LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Linus Torvalds <torvalds@linux-foundation.org>
To: Jonathan Corbet <corbet@lwn.net>
Cc: Yinghai Lu <yinghai@kernel.org>, Ingo Molnar <mingo@elte.hu>,
	Robert Hancock <hancockr@shaw.ca>,
	e1000-devel@lists.sourceforge.net,
	LKML <linux-kernel@vger.kernel.org>,
	Steven Rostedt <rostedt@goodmis.org>
Subject: Re: 2.6.28-rc2 hates my e1000e
Date: Sat, 1 Nov 2008 10:16:14 -0700 (PDT)	[thread overview]
Message-ID: <alpine.LFD.2.00.0811010958230.6226@nehalem.linux-foundation.org> (raw)
In-Reply-To: <20081101090154.3d014f57@bike.lwn.net>



On Sat, 1 Nov 2008, Jonathan Corbet wrote:

> Networking is fine in the absence of NFS.  I retried things and
> stress-tested it in a few ways with no trouble.  I think your last patch
> fixes the network card just fine.
> 
> Then I tried NFS again, watching more closely this time around.
> Everything locks up.  In fact, the soft lockup watchdog starts to
> scream:

Interesting. I wonder why it happens for NFS, but not apparently for all 
your other modules.

It does look very much like a ftrace issue, though, not NFS or 
network-related. Steven? Is this something that you are aware of already, 
with what looks like a lockup in ftrace_record_ip()?

> So methinks I'll add Steven to the Cc on this one :)  Looks like a
> different problem for sure.

Agreed. Looks unlikely to be related.

> > Oh, and getting the old (2.6.27) and new (2.6.28-rc2+patch) 
> > /proc/iomem would be nice.
> 
> For completeness, here they are.

Wow. Your BIOS really does screw up massively. The one reserved region 
difference is:

Old kernel (with lots of resources just re-assigned elsewhere):

> e0000000-fed003ff : reserved
>   fec00000-fec00fff : IOAPIC 0
>   fed00000-fed003ff : HPET 0

New kernel:

> e0000000-fed003ff : reserved
>   fe800000-fe8fffff : PCI Bus 0000:01
>   fe9d9b00-fe9d9bff : 0000:00:1f.3
>   fe9d9c00-fe9d9fff : 0000:00:1a.7
>     fe9d9c00-fe9d9fff : ehci_hcd
>   fe9da000-fe9dafff : 0000:00:03.3
>   fe9db000-fe9dbfff : 0000:00:19.0
>     fe9db000-fe9dbfff : e1000e
>   fe9dc000-fe9dffff : 0000:00:1b.0
>     fe9dc000-fe9dffff : ICH HD audio
>   fe9e0000-fe9fffff : 0000:00:19.0
>     fe9e0000-fe9fffff : e1000e
>   fea00000-fea7ffff : 0000:00:02.0
>   fea80000-feafffff : 0000:00:02.1
>   feb00000-febfffff : 0000:00:02.0
>   fec00000-fec00fff : IOAPIC 0
>   fed00000-fed003ff : HPET 0

ie the BIOS had marked a _lot_ of PCI allocations that it did as being 
reserved, and there was actually no partial overlap in your case. The old 
kernel would end up re-assigning all the resources (except for the magic 
non-PCI-BAR ones like the IOAPIC and the HPET) because of that BIOS 
reservation.

I do think that the new layout looks better, and I also think that 
"insert_resource_expand_to_fit()" did a much better and more logical job 
than "reserve_region_with_split()" did. So it looks like an improvement. I 
wonder who else with have breakage though - EVERY SINGLE TIME we do 
resource allocation cleanups/fixes, some odd firmware inevtiably breaks.

It's really sad. I worry that the old-style reserved handling hid bus 
where the firmware had assigned resources to insane locations (and then 
the reserved area code ended up forcing us to re-assign them to better 
ones). But my second patch at least -conceptually- makes sense, and 
obviously fixes your case, so I'm inclined to just commit it.

And either of the above two resource listings look saner than the plain 
-rc2 version (using reserve_region_with_split):

> e0000000-fe7fffff : reserved
> fe800000-fe8fffff : PCI Bus 0000:01
>  fe800000-fe8fffff : reserved
> fe900000-fe9d9aff : reserved
> fe9d9b00-fe9d9bff : 0000:00:1f.3
>  fe9d9b00-fe9d9bff : reserved
> fe9d9c00-fe9d9fff : 0000:00:1a.7
>  fe9d9c00-fe9d9fff : reserved
> fe9da000-fe9dafff : 0000:00:03.3
>  fe9da000-fe9dafff : reserved
> fe9db000-fe9dbfff : 0000:00:19.0
>  fe9db000-fe9dbfff : reserved
> fe9dc000-fe9dffff : 0000:00:1b.0
>  fe9dc000-fe9dffff : reserved
> fe9e0000-fe9fffff : 0000:00:19.0
>  fe9e0000-fe9fffff : reserved
> fea00000-fea7ffff : 0000:00:02.0
>  fea00000-fea7ffff : reserved
> fea80000-feafffff : 0000:00:02.1
>  fea80000-feafffff : reserved
> feb00000-febfffff : 0000:00:02.0
>  feb00000-febfffff : reserved
> fec00000-fed003ff : reserved
>  fec00000-fec00fff : IOAPIC 0
>  fed00000-fed003ff : HPET 0

.. which is just really messy, but is the same e0000000-fed003ff
"reserved" e820 entry just split and moved into each resource. 

I hate firmware. 

			Linus

  reply	other threads:[~2008-11-01 17:17 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <fa.P5jmIAOhSJ192sfWD+qlec07fe8@ifi.uio.no>
2008-10-31  0:45 ` Robert Hancock
2008-10-31  2:58   ` Jonathan Corbet
2008-10-31  4:08     ` Yinghai Lu
2008-10-31 15:07       ` Linus Torvalds
2008-10-31 15:44         ` Linus Torvalds
2008-10-31 16:00           ` Jonathan Corbet
2008-10-31 16:24             ` Linus Torvalds
2008-10-31 16:35               ` Yinghai Lu
2008-10-31 16:46                 ` Linus Torvalds
2008-10-31 16:51               ` Jonathan Corbet
2008-10-31 16:58                 ` Linus Torvalds
2008-10-31 20:20                   ` Yinghai Lu
2008-11-01 15:01                   ` Jonathan Corbet
2008-11-01 17:16                     ` Linus Torvalds [this message]
2008-11-01 17:35                       ` Steven Rostedt
2008-11-01 19:50                       ` Yinghai Lu
2008-11-01 22:45                         ` Jonathan Corbet
2008-11-01 22:47                           ` Yinghai Lu
2008-11-01 23:18                           ` Linus Torvalds
2008-11-02  1:26                             ` Robert Hancock
2008-10-30 23:44 Jonathan Corbet

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=alpine.LFD.2.00.0811010958230.6226@nehalem.linux-foundation.org \
    --to=torvalds@linux-foundation.org \
    --cc=corbet@lwn.net \
    --cc=e1000-devel@lists.sourceforge.net \
    --cc=hancockr@shaw.ca \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=rostedt@goodmis.org \
    --cc=yinghai@kernel.org \
    --subject='Re: 2.6.28-rc2 hates my e1000e' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).