457e77b2 effectively replaces (... & 0xffffff00) << 8 with (... >> 8) << 8. Which does not do the same and breaks boot on my machine.
Restore the old behaviour and remove the unnecessary cast.
Signed-off-by: Andreas Noever andreas.noever@gmail.com --- drivers/gpu/drm/nouveau/core/subdev/bios/base.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c index e9df94f..2885b7c 100644 --- a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c +++ b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c @@ -109,7 +109,7 @@ nouveau_bios_shadow_pramin(struct nouveau_bios *bios) return; }
- addr = (u64)(addr >> 8) << 8; + addr = (addr >> 8) << 16; if (!addr) { addr = (u64)nv_rd32(bios, 0x001700) << 16; addr += 0xf0000;
Hello Andreas, after pulling and rebooting my machine this morning, nouveau was no longer working:
[ 6.455247] nouveau [ DEVICE][0000:02:00.0] BOOT0 : 0x0ac080b1 [ 6.455312] nouveau [ DEVICE][0000:02:00.0] Chipset: MCP79/MCP7A (NVAC) [ 6.455374] nouveau [ DEVICE][0000:02:00.0] Family : NV50 [ 6.456730] nouveau [ VBIOS][0000:02:00.0] checking PRAMIN for image... [ 6.456796] nouveau [ VBIOS][0000:02:00.0] ... signature not found [ 6.456858] nouveau [ VBIOS][0000:02:00.0] checking PROM for image... [ 6.471198] nouveau [ VBIOS][0000:02:00.0] ... signature not found [ 6.471265] nouveau [ VBIOS][0000:02:00.0] checking ACPI for image... [ 6.471328] nouveau [ VBIOS][0000:02:00.0] ... signature not found [ 6.471390] nouveau [ VBIOS][0000:02:00.0] checking PCIROM for image... [ 6.477478] nouveau [ VBIOS][0000:02:00.0] ... appears to be valid [ 6.477549] nouveau [ VBIOS][0000:02:00.0] using image from PCIROM [ 6.477789] nouveau [ VBIOS][0000:02:00.0] BIT signature found [ 6.477852] nouveau [ VBIOS][0000:02:00.0] version 62.79.4e.00.01 [ 6.486539] nouveau E[ VBIOS][0000:02:00.0] 0xd97c[ ]: unknown opcode 0x00 [ 6.486611] nouveau E[ DEVINIT][0000:02:00.0] init failed, -22 [ 6.486673] nouveau E[ DRM] failed to create 0x80000080, -22 [ 6.488470] nouveau: probe of 0000:02:00.0 failed with error -22
I bisected the problem:
# bad: [39de65aa2c3eee901db020a4f1396998e09602a3] Merge branch 'i2c/for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux # good: [b003d7706abc5d75cb58de0c9de8f1fc77e57008] Merge branch 'kbuild' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild git bisect start 'HEAD' 'v3.14-11011-gb003d77' 'drivers/gpu/drm/' # good: [e19b9137142988bec5a76c5f8bdf12a77ea802b0] Merge remote-tracking branch 'airlied/drm-next' into drm-intel-next git bisect good e19b9137142988bec5a76c5f8bdf12a77ea802b0 # bad: [60f2b4af1258c05e6b037af866be81abc24438f7] drm/i915: fix build warning on 32-bit (v2) git bisect bad 60f2b4af1258c05e6b037af866be81abc24438f7 # good: [f3381dfc9745bcd8b6be676ec4f68c52e71d24f1] drm/radeon/dp: use i2c_get_adapdata rather than casting git bisect good f3381dfc9745bcd8b6be676ec4f68c52e71d24f1 # good: [420b94697722512a2c0732970dc1530197a49adb] support for platform devices git bisect good 420b94697722512a2c0732970dc1530197a49adb # bad: [fc243d7f92d95d961186126efaad36197f133ab1] drm/nouveau/disp: limit dp capabilities as per dcb git bisect bad fc243d7f92d95d961186126efaad36197f133ab1 # bad: [88e98d49a1b3e0f8103cdd4fd80d576ec33133ab] drm/gf100-/gr: split ppc state into its subunits git bisect bad 88e98d49a1b3e0f8103cdd4fd80d576ec33133ab # bad: [e21bfd171a192dfba4a8907f2fcc41acac0f685f] drm/gf110/gr: fixup gpc/tpc initvals lists git bisect bad e21bfd171a192dfba4a8907f2fcc41acac0f685f # bad: [eeb0558e074215656ae11a170059a5f2ce29963f] drm/gf104/gr: rename gf104 (nvc4), it came before gf106 (nvc3) git bisect bad eeb0558e074215656ae11a170059a5f2ce29963f # bad: [6acc09b99d5d8f276a4f9bffc32f0bb0f939c7ca] drm/nvc0-/graph: fix gpccs fuc stack setup git bisect bad 6acc09b99d5d8f276a4f9bffc32f0bb0f939c7ca # bad: [457e77b26428ab4a24998eecfb99f27fa4195397] drm/nouveau/bios: add more checks to PRAMIN image fetching git bisect bad 457e77b26428ab4a24998eecfb99f27fa4195397
Than I saw your posting on LKML and tried your fix and your fix resolves my problem on top of Linus tip.
Tested-by: Thomas Glanzmann thomas@glanzmann.de
Cheers, Thomas
On Thu, Apr 10, 2014 at 5:33 AM, Andreas Noever andreas.noever@gmail.com wrote:
457e77b2 effectively replaces (... & 0xffffff00) << 8 with (... >> 8) << 8. Which does not do the same and breaks boot on my machine.
Restore the old behaviour and remove the unnecessary cast.
Signed-off-by: Andreas Noever andreas.noever@gmail.com
Signed-off-by: Ben Skeggs bskeggs@redhat.com
drivers/gpu/drm/nouveau/core/subdev/bios/base.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c index e9df94f..2885b7c 100644 --- a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c +++ b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c @@ -109,7 +109,7 @@ nouveau_bios_shadow_pramin(struct nouveau_bios *bios) return; }
addr = (u64)(addr >> 8) << 8;
addr = (addr >> 8) << 16; if (!addr) { addr = (u64)nv_rd32(bios, 0x001700) << 16; addr += 0xf0000;-- 1.9.1
dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
dri-devel@lists.freedesktop.org
-
Andreas Noever -
Ben Skeggs -
Thomas Glanzmann