for bnx2/bnx2-mips-06-6.2.3.fw
> failed with error -2
> 
> Reverting this commit does fix it (tested on 4.8.6 too) - but it's ofcourse
> not what you've initially intended.
> 
> TIA,
> Frank Reppin
> 
> -- 
> 43rd Law of Computing:
>         Anything that can go wr
> fortune: Segmentation violation -- Core dumped

Re: bnx2: broken firware loading

[Jay Rouman]

Did this get pushed upstream?  I see a new patch for 4.8.7 but the BNX2
regression doesn't seem to be fixed.  Or am I missing something?
-- 
Jay Rouman (jsr@edzone.net)
Gratiot/Isabella RESD, Ithaca, MI

> Yes, Paul has reported this. Sorry about this!
> 
> Will work on it right away.
> 
> Thanks
> Baoquan
> 
> On 11/08/16 at 11:50pm, Frank Reppin wrote:
> > Hellos,
> > 
> > ... firmware loading seems to be somewhat broken now with/since
> > 
> > https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/drivers/net/ethernet/broadcom/bnx2.c?id=3e1be7ad2d38c6bd6aeef96df9bd0a7822f4e51c
> > 
> > which gives me - on 4.8.6:
> > 
> > bnx2 0000:05:00.0: Direct firmware load for bnx2/bnx2-mips-06-6.2.3.fw
> > failed with error -2
> > 
> > Reverting this commit does fix it (tested on 4.8.6 too) - but it's ofcourse
> > not what you've initially intended.
> > 
> > TIA,
> > Frank Reppin
> > 
> > -- 
> > 43rd Law of Computing:
> >         Anything that can go wr
> > fortune: Segmentation violation -- Core dumped

Re: bnx2: broken firware loading

[Paul Menzel]

Dear Jay,


On 11/10/16 19:19, Jay Rouman wrote:
> Did this get pushed upstream?  I see a new patch for 4.8.7 but the BNX2
> regression doesn't seem to be fixed.  Or am I missing something?

I don’t think Baoquan’s patch is ready yet.

I guess he will put you in the receivers list, once he sends it for review.

As a workaround, you can put the firmware in the initramfs.


Kind regards,

Paul

Re: bnx2: broken firware loading

[Jay Rouman]

The part that confuses me is that the kernels that I use are custom
configured for a particular machine.  I don't use initramfs or even
allow module loading.  All the drivers are actually in the kernel.  As
far as I know, no firmware is loaded.  But the regression stops the BNX2
NIC from being discovered by the kernel.  Does this make sense?  It
seems like removing the firmware loading stops the discovery process
even though no firmware is actually loaded either way.
-- 
Jay Rouman (jsr@edzone.net)
Gratiot/Isabella RESD, Ithaca, MI

> On 11/10/16 19:19, Jay Rouman wrote:
> > Did this get pushed upstream?  I see a new patch for 4.8.7 but the BNX2
> > regression doesn't seem to be fixed.  Or am I missing something?
> 
> I don?t think Baoquan?s patch is ready yet.
> 
> I guess he will put you in the receivers list, once he sends it for review.
> 
> As a workaround, you can put the firmware in the initramfs.
> 
> 
> Kind regards,
> 
> Paul

Re: bnx2: broken firware loading

[Baoquan He]

On 11/11/16 at 08:44am, Jay Rouman wrote:
> The part that confuses me is that the kernels that I use are custom
> configured for a particular machine.  I don't use initramfs or even
> allow module loading.  All the drivers are actually in the kernel.  As
> far as I know, no firmware is loaded.  But the regression stops the BNX2
> NIC from being discovered by the kernel.  Does this make sense?  It
> seems like removing the firmware loading stops the discovery process
> even though no firmware is actually loaded either way.

I tried but didn't find a bnx2 chip spec. Now I post a patches to revert
my previous commit and add a function to only try to hard reset bnx2
chip but not involving firmware issues.

Will wait and see what bnx2 experts will say.

Sorry about the thouble I made to you.

> -- 
> Jay Rouman (jsr@edzone.net)
> Gratiot/Isabella RESD, Ithaca, MI
> 
> > On 11/10/16 19:19, Jay Rouman wrote:
> > > Did this get pushed upstream?  I see a new patch for 4.8.7 but the BNX2
> > > regression doesn't seem to be fixed.  Or am I missing something?
> > 
> > I don?t think Baoquan?s patch is ready yet.
> > 
> > I guess he will put you in the receivers list, once he sends it for review.
> > 
> > As a workaround, you can put the firmware in the initramfs.
> > 
> > 
> > Kind regards,
> > 
> > Paul