Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.kernel > #58210 > unrolled thread

Bug#866511: kernel BUG at nvme/host/pci.c

Back to article view | Back to linux.debian.kernel

Contents

  Bug#866511: kernel BUG at nvme/host/pci.c Andreas Pflug <pgadmin@pse-consulting.de> - 2017-06-29 19:40 +0200
    Processed: Re: Bug#866511: kernel BUG at nvme/host/pci.c owner@bugs.debian.org (Debian Bug Tracking System) - 2017-07-05 19:40 +0200
    Bug#866511: kernel BUG at nvme/host/pci.c Ben Hutchings <ben@decadent.org.uk> - 2017-07-05 19:40 +0200
      Bug#866511: kernel BUG at nvme/host/pci.c Andreas Pflug <pgadmin@pse-consulting.de> - 2017-07-06 03:00 +0200
        Bug#866511: kernel BUG at nvme/host/pci.c Ben Hutchings <ben@decadent.org.uk> - 2017-07-06 13:20 +0200
          Bug#866511: kernel BUG at nvme/host/pci.c Andreas Pflug <pgadmin@pse-consulting.de> - 2017-07-10 17:40 +0200
            Bug#866511: kernel BUG at nvme/host/pci.c Ben Hutchings <ben@decadent.org.uk> - 2017-07-10 19:40 +0200
              Bug#866511: kernel BUG at nvme/host/pci.c Andreas Pflug <pgadmin@pse-consulting.de> - 2017-07-28 19:40 +0200
    Processed: Re: Bug#866511: kernel BUG at nvme/host/pci.c owner@bugs.debian.org (Debian Bug Tracking System) - 2017-07-06 13:20 +0200

#58210 — Bug#866511: kernel BUG at nvme/host/pci.c

Package: linux-image-4.9.0-3-amd64

Version: 4.9.30-2+deb9u2

When starting an ancient Windows Server 2003 virtual machine on a fully
updated stretch host, the host will crash repeatably about 10 seconds
after start. Kernel logging was implemented over netconsole, so I could
capture the kernel output stating

kernel BUG at /build/linux-9uDFZV/linux-4.9.30/drivers/nvme/host/pci.c:495!

Hardware is dual E5-2620v4 on Supermicro 10DRI-T with two SAMSUNG
MZQLW960HMJP-00003 NVME disks (mdadm RAID-1).

drivers/nvme/host/pci.c wasn't changed up to 4.11.8, the crashing
function nvme_setup_prps still the same in 4.12rc7, so I guess the bug
is still there.

[toc] | [next] | [standalone]

#58273 — Processed: Re: Bug#866511: kernel BUG at nvme/host/pci.c

Processing control commands:

> reassign -1 src:linux 4.9.30-2+deb9u2
Bug #866511 [linux-image-4.9.0-3-amd64] kernel BUG at nvme/host/pci.c
Bug reassigned from package 'linux-image-4.9.0-3-amd64' to 'src:linux'.
No longer marked as found in versions linux/4.9.30-2.
Ignoring request to alter fixed versions of bug #866511 to the same values previously set
Bug #866511 [src:linux] kernel BUG at nvme/host/pci.c
Marked as found in versions linux/4.9.30-2+deb9u2.
> tag -1 upstream patch moreinfo
Bug #866511 [src:linux] kernel BUG at nvme/host/pci.c
Added tag(s) upstream, patch, and moreinfo.

-- 
866511: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=866511
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems

[toc] | [prev] | [next] | [standalone]

#58274

[Multipart message — attachments visible in raw view] — view raw

Control: reassign -1 src:linux 4.9.30-2+deb9u2
Control: tag -1 upstream patch moreinfo

On Thu, 2017-06-29 at 19:30 +0200, Andreas Pflug wrote:
> Package: linux-image-4.9.0-3-amd64
> 
> Version: 4.9.30-2+deb9u2
> 
> When starting an ancient Windows Server 2003 virtual machine on a fully
> updated stretch host, the host will crash repeatably about 10 seconds
> after start. Kernel logging was implemented over netconsole, so I could
> capture the kernel output stating
> 
> kernel BUG at /build/linux-9uDFZV/linux-4.9.30/drivers/nvme/host/pci.c:495!

A single log line is rarely enough to be useful...

> Hardware is dual E5-2620v4 on Supermicro 10DRI-T with two SAMSUNG
> MZQLW960HMJP-00003 NVME disks (mdadm RAID-1).
> 
> drivers/nvme/host/pci.c wasn't changed up to 4.11.8, the crashing
> function nvme_setup_prps still the same in 4.12rc7, so I guess the bug
> is still there.

...but I happened to see a fix that mentioned this function.

This should make its way into a stable update eventually, but perhaps
you could confirm that the attached patch fixes your problem by
following these instructions:
https://kernel-handbook.alioth.debian.org/ch-common-tasks.html#s-common-official

Ben.

-- 
Ben Hutchings
Anthony's Law of Force: Don't force it, get a larger hammer.

