From: Dave Jones <davej@redhat.com>
To: Linux Kernel <linux-kernel@vger.kernel.org>, bskeggs@redhat.com
Cc: maarten.lankhorst@canonical.com
Subject: nouveau: check client.base.vm before dereferencing it.
Date: Tue, 30 Jul 2013 18:55:21 -0400 [thread overview]
Message-ID: <20130730225521.GA25895@redhat.com> (raw)
In-Reply-To: <20130725162203.GA32273@redhat.com>
On Thu, Jul 25, 2013 at 12:22:03PM -0400, Dave Jones wrote:
> This recently started happening (since the last DRM merge, 3.10 was fine).
>
> [ 17.751970] Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
> [ 17.753911] CPU: 3 PID: 292 Comm: systemd-udevd Not tainted 3.11.0-rc2+ #13
> [ 17.754392] EIP: 0060:[<f86dde0f>] EFLAGS: 00010296 CPU: 3
> [ 17.754542] EIP is at nouveau_bo_new+0x1f/0x28c [nouveau]
> [ 17.754647] EAX: 00000000 EBX: ed8610b0 ECX: 00000100 EDX: 00004000
> ..
> [ 17.756729] Call Trace:
> [ 17.756849] [<f850280f>] ? drm_mode_crtc_set_gamma_size+0x23/0x43 [drm]
> [ 17.756993] [<f86ef2f6>] nv04_crtc_create+0xd4/0x142 [nouveau]
> [ 17.757138] [<f86f2758>] nv04_display_create+0xf2/0x35a [nouveau]
> [ 17.757281] [<f86e801a>] nouveau_display_create+0x33f/0x553 [nouveau]
> [ 17.757422] [<f86d99fb>] nouveau_drm_load+0x22f/0x5dc [nouveau]
> [ 17.757534] [<c1391816>] ? device_register+0x17/0x1a
> [ 17.757648] [<f84ff77b>] ? drm_sysfs_device_add+0x76/0xa3 [drm]
> [ 17.757764] [<f84fe6bd>] drm_get_pci_dev+0x138/0x238 [drm]
> [ 17.757902] [<f86ab3ce>] ? nouveau_device_create_+0x65/0x11b [nouveau]
> [ 17.758044] [<f86da081>] nouveau_drm_probe+0x2d9/0x360 [nouveau]
> [ 17.758155] [<c12d55ac>] pci_device_probe+0x6c/0xb0
> [ 17.758261] [<c1393dae>] driver_probe_device+0x7f/0x356
> [ 17.758367] [<c139412c>] __driver_attach+0x74/0x76
> [ 17.758473] [<c13940b8>] ? __device_attach+0x33/0x33
> [ 17.758579] [<c13922b6>] bus_for_each_dev+0x49/0x74
> [ 17.758684] [<c13938ec>] driver_attach+0x1e/0x20
> [ 17.758791] [<c13940b8>] ? __device_attach+0x33/0x33
> [ 17.758896] [<c1393550>] bus_add_driver+0x1d0/0x27c
> [ 17.759002] [<c12d51c2>] ? pci_pm_suspend+0x111/0x111
> [ 17.759109] [<c12d51c2>] ? pci_pm_suspend+0x111/0x111
> [ 17.759215] [<c1394667>] driver_register+0x6a/0x123
> [ 17.759321] [<c12bdfad>] ? __raw_spin_lock_init+0x2d/0x4e
> [ 17.759428] [<c12d45f3>] __pci_register_driver+0x4a/0x4d
> [ 17.760008] [<f84fe8a3>] drm_pci_init+0xe6/0xee [drm]
> [ 17.760008] [<f8752000>] ? 0xf8751fff
> [ 17.760008] [<f8752048>] nouveau_drm_init+0x48/0x1000 [nouveau]
> [ 17.760008] [<c10003f0>] do_one_initcall+0xc0/0x180
> [ 17.760008] [<f8752000>] ? 0xf8751fff
> [ 17.760008] [<c1033765>] ? set_memory_nx+0x5a/0x5c
> [ 17.760008] [<c15aca75>] ? set_section_ro_nx+0x54/0x59
> [ 17.760008] [<c10b4174>] load_module+0x1ad6/0x2519
> [ 17.760008] [<c10b05ed>] ? copy_module_from_fd.isra.49+0x34/0x13b
> [ 17.760008] [<c10b4d0b>] SyS_finit_module+0x73/0xac
> [ 17.760008] [<c106b0cb>] ? up_write+0x1b/0x30
> [ 17.760008] [<c11343ca>] ? vm_mmap_pgoff+0x7a/0x97
> [ 17.760008] [<c15bf03b>] sysenter_do_call+0x12/0x32
> [ 17.760008] Code: c7 83 1c 01 00 00 ff ff ff ff eb aa 55 89 e5 57 56 53 83 ec 2c 66 66 66 66 90 89 d6 89 4d e8 8b b8 ec 03 00 00 8b 87 8c 00 00 00 <8b> 00 0f b6 88 91 00 00 00 b8 ff ff ff ff d3 e0 25 ff ff ff 7f
0108bc808107b97e101b15af9705729626be6447 introduced an oops due
to use doing a dereference on a variable that can validly be null.
Split up the new check, and add an additional check before
the potential NULL dereference.
Signed-off-by: Dave Jones <davej@redhat.com>
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 9b2c939..6c1ba6b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -198,13 +198,22 @@ nouveau_bo_new(struct drm_device *dev, int size, int align,
size_t acc_size;
int ret;
int type = ttm_bo_type_device;
- int max_size = INT_MAX & ~((1 << drm->client.base.vm->vmm->lpg_shift) - 1);
+ int max_size;
- if (size <= 0 || size > max_size) {
+ if (size <= 0) {
nv_warn(drm, "skipped size %x\n", (u32)size);
return -EINVAL;
}
+ if (drm->client.base.vm) {
+ max_size = INT_MAX & ~((1 << drm->client.base.vm->vmm->lpg_shift) - 1);
+
+ if (size > max_size) {
+ nv_warn(drm, "skipped size %x\n", (u32)size);
+ return -EINVAL;
+ }
+ }
+
if (sg)
type = ttm_bo_type_sg;
next prev parent reply other threads:[~2013-07-30 22:55 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-07-25 16:22 nouveau oops in nouveau_bo_new Dave Jones
2013-07-29 6:56 ` Johannes Hirte
2013-07-29 7:35 ` Johannes Hirte
2013-07-30 22:55 ` Dave Jones [this message]
2013-07-31 6:07 ` nouveau: check client.base.vm before dereferencing it Maarten Lankhorst
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=20130730225521.GA25895@redhat.com \
--to=davej@redhat.com \
--cc=bskeggs@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=maarten.lankhorst@canonical.com \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.