similar to: [Bug 65554] New: CPU lock with nouveau_fan_update

From: Martin Peres <martin.peres at labri.fr> This should fix a deadlock that has been reported to us where fan_update() would hold the fan lock and try to grab the alarm_program_lock to reschedule an update. On an other CPU, the alarm_program_lock would have been taken before calling fan_update(), leading to a deadlock. We should Cc: <stable at vger.kernel.org> # 3.9+ Reported-by:

On Sun, Mar 9, 2014 at 10:51 AM, Marcin Slusarz <marcin.slusarz at gmail.com> wrote: > [ 326.168487] ====================================================== > [ 326.168491] [ INFO: possible circular locking dependency detected ] > [ 326.168496] 3.13.6 #1270 Not tainted > [ 326.168500] ------------------------------------------------------- > [ 326.168504] ldconfig/22297 is

Hello there, [linux-4.13/drivers/gpu/drm/nouveau/nvkm/subdev/therm/fan.c:93]: (warning) Opposite inner 'if' condition leads to a dead code block. Source code is if (target != duty) { u16 bump_period = fan->bios.bump_period; u16 slow_down_period = fan->bios.slow_down_period; u64 delay; if (duty > target) delay = slow_down_period;

From: Martin Peres <martin.peres at labri.fr> This patchset is mostly about fixing fdo bug #66177, reported by Dash Four. This bug is about fan/temp management not working after a suspend/resume cycle. Fan/therm management relies on ptimer's alarm feature to call periodically multiple callbacks that poll the temperature and update the fan speed if needed. The problem is that there is

The check for duty == target is always false because it is in an if block where duty != target. A previous change added the duty != target check and so the check duty == target check is now redundant and can be removed. Cleans up a cppcheck warning: drivers/gpu/drm/nouveau/nvkm/subdev/therm/fan.c:93:17: warning: Opposite inner 'if' condition leads to a dead code block.

Hi, I used to experience kernel panics caused by CPUx hard lockup once almost everyday. I'm running Ubuntu 14.10 on vanilla linux kernel 3.19.2 on Core i7-3770 with Gallium 0.4 on NV108. The panic log looked like: [ 9227.509744] ------------[ cut here ]------------ [ 9227.509750] WARNING: CPU: 0 PID: 0 at kernel/watchdog.c:290 watchdog_overflow_callback+0x92/0xc0() [ 9227.509751] Watchdog

From: Martin Peres <martin.peres at labri.fr> This should fix a deadlock that has been reported to us where fan_update() would hold the fan lock and try to grab the alarm_program_lock to reschedule an update. On an other CPU, the alarm_program_lock would have been taken before calling fan_update(), leading to a deadlock. Reported-by: Marcin Slusarz <marcin.slusarz at gmail.com>

Now it outputs: nouveau E[ PGRAPH][0000:02:00.0] PGRAPH TLB flush idle timeout fail nouveau E[ PGRAPH][0000:02:00.0] PGRAPH_STATUS: BUSY DISPATCH VFETCH CCACHE_UNK4 STRMOUT_GSCHED_UNK5 UNK14XX UNK1CXX CLIPID ZCULL ENG2D UNK34XX TPRAST TPROP ROP (0x011fde03) nouveau E[ PGRAPH][0000:02:00.0] PGRAPH_VSTATUS: CCACHE (0x00145b4d) (0x0000002d) ENG2D ROP (0x0034db40) instead of: [drm] nouveau

Hey, This patch causes a regression on my display-less nvd7. Commenting out the aux, aux_stat and aux_mask members in nvd0_i2c_oclass fixes boot, and makes things work normally again. broken dmesg: [ 40.314470] ACPI Warning: \_SB_.PCI0.GFX0._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20140424/nsarguments-95) [ 40.314729] ACPI Warning: \_SB_.PCI0.GFX0._DSM:

Now it outputs: nouveau E[ PGRAPH][0000:02:00.0] PGRAPH TLB flush idle timeout fail nouveau E[ PGRAPH][0000:02:00.0] PGRAPH_STATUS: BUSY DISPATCH VFETCH CCACHE_UNK4 STRMOUT_GSCHED_UNK5 UNK14XX UNK1CXX CLIPID ZCULL ENG2D UNK34XX TPRAST TPROP ROP (0x011fde03) nouveau E[ PGRAPH][0000:02:00.0] PGRAPH_VSTATUS_0: CCACHE (0x00145b4d) nouveau E[ PGRAPH][0000:02:00.0] PGRAPH_VSTATUS_1: (0x0000002d)

