From: Michel Thierry <michel.thierry@intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>, intel-gfx@lists.freedesktop.org
Subject: Re: [RFC] drm/i915: Handle E2BIG error in i915_gem_fault
Date: Wed, 9 Sep 2015 17:53:57 +0100 [thread overview]
Message-ID: <55F06425.8020604@intel.com> (raw)
In-Reply-To: <20150909103319.GC32324@nuc-i3427.alporthouse.com>
On 9/9/2015 11:33 AM, Chris Wilson wrote:
> On Wed, Sep 09, 2015 at 11:29:52AM +0100, Michel Thierry wrote:
>> i915_gem_object_bind_to_vm returns -E2BIG when the user tries to bind an
>> object larger than the aperture, but i915_gem_fault does not handle this
>> return code:
>>
>> [501906.530985] gem_mmap_gtt: starting subtest big-bo-tiledX
>> [501906.541992] gem_mmap_gtt (22362): drop_caches: 3
>> [501906.610774] WARNING: CPU: 2 PID: 22362 at
>> drivers/gpu/drm/i915/i915_gem.c:1880 i915_gem_fault+0x24f/0x470 [i915]()
>> [501906.623568] unhandled error in i915_gem_fault: -7
>> [501906.825115] Call Trace:
>> [501906.830322] [<ffffffff8178ffcc>] dump_stack+0x45/0x57
>> [501906.838589] [<ffffffff810759aa>] warn_slowpath_common+0x8a/0xc0
>> [501906.847846] [<ffffffff81075a26>] warn_slowpath_fmt+0x46/0x50
>> [501906.856776] [<ffffffffc079591f>] i915_gem_fault+0x24f/0x470 [i915]
>> [501906.866276] [<ffffffff8119e11d>] __do_fault+0x3d/0xa0
>> [501906.874464] [<ffffffff81068cc0>] ? pte_alloc_one+0x30/0x50
>> [501906.883169] [<ffffffff811a26a7>] handle_mm_fault+0xe27/0x1810
>> [501906.892202] [<ffffffff81306e8a>] ? security_mmap_file+0xca/0xe0
>> [501906.900389] [<ffffffff811fb6ad>] ? do_vfs_ioctl+0x2cd/0x4b0
>> [501906.908143] [<ffffffff81063ada>] __do_page_fault+0x19a/0x430
>> [501906.916024] [<ffffffff81063d92>] do_page_fault+0x22/0x30
>> [501906.923532] [<ffffffff81799248>] page_fault+0x28/0x30
>>
>> RFC about the error code that should be returned by i915_gem_fault.
>
> There are two fixes here, change E2BIG to ENOSPC. The differentiate is
> painful to all consumers (and missing in userspace).
I'll change the return code in gem_object_bind_to_vm instead.
>
> The second is that gem_fault was supposed to be fixed to handle it and
> sigbus is the legimate error for that case (not enomem/sigsegv).
These subtests use objs tiling x/y, so the driver won't use partial views.
Should it be better if i915_gem_mmap_gtt checks for this (obj size
bigger than apperture and some tiling mode), and return an error at that
point? Instead of waiting for the sigbus error in gem_fault.
-Michel
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2015-09-09 16:53 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-09 10:29 [RFC] drm/i915: Handle E2BIG error in i915_gem_fault Michel Thierry
2015-09-09 10:33 ` Chris Wilson
2015-09-09 16:53 ` Michel Thierry [this message]
2015-09-09 17:44 ` Chris Wilson
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=55F06425.8020604@intel.com \
--to=michel.thierry@intel.com \
--cc=chris@chris-wilson.co.uk \
--cc=intel-gfx@lists.freedesktop.org \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox