| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Thu, 2 Dec 2021 10:44:29 +0100
From: Claudio Suarez <cssk@...-c.es>
To: dri-devel@...ts.freedesktop.org, 0day robot <lkp@...el.com>,
LKML <linux-kernel@...r.kernel.org>, lkp@...ts.01.org,
Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
Maxime Ripard <mripard@...nel.org>,
Thomas Zimmermann <tzimmermann@...e.de>,
David Airlie <airlied@...ux.ie>
Subject: Re: [PATCH] drm: fix error found in some cases after the patch
d1af5cd86997
On Tue, Nov 30, 2021 at 09:38:11AM +0100, Daniel Vetter wrote:
> On Mon, Nov 29, 2021 at 08:27:45PM +0100, Claudio Suarez wrote:
> > The patch d1af5cd86997 ("drm: get rid of DRM_DEBUG_* log
> > calls in drm core, files drm_a*.c") fails when the drm_device
> > cannot be found in the parameter plane_state. Fix it.
> >
> > Reported-by: kernel test robot <oliver.sang@...el.com>
> > Fixes: d1af5cd86997 ("drm: get rid of DRM_DEBUG_* log calls in drm core, files drm_a*.c")
> > Signed-off-by: Claudio Suarez <cssk@...-c.es>
> > ---
> > drivers/gpu/drm/drm_atomic_helper.c | 12 +++++-------
> > 1 file changed, 5 insertions(+), 7 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
> > index aef2fbd676e5..8bd4472d7949 100644
> > --- a/drivers/gpu/drm/drm_atomic_helper.c
> > +++ b/drivers/gpu/drm/drm_atomic_helper.c
> > @@ -312,7 +312,7 @@ update_connector_routing(struct drm_atomic_state *state,
> >
> > if (!new_connector_state->crtc) {
> > drm_dbg_atomic(connector->dev, "Disabling [CONNECTOR:%d:%s]\n",
> > - connector->base.id, connector->name);
> > + connector->base.id, connector->name);
>
> Can you pls split this checkpatch fix out?
Of course.
>
> >
> > set_best_encoder(state, new_connector_state, NULL);
> >
> > @@ -805,6 +805,7 @@ int drm_atomic_helper_check_plane_state(struct drm_plane_state *plane_state,
> > bool can_update_disabled)
> > {
> > struct drm_framebuffer *fb = plane_state->fb;
> > + struct drm_device *dev = plane_state->plane ? plane_state->plane->dev : NULL;
>
> For real drivers plane_state->plane really should never be NULL, and
> looking at the test report we blow up in an selftest. I think the right
> fix here is to make the selftest more robust and also mock a drm_plane
> (with a NULL plane->dev pointer, that should be fine).
>
> Can you pls spin that and test it with the selftests enabled in the
> config?
You are right. I made this change in the test and the error was gone.
Output before and after follow.
The code in the kernel should be fixed though. Currently plane_state->crtc
is used. It should be plane_state->plane
I going to send two patches: one for the test and a v2 for the drm code.
Output before changing the test:
[ 38.161315][ T1] drm_mm: bottom-up fragmented insert of 10000 and 20000 insertions took 7047892 and 14424064 nsecs
[ 38.191061][ T1] drm_mm: top-down fragmented insert of 10000 and 20000 insertions took 7114502 and 14582794 nsecs
[ 40.260837][ T1] BUG: kernel NULL pointer dereference, address: 00000000
[ 40.261381][ T1] #PF: supervisor read access in kernel mode
[ 40.261845][ T1] #PF: error_code(0x0000) - not-present page
[ 40.262346][ T1] *pde = 00000000
[ 40.262692][ T1] Oops: 0000 [#1] PREEMPT SMP
[ 40.263061][ T1] CPU: 1 PID: 1 Comm: swapper/0 Tainted: G S W 5.16.0-rc2-00002-g58f3ee4dc1e9-dirty #1
[ 40.264036][ T1] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-2 04/01/2014
[ 40.264036][ T1] EIP: drm_atomic_helper_check_plane_state+0x15e/0x280
[ 40.264036][ T1] Code: 8b 45 d4 50 89 f0 e8 c1 0c 03 00 5a 80 7b 6c 00 74 6a 80 7d cc 00 75 64 8b 45 e4 39 43 5c 75 08 8b 45 ec 39 43 64 74 44 8b 03 <8b> 00 85 c0 74 03 8b 40 08 68 33 c2 70 c2 6a 04 50 e8 4c 14 04 00
[ 40.264036][ T1] EAX: 00000000 EBX: c036dce8 ECX: 08000000 EDX: 00000001
Output after:
[ 38.230609][ T1] drm_mm: bottom-up fragmented insert of 10000 and 20000 insertions took 7235202 and 14675224 nsecs
[ 38.259900][ T1] drm_mm: top-down fragmented insert of 10000 and 20000 insertions took 7078062 and 14642524 nsecs
[ 40.347207][ T1] random: get_random_bytes called from igt_dp_mst_sideband_msg_req_decode+0x25f/0x300 with crng_init=0
[ 40.347248][ T1] ACPI: bus type drm_connector registered
[ 40.350980][ T1] [drm] Initialized vgem 1.0.0 20120112 for vgem on minor 0
[ 40.351670][ T2] random: get_random_u32 called from copy_process+0x21c/0x1a00 with crng_init=0
[ 40.353310][ T1] [drm] Initialized vkms 1.0.0 20180514 for vkms on minor 1
BR
Claudio Suarez
>
> Thanks, Daniel
>
> > struct drm_rect *src = &plane_state->src;
> > struct drm_rect *dst = &plane_state->dst;
> > unsigned int rotation = plane_state->rotation;
> > @@ -828,8 +829,7 @@ int drm_atomic_helper_check_plane_state(struct drm_plane_state *plane_state,
> > }
> >
> > if (!crtc_state->enable && !can_update_disabled) {
> > - drm_dbg_kms(plane_state->crtc->dev,
> > - "Cannot update plane of a disabled CRTC.\n");
> > + drm_dbg_kms(dev, "Cannot update plane of a disabled CRTC.\n");
> > return -EINVAL;
> > }
> >
> > @@ -839,8 +839,7 @@ int drm_atomic_helper_check_plane_state(struct drm_plane_state *plane_state,
> > hscale = drm_rect_calc_hscale(src, dst, min_scale, max_scale);
> > vscale = drm_rect_calc_vscale(src, dst, min_scale, max_scale);
> > if (hscale < 0 || vscale < 0) {
> > - drm_dbg_kms(plane_state->crtc->dev,
> > - "Invalid scaling of plane\n");
> > + drm_dbg_kms(dev, "Invalid scaling of plane\n");
> > drm_rect_debug_print("src: ", &plane_state->src, true);
> > drm_rect_debug_print("dst: ", &plane_state->dst, false);
> > return -ERANGE;
> > @@ -864,8 +863,7 @@ int drm_atomic_helper_check_plane_state(struct drm_plane_state *plane_state,
> > return 0;
> >
> > if (!can_position && !drm_rect_equals(dst, &clip)) {
> > - drm_dbg_kms(plane_state->crtc->dev,
> > - "Plane must cover entire CRTC\n");
> > + drm_dbg_kms(dev, "Plane must cover entire CRTC\n");
> > drm_rect_debug_print("dst: ", dst, false);
> > drm_rect_debug_print("clip: ", &clip, false);
> > return -EINVAL;
> > --
> > 2.33.0
> >
> >
> >
>
> --
> Daniel Vetter
> Software Engineer, Intel Corporation
> http://blog.ffwll.ch
Powered by blists - more mailing lists