[toc] | [prev] | [next] | [standalone]

#58276

[Multipart message — attachments visible in raw view] — view raw

Am 05.07.17 um 19:31 schrieb Ben Hutchings:
> Control: reassign -1 src:linux 4.9.30-2+deb9u2
> Control: tag -1 upstream patch moreinfo
>
> On Thu, 2017-06-29 at 19:30 +0200, Andreas Pflug wrote:
>> Package: linux-image-4.9.0-3-amd64
>>
>> Version: 4.9.30-2+deb9u2
>>
>> When starting an ancient Windows Server 2003 virtual machine on a fully
>> updated stretch host, the host will crash repeatably about 10 seconds
>> after start. Kernel logging was implemented over netconsole, so I could
>> capture the kernel output stating
>>
>> kernel BUG at /build/linux-9uDFZV/linux-4.9.30/drivers/nvme/host/pci.c:495!
> A single log line is rarely enough to be useful...
>
>> Hardware is dual E5-2620v4 on Supermicro 10DRI-T with two SAMSUNG
>> MZQLW960HMJP-00003 NVME disks (mdadm RAID-1).
>>
>> drivers/nvme/host/pci.c wasn't changed up to 4.11.8, the crashing
>> function nvme_setup_prps still the same in 4.12rc7, so I guess the bug
>> is still there.
> ...but I happened to see a fix that mentioned this function.
>
> This should make its way into a stable update eventually, but perhaps
> you could confirm that the attached patch fixes your problem by
> following these instructions:
> https://kernel-handbook.alioth.debian.org/ch-common-tasks.html#s-common-official

The description of the patch seems very promising, but using the
4.9.30-2+deb9u2a~test kernel the bug happens as well.
Log is attached, cut after 460 lines: the last trace on CPU15 is
repeated all over again, eventually leading to "Fixing recursive fault
but reboot is needed!"

Regards,
Andreas

[toc] | [prev] | [next] | [standalone]

#58282

[Multipart message — attachments visible in raw view] — view raw

Control: tag -1 - moreinfo patch

On Thu, 2017-07-06 at 02:47 +0200, Andreas Pflug wrote:
> Am 05.07.17 um 19:31 schrieb Ben Hutchings:
> > Control: reassign -1 src:linux 4.9.30-2+deb9u2
> > Control: tag -1 upstream patch moreinfo
> > 
> > On Thu, 2017-06-29 at 19:30 +0200, Andreas Pflug wrote:
> > > Package: linux-image-4.9.0-3-amd64
> > > 
> > > Version: 4.9.30-2+deb9u2
> > > 
> > > When starting an ancient Windows Server 2003 virtual machine on a fully
> > > updated stretch host, the host will crash repeatably about 10 seconds
> > > after start. Kernel logging was implemented over netconsole, so I could
> > > capture the kernel output stating
> > > 
> > > kernel BUG at /build/linux-9uDFZV/linux-4.9.30/drivers/nvme/host/pci.c:495!
> > 
> > A single log line is rarely enough to be useful...
> > 
> > > Hardware is dual E5-2620v4 on Supermicro 10DRI-T with two SAMSUNG
> > > MZQLW960HMJP-00003 NVME disks (mdadm RAID-1).
> > > 
> > > drivers/nvme/host/pci.c wasn't changed up to 4.11.8, the crashing
> > > function nvme_setup_prps still the same in 4.12rc7, so I guess the bug
> > > is still there.
> > 
> > ...but I happened to see a fix that mentioned this function.
> > 
> > This should make its way into a stable update eventually, but perhaps
> > you could confirm that the attached patch fixes your problem by
> > following these instructions:
> > https://kernel-handbook.alioth.debian.org/ch-common-tasks.html#s-common-official
> 
> The description of the patch seems very promising, but using the
> 4.9.30-2+deb9u2a~test kernel the bug happens as well.
> Log is attached, cut after 460 lines: the last trace on CPU15 is
> repeated all over again, eventually leading to "Fixing recursive fault
> but reboot is needed!"

Thanks for checking.

Ben.

-- 
Ben Hutchings
Man invented language to satisfy his deep need to complain. - Lily
Tomlin

[toc] | [prev] | [next] | [standalone]

#58332

Bug reported upstream as https://bugzilla.kernel.org/show_bug.cgi?id=196315

[toc] | [prev] | [next] | [standalone]

#58333

[Multipart message — attachments visible in raw view] — view raw

On Mon, 2017-07-10 at 17:30 +0200, Andreas Pflug wrote:
> Bug reported upstream as https://bugzilla.kernel.org/show_bug.cgi?id=196315

That seems to be the wrong place.  "Flash/Memory Technology Devices" is
for host-managed flash chips, not high-level interfaces like NVMe.  It
doesn't look like nvme driver bugs are accepted through Bugzilla.  Try
sending mail to linux-nvme@lists.infradead.org instead.

Ben.

-- 
Ben Hutchings
Absolutum obsoletum. (If it works, it's out of date.) - Stafford Beer

[toc] | [prev] | [next] | [standalone]

#58542

After discussing the issue on linux-nvme, the bug was located in
xen/biomerge.c. It can be fixed in 4.9.30 as well as 4.13RC1 using the
patch from Roger Pau Monne below.

---
 drivers/xen/biomerge.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/xen/biomerge.c b/drivers/xen/biomerge.c
index 4da69dbf7dca..1bdd02a6d6ac 100644
--- a/drivers/xen/biomerge.c
+++ b/drivers/xen/biomerge.c
@@ -10,8 +10,7 @@ bool xen_biovec_phys_mergeable(const struct bio_vec *vec1,
 	unsigned long bfn1 = pfn_to_bfn(page_to_pfn(vec1->bv_page));
 	unsigned long bfn2 = pfn_to_bfn(page_to_pfn(vec2->bv_page));
 
-	return __BIOVEC_PHYS_MERGEABLE(vec1, vec2) &&
-		((bfn1 == bfn2) || ((bfn1+1) == bfn2));
+	return bfn1 + PFN_DOWN(vec1->bv_offset + vec1->bv_len) == bfn2;
 #else
 	/*
 	 * XXX: Add support for merging bio_vec when using different page

[toc] | [prev] | [next] | [standalone]

#58281 — Processed: Re: Bug#866511: kernel BUG at nvme/host/pci.c

Processing control commands:

> tag -1 - moreinfo patch
Bug #866511 [src:linux] kernel BUG at nvme/host/pci.c
Removed tag(s) moreinfo and patch.

-- 
866511: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=866511
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.kernel

csiph-web