Re: [RFC PATCH v2 10/17] WIP: gpu: host1x: Add no-recovery mode

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



05.09.2020 13:34, Mikko Perttunen пишет:
> +	} else {
> +		struct host1x_job *failed_job = job;
> +
> +		host1x_job_dump(dev, job);
> +
> +		host1x_syncpt_set_locked(job->syncpt);
> +		failed_job->cancelled = true;
> +
> +		list_for_each_entry_continue(job, &cdma->sync_queue, list) {
> +			unsigned int i;
> +
> +			if (job->syncpt != failed_job->syncpt)
> +				continue;
> +
> +			for (i = 0; i < job->num_slots; i++) {
> +				unsigned int slot = (job->first_get/8 + i) %
> +						    HOST1X_PUSHBUFFER_SLOTS;
> +				u32 *mapped = cdma->push_buffer.mapped;
> +
> +				mapped[2*slot+0] = 0x1bad0000;
> +				mapped[2*slot+1] = 0x1bad0000;

The 0x1bad0000 is a valid memory address on Tegra20.

The 0x60000000 is invalid phys address for all hardware generations.
It's used by grate-kernel [1] and VDE driver [2]. Note that the 0x6 <<
28 is also invalid Host1x opcode, while 0x1 should break CDMA parser
during of PB debug-dumping.

[1]
https://github.com/grate-driver/linux/blob/master/drivers/gpu/drm/tegra/gem.h#L16

[2]
https://elixir.bootlin.com/linux/v5.9-rc4/source/drivers/staging/media/tegra-vde/iommu.c#L99

The VDE driver reserves the trapping IOVA addresses, I assume the Host1x
driver should do the same.



[Index of Archives]     [ARM Kernel]     [Linux ARM]     [Linux ARM MSM]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux