without the obj[0] set, we can see the following panic:
[ 29.236764] pstate: 20400009 (nzCv daif +PAN -UAO) [ 29.241532] pc : drm_gem_vram_offset+0x10/0x28 [drm_vram_helper] [ 29.247511] lr : hibmc_plane_atomic_update+0x30/0xe0 [hibmc_drm] [ 29.253490] sp : ffff003fab713a50 [ 29.256789] x29: ffff003fab713a50 x28: 0000000000000028 [ 29.262077] x27: 0000000000000018 x26: 0000aaaaced52780 [ 29.267363] x25: ffff203f98dcde00 x24: 0000000000000001 [ 29.272650] x23: 0000000000000001 x22: 0000000000000000 [ 29.277937] x21: 0000000000000000 x20: ffff002fb4899580 [ 29.283224] x19: ffff003fb8767c00 x18: 0000000000000000 [ 29.288510] x17: 0000000000000000 x16: 0000000000000000 [ 29.293797] x15: 0000000000000010 x14: 332d4341443a3433 [ 29.299084] x13: 3a5245444f434e45 x12: 5b206e6f20746573 [ 29.304371] x11: 65646f6d205d7365 x10: 6c62617369645f74 [ 29.309657] x9 : ffff800008e4d650 x8 : ffff8000107241d0 [ 29.314944] x7 : 0000000000000000 x6 : ffff003fb8766400 [ 29.320231] x5 : ffff003fb8767c00 x4 : ffff800008e4f320 [ 29.325517] x3 : ffff002fb3ae5080 x2 : ffff002fb3ae0000 [ 29.330804] x1 : 0000000000000000 x0 : ffff002fb07f5c00 [ 29.336090] Call trace: [ 29.338528] drm_gem_vram_offset+0x10/0x28 [drm_vram_helper] [ 29.344160] drm_atomic_helper_commit_planes+0x144/0x228 [ 29.349447] drm_atomic_helper_commit_tail+0x38/0x78 [ 29.354389] commit_tail+0xb4/0x160 [ 29.357861] drm_atomic_helper_commit+0x134/0x168 [ 29.362543] drm_atomic_commit+0x50/0x60 [ 29.366448] drm_mode_atomic_ioctl+0x724/0xb00 [ 29.370871] drm_ioctl_kernel+0x98/0xe8 [ 29.374690] drm_ioctl+0x26c/0x3d0 [ 29.378077] do_vfs_ioctl+0xafc/0xbb0 [ 29.381722] ksys_ioctl+0x64/0x98 [ 29.385022] __arm64_sys_ioctl+0x24/0x50 [ 29.388929] el0_svc_common.constprop.3+0xf0/0x158 [ 29.393697] el0_svc_handler+0x68/0x80 [ 29.397429] el0_sync_handler+0xa8/0x198 [ 29.401333] el0_sync+0x140/0x180 [ 29.404632] ---[ end trace 4ba9406972d937dd ]--- [ 29.409245] Unexpected kernel BRK exception at EL1
Signed-off-by: Zhihui Chen chenzhihui4@huawei.com --- drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c | 1 + 1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c b/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c index 21b684eab5c9..19dc525a73b0 100644 --- a/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c +++ b/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c @@ -126,6 +126,7 @@ hibmc_framebuffer_init(struct drm_device *dev,
drm_helper_mode_fill_fb_struct(dev, &hibmc_fb->fb, mode_cmd); hibmc_fb->obj = obj; + hibmc_fb->fb.obj[0] = obj; ret = drm_framebuffer_init(dev, &hibmc_fb->fb, &hibmc_fb_funcs); if (ret) { DRM_ERROR("drm_framebuffer_init failed: %d\n", ret);
Hi
Am 23.12.19 um 04:08 schrieb Zhihui Chen:
without the obj[0] set, we can see the following panic:
[ 29.236764] pstate: 20400009 (nzCv daif +PAN -UAO) [ 29.241532] pc : drm_gem_vram_offset+0x10/0x28 [drm_vram_helper] [ 29.247511] lr : hibmc_plane_atomic_update+0x30/0xe0 [hibmc_drm] [ 29.253490] sp : ffff003fab713a50 [ 29.256789] x29: ffff003fab713a50 x28: 0000000000000028 [ 29.262077] x27: 0000000000000018 x26: 0000aaaaced52780 [ 29.267363] x25: ffff203f98dcde00 x24: 0000000000000001 [ 29.272650] x23: 0000000000000001 x22: 0000000000000000 [ 29.277937] x21: 0000000000000000 x20: ffff002fb4899580 [ 29.283224] x19: ffff003fb8767c00 x18: 0000000000000000 [ 29.288510] x17: 0000000000000000 x16: 0000000000000000 [ 29.293797] x15: 0000000000000010 x14: 332d4341443a3433 [ 29.299084] x13: 3a5245444f434e45 x12: 5b206e6f20746573 [ 29.304371] x11: 65646f6d205d7365 x10: 6c62617369645f74 [ 29.309657] x9 : ffff800008e4d650 x8 : ffff8000107241d0 [ 29.314944] x7 : 0000000000000000 x6 : ffff003fb8766400 [ 29.320231] x5 : ffff003fb8767c00 x4 : ffff800008e4f320 [ 29.325517] x3 : ffff002fb3ae5080 x2 : ffff002fb3ae0000 [ 29.330804] x1 : 0000000000000000 x0 : ffff002fb07f5c00 [ 29.336090] Call trace: [ 29.338528] drm_gem_vram_offset+0x10/0x28 [drm_vram_helper] [ 29.344160] drm_atomic_helper_commit_planes+0x144/0x228 [ 29.349447] drm_atomic_helper_commit_tail+0x38/0x78 [ 29.354389] commit_tail+0xb4/0x160 [ 29.357861] drm_atomic_helper_commit+0x134/0x168 [ 29.362543] drm_atomic_commit+0x50/0x60 [ 29.366448] drm_mode_atomic_ioctl+0x724/0xb00 [ 29.370871] drm_ioctl_kernel+0x98/0xe8 [ 29.374690] drm_ioctl+0x26c/0x3d0 [ 29.378077] do_vfs_ioctl+0xafc/0xbb0 [ 29.381722] ksys_ioctl+0x64/0x98 [ 29.385022] __arm64_sys_ioctl+0x24/0x50 [ 29.388929] el0_svc_common.constprop.3+0xf0/0x158 [ 29.393697] el0_svc_handler+0x68/0x80 [ 29.397429] el0_sync_handler+0xa8/0x198 [ 29.401333] el0_sync+0x140/0x180 [ 29.404632] ---[ end trace 4ba9406972d937dd ]--- [ 29.409245] Unexpected kernel BRK exception at EL1
Signed-off-by: Zhihui Chen chenzhihui4@huawei.com
drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c | 1 + 1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c b/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c index 21b684eab5c9..19dc525a73b0 100644 --- a/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c +++ b/drivers/gpu/drm/hisilicon/hibmc/hibmc_ttm.c @@ -126,6 +126,7 @@ hibmc_framebuffer_init(struct drm_device *dev,
drm_helper_mode_fill_fb_struct(dev, &hibmc_fb->fb, mode_cmd); hibmc_fb->obj = obj;
- hibmc_fb->fb.obj[0] = obj; ret = drm_framebuffer_init(dev, &hibmc_fb->fb, &hibmc_fb_funcs); if (ret) { DRM_ERROR("drm_framebuffer_init failed: %d\n", ret);
Thanks for the patch. I have meanwhile merged
commit 3e10d2ffd245f996c36f87b0a8027a719e6135c3 Author: Thomas Zimmermann tzimmermann@suse.de Date: Tue Dec 3 09:38:16 2019 +0100
drm/hisilicon/hibmc: Replace struct hibmc_framebuffer with generic code
which replaces the code entirely.
Best regards Thomas
dri-devel@lists.freedesktop.org
-
Thomas Zimmermann -
Zhihui Chen