Re: 2.6.15-rc1-mm1

[Andrew Morton <akpm@osdl.org>]

Reuben Farrelly <reuben-lkml@reub.net> wrote:
>
> Hi,
> 
> On 18/11/2005 8:18 a.m., Andrew Morton wrote:
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.15-rc1/2.6.15-rc1-mm1
> > 
> > - reiser4 significantly updated
> > 
> > 
> > 
> > 
> > Changes since 2.6.14-mm2:
> 
> This has been one of the best -mm releases in a while.  No problems compiling
> or running - and so far nearly 18 hours uptime without any surprises.

We'll have to try harder.  -mm2 is up there now, to break everything again.

> Following up on a posting from the last -mm release,  I'm still seeing errors 
> loading multiple network drivers as modules (e100 and sky2) when 
> CONFIG_PREEMPT_BKL is enabled, with 2.6.14-mm1, 2.6.14-mm2 and now
> 2.6.15-rc1-mm1.  Mainline git doesn't exhibit the problem, so it's -mm specific.
> 
> This is what is logged:
> 
> Nov 18 17:40:42 tornado kernel: e100: 0000:06:03.0: e100_eeprom_load: EEPROM
> corrupted
> Nov 18 17:40:42 tornado kernel: ACPI: PCI interrupt for device 0000:06:03.0
> disabled
> Nov 18 17:40:42 tornado kernel: e100: probe of 0000:06:03.0 failed with error -11
> Nov 18 17:40:43 tornado kernel: ACPI: PCI Interrupt 0000:04:00.0[A] -> GSI 17
> (level, low) -> IRQ 177
> Nov 18 17:40:43 tornado kernel: sky2 0000:04:00.0: unsupported chip type 0xff
> Nov 18 17:40:43 tornado kernel: ACPI: PCI interrupt for device 0000:04:00.0
> disabled
> Nov 18 17:40:43 tornado kernel: sky2: probe of 0000:04:00.0 failed with error -95
> 
> I'm certain that both of these NIC's are OK as they work fine with 
> CONFIG_PREEMPT_BKL not selected.
> 
> With CONFIG_PREEMPT_BKL disabled and an otherwise identical config, the
> driver modules load up just fine.
> 
> A known good kernel with this config was 2.6.14-rc5-mm1.
> I have backed out git-netdev-all but it made no difference, as well as backed 
> out the e100 changes in -mm on 2.6.15-mm2, again no difference.  So I suspect 
> it's not a netdev driver problem.
> 
> What else can I do to help narrow down the problem?  What other trees or patches 
> would be worth backing out to try and narrow it down?

I'd be suspecting the PCI changes firstly.  That's gregkh-pci-*. 

Conceivably git-acpi, but that hasn't changed in quite some time.  In fact,
no ACPI changes since 2.6.14-rc5-mm1.

After that I don't know, sorry.   Binary search time?