Re: [PATCH] kexec: Fix reboot race during device_shutdown()

[Eric W. Biederman]

"Joel Fernandes (Google)" <joel@xxxxxxxxxxxxxxxxx> writes:

> During kexec reboot, it is possible for a race to occur between
> device_shutdown() and userspace.  This causes accesses to GPU after pm_runtime
> suspend has already happened. Fix this by calling freeze_processes() before
> device_shutdown().

Is there any reason why this same race with between sys_kexec and the
adreno_ioctl can not happen during a normal reboot?

Is there any reason why there is not a .shutdown method to prevent the
race?

I would think the thing to do is to prevent this race in
kernel_restart_prepare or in the GPUs .shutdown method.  As I don't see
anything that would prevent this during a normal reboot.

>
> Such freezing is already being done if kernel supports KEXEC_JUMP and
> kexec_image->preserve_context is true. However, doing it if either of these are
> not true prevents crashes/races.

The KEXEC_JUMP case is something else entirely.  It is supposed to work
like suspend to RAM.  Maybe reboot should as well, but I am
uncomfortable making a generic device fix kexec specific.


> This fixes the following crash during kexec reboot on an ARM64 device
> with adreno GPU:
>
> [  292.534314] Kernel panic - not syncing: Asynchronous SError Interrupt
> [  292.534323] Hardware name: Google Lazor (rev3 - 8) with LTE (DT)
> [  292.534326] Call trace:
> [  292.534328]  dump_backtrace+0x0/0x1d4
> [  292.534337]  show_stack+0x20/0x2c
> [  292.534342]  dump_stack_lvl+0x60/0x78
> [  292.534347]  dump_stack+0x18/0x38
> [  292.534352]  panic+0x148/0x3b0
> [  292.534357]  nmi_panic+0x80/0x94
> [  292.534364]  arm64_serror_panic+0x70/0x7c
> [  292.534369]  do_serror+0x0/0x7c
> [  292.534372]  do_serror+0x54/0x7c
> [  292.534377]  el1h_64_error_handler+0x34/0x4c
> [  292.534381]  el1h_64_error+0x7c/0x80
> [  292.534386]  el1_interrupt+0x20/0x58
> [  292.534389]  el1h_64_irq_handler+0x18/0x24
> [  292.534395]  el1h_64_irq+0x7c/0x80
> [  292.534399]  local_daif_inherit+0x10/0x18
> [  292.534405]  el1h_64_sync_handler+0x48/0xb4
> [  292.534410]  el1h_64_sync+0x7c/0x80
> [  292.534414]  a6xx_gmu_set_oob+0xbc/0x1fc
> [  292.534422]  a6xx_get_timestamp+0x40/0xb4
> [  292.534426]  adreno_get_param+0x12c/0x1e0
> [  292.534433]  msm_ioctl_get_param+0x64/0x70
> [  292.534440]  drm_ioctl_kernel+0xe8/0x158
> [  292.534448]  drm_ioctl+0x208/0x320
> [  292.534453]  __arm64_sys_ioctl+0x98/0xd0
> [  292.534461]  invoke_syscall+0x4c/0x118
>
> Cc: Steven Rostedt <rostedt@xxxxxxxxxxx>
> Cc: Ricardo Ribalda <ribalda@xxxxxxxxxx>
> Cc: Ross Zwisler <zwisler@xxxxxxxxxx>
> Cc: Rob Clark <robdclark@xxxxxxxxx>
> Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
> Tested-by: Ricardo Ribalda <ribalda@xxxxxxxxxx>
> Signed-off-by: Joel Fernandes (Google) <joel@xxxxxxxxxxxxxxxxx>
> ---
>  kernel/kexec_core.c | 6 ++++++
>  1 file changed, 6 insertions(+)
>
> diff --git a/kernel/kexec_core.c b/kernel/kexec_core.c
> index e2f2574d8b74..6599f485e42d 100644
> --- a/kernel/kexec_core.c
> +++ b/kernel/kexec_core.c
> @@ -1299,6 +1299,12 @@ int kernel_kexec(void)
>  	} else
>  #endif
>  	{
> +		error = freeze_processes();
> +		if (error) {
> +			error = -EBUSY;
> +			goto Unlock;
> +		}
> +
>  		kexec_in_progress = true;
>  		kernel_restart_prepare("kexec reboot");
>  		migrate_to_reboot_cpu();

Eric