diff options
| author | Ben Skeggs <bskeggs@redhat.com> | 2017-07-25 03:06:47 +0200 |
|---|---|---|
| committer | Ben Skeggs <bskeggs@redhat.com> | 2017-07-25 07:30:27 +0200 |
| commit | 38bcb208f60924a031b9f809f7cd252ea4a94e5f (patch) | |
| tree | 890007acc9d7f8a8b7ec2a923434a7845818a8c1 | |
| parent | drm/nouveau/disp/nv50-: bump max chans to 21 (diff) | |
| download | linux-38bcb208f60924a031b9f809f7cd252ea4a94e5f.tar.xz linux-38bcb208f60924a031b9f809f7cd252ea4a94e5f.zip | |
drm/nouveau/bar/gf100: fix access to upper half of BAR2
Bit 30 being set causes the upper half of BAR2 to stay in physical mode,
mapped over the end of VRAM, even when the rest of the BAR has been set
to virtual mode.
We inherited our initial value from RM, but I'm not aware of any reason
we need to keep it that way.
This fixes severe GPU hang/lockup issues revealed by Wayland on F26.
Shout-out to NVIDIA for the quick response with the potential cause!
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Cc: stable@vger.kernel.org # 4.3+
| -rw-r--r-- | drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c index c794b2c2d21e..6d8f21290aa2 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bar/gf100.c @@ -129,7 +129,7 @@ gf100_bar_init(struct nvkm_bar *base) if (bar->bar[0].mem) { addr = nvkm_memory_addr(bar->bar[0].mem) >> 12; - nvkm_wr32(device, 0x001714, 0xc0000000 | addr); + nvkm_wr32(device, 0x001714, 0x80000000 | addr); } return 0; |