[PATCH] drm/amdgpu: fix modprobe failure of the secondary GPU when GDDR6 training enabled(V4)
Alex Deucher
alexdeucher at gmail.com
Fri Jan 10 15:21:24 UTC 2020
On Fri, Jan 10, 2020 at 12:30 AM Tianci Yin <tianci.yin at amd.com> wrote:
>
> From: "Tianci.Yin" <tianci.yin at amd.com>
>
> [why]
> In dual GPUs scenario, stolen_size is assigned to zero on the secondary GPU,
> since there is no pre-OS console using that memory. Then the bottom region of
> VRAM was allocated as GTT, unfortunately a small region of bottom VRAM was
> encroached by UMC firmware during GDDR6 BIST training, this cause page fault.
>
> [how]
> Forcing stolen_size to 3MB, then the bottom region of VRAM was
> allocated as stolen memory, GTT corruption avoid.
>
> Change-Id: I310a72ba0402994defbe50839842a8edb025a868
> Signed-off-by: Tianci.Yin <tianci.yin at amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h | 5 +++++
> drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c | 27 ++++++++++++++++++++++++-
> 2 files changed, 31 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
> index c91dd602d5f1..e4b2f9bcaeb7 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
> @@ -60,6 +60,11 @@
> */
> #define AMDGPU_GMC_FAULT_TIMEOUT 5000ULL
>
> +/*
> + * Default stolen memory size, 1024 * 768 * 4
> + */
> +#define AMDGPU_STOLEN_VGA_DEFAULT_SIZE 0x300000ULL
It would be better to call this
AMDGPU_STOLEN_BIST_TRAINING_DEFAULT_SIZE. Also, I sounds like we are
probably going to have to assume that all of vram could be touched and
adjust our sequnce appropriately. I think we are just getting lucky
with this size because most imporant kernel structures are in the
lower area of vram.
Alex
> +
> struct firmware;
>
> /*
> diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
> index 5ad89bb6f3ba..14961f1ebfab 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
> @@ -566,7 +566,12 @@ static int gmc_v10_0_late_init(void *handle)
> struct amdgpu_device *adev = (struct amdgpu_device *)handle;
> int r;
>
> - amdgpu_bo_late_init(adev);
> + /*
> + * Can't free the stolen VGA memory when it might be used for memory
> + * training again.
> + */
> + if (!adev->fw_vram_usage.mem_train_support)
> + amdgpu_bo_late_init(adev);
>
> r = amdgpu_gmc_allocate_vm_inv_eng(adev);
> if (r)
> @@ -750,6 +755,19 @@ static int gmc_v10_0_sw_init(void *handle)
>
> adev->gmc.stolen_size = gmc_v10_0_get_vbios_fb_size(adev);
>
> + /*
> + * In dual GPUs scenario, stolen_size is assigned to zero on the
> + * secondary GPU, since there is no pre-OS console using that memory.
> + * Then the bottom region of VRAM was allocated as GTT, unfortunately a
> + * small region of bottom VRAM was encroached by UMC firmware during
> + * GDDR6 BIST training, this cause page fault.
> + * The page fault can be fixed by forcing stolen_size to 3MB, then the
> + * bottom region of VRAM was allocated as stolen memory, GTT corruption
> + * avoid.
> + */
> + adev->gmc.stolen_size = max(adev->gmc.stolen_size,
> + AMDGPU_STOLEN_VGA_DEFAULT_SIZE);
> +
> /* Memory manager */
> r = amdgpu_bo_init(adev);
> if (r)
> @@ -789,6 +807,13 @@ static void gmc_v10_0_gart_fini(struct amdgpu_device *adev)
> static int gmc_v10_0_sw_fini(void *handle)
> {
> struct amdgpu_device *adev = (struct amdgpu_device *)handle;
> + void *stolen_vga_buf;
> +
> + /*
> + * Free the stolen memory if it wasn't already freed in late_init
> + * because of memory training.
> + */
> + amdgpu_bo_free_kernel(&adev->stolen_vga_memory, NULL, &stolen_vga_buf);
>
> amdgpu_vm_manager_fini(adev);
> gmc_v10_0_gart_fini(adev);
> --
> 2.17.1
>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
More information about the amd-gfx
mailing list