On Tue, 2018-06-05 at 10:05 -0700, Nick Desaulniers wrote: > Functions marked extern inline do not emit an externally visible > function when the gnu89 C standard is used. Some KBUILD Makefiles > overwrite KBUILD_CFLAGS. This is an issue for GCC 5.1+ users as without > an explicit C standard specified, the default is gnu11. Since c99, the > semantics of extern inline have changed

On Tue, 2018-06-05 at 10:05 -0700, Nick Desaulniers wrote: > Functions marked extern inline do not emit an externally visible > function when the gnu89 C standard is used. Some KBUILD Makefiles > overwrite KBUILD_CFLAGS. This is an issue for GCC 5.1+ users as without > an explicit C standard specified, the default is gnu11. Since c99, the > semantics of extern inline have changed

From: Martin Peres <martin.peres at labri.fr> Some vbioses have extra useless entries after "the end" of the table. This is problematic since all of the vbios I found with this issue redefine the pwm freq divider to insane levels (52750 Hz instead of 2500), thus breaking fan management. The first solution to solve this mess would be to change the length of the table. The solution

Signed-off-by: Marcin Slusarz <marcin.slusarz at gmail.com> --- drivers/gpu/drm/nouveau/nv50_graph.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nv50_graph.c b/drivers/gpu/drm/nouveau/nv50_graph.c index 6899547..a61853f 100644 --- a/drivers/gpu/drm/nouveau/nv50_graph.c +++ b/drivers/gpu/drm/nouveau/nv50_graph.c @@ -435,6 +435,11 @@

On Mon, Mar 24, 2014 at 6:42 PM, Alexandre Courbot <acourbot at nvidia.com> wrote: > Set the correct subdev/engine classes when GK20A (0xea) is probed. > > Signed-off-by: Alexandre Courbot <acourbot at nvidia.com> > --- > drivers/gpu/drm/nouveau/core/engine/device/nve0.c | 20 ++++++++++++++++++++ > 1 file changed, 20 insertions(+) > > diff --git

Overall idea: Detect lockups by watching for timeouts (vm flush / fence), return -EIOs, handle them at ioctl level, reset the GPU and repeat last ioctl. GPU reset is done by doing suspend / resume cycle with few tweaks: - CPU-only bo eviction - ignoring vm flush / fence timeouts - shortening waits Signed-off-by: Marcin Slusarz <marcin.slusarz at gmail.com> ---

https://bugs.freedesktop.org/show_bug.cgi?id=67161 Priority: medium Bug ID: 67161 Assignee: nouveau at lists.freedesktop.org Summary: Blank video after resuming from S3 or S4 QA Contact: xorg-team at lists.x.org Severity: normal Classification: Unclassified OS: Linux (All) Reporter: mauromol at tiscali.it

From: Martin Peres <martin.peres at labri.fr> Re-use the therm-exported fan structure with only two minor modifications: - pwm_freq: u16 -> u32; - add fan_type (toggle or PWM) Signed-off-by: Martin Peres <martin.peres at free.fr> --- drm/Kbuild | 1 + drm/core/include/subdev/bios/fan.h | 1 + drm/core/subdev/bios/fan.c | 1 +

System was booted with PCI graphics card first in VGA text mode. Loading nouveau from there on causes the following BUG, after what kernel produces trace over trace until overflowing its stack. (trace captured via netconsole) [ 154.323717] wmi: Mapper loaded [ 154.735793] nouveau 0000:02:00.0: enabling device (0004 -> 0006) [ 154.743189] ACPI: PCI Interrupt Link [LNK5] enabled at IRQ 16 [

Signed-off-by: Marcin Ko?cielnicki <koriakin at 0x04.net> --- drivers/gpu/drm/nouveau/nv04_graph.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nv04_graph.c b/drivers/gpu/drm/nouveau/nv04_graph.c index 396ee92..d561d77 100644 --- a/drivers/gpu/drm/nouveau/nv04_graph.c +++ b/drivers/gpu/drm/nouveau/nv04_graph.c @@ -543,7 +543,7 @@