Re: [PATCH] x86, irq: Keep IRQ assignment for PCI devices during suspend/hibernation

[Borislav Petkov <bp-Gina5bIWoIWzQB+pC5nmwQ@public.gmane.org>]

Hi,

On Thu, Jul 31, 2014 at 10:41:36PM +0800, Jiang Liu wrote:
> 	Really appreciate your help. There are two issues left
> according to the log messages.

first of all:

A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?

So please refrain from top-posting on lkml.

> 	The first issue is " WARNING: CPU: 2 PID: 3786 at
> fs/proc/generic.c:521 remove_proc_entry+0x19c/0x1b0()".
> It's caused by an issue in usb driver and a patch is ready to fix
> this issue, please refer to https://lkml.org/lkml/2014/7/20/277.
> This patch has been merged into Greg's usb repository, so it should
> be merge into v3.17 too.

I can apply this patch to test.

> 	The second issue is "AMD-Vi: Event logged [IO_PAGE_FAULT
> device=00:12.0 domain=0x0009 address=0x0000000000000000 flags=0x0000]".
> I feel it's not caused by IRQ related changes, but may be caused
> by IO page fault related work for AMD IOMMU. I need more time
> to investigate this and also CC Joerg for help too.

I don't think that's the issue - plain rc7 doesn't trigger the issue.
Only if I merge tip/master in it happens. And there are no AMD iommu
patches in tip/master since Joerg doesn't send stuff through tip
anymore.

> 	According to the log message, I feel my patch has fixed the
> original issue caused by IRQ related changes. What's your thoughts?

I still think the issue is in tip/master. I'm currently trying to
reproduce on another box so that I can bisect it. (Bisecting on my
workstation is always nasty :-\).

Thanks.

-- 
Regards/Gruss,
    Boris.

Sent from a fat crate under my desk. Formatting is fine.
--