[PATCH AUTOSEL 5.6 21/50] drm/amd/display: blank dp stream before re-train the link

[PATCH AUTOSEL 5.6 21/50] drm/amd/display: blank dp stream before re-train the link

[Sasha Levin]

From: Xiaodong Yan <Xiaodong.Yan@amd.com>

[ Upstream commit 718a5569b6fa6e1f49f1ae76a3c18acb4ddb74f1 ]

[Why]
When link loss happened, monitor can not light up if only re-train the
link.

[How]
Blank all the DP streams on this link before re-train the link, and then
unblank the stream

Signed-off-by: Xiaodong Yan <Xiaodong.Yan@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
index fd9e69634c50a..1b6c75a4dd60a 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
@@ -2885,6 +2885,12 @@ bool dc_link_handle_hpd_rx_irq(struct dc_link *link, union hpd_irq_data *out_hpd
 					sizeof(hpd_irq_dpcd_data),
 					"Status: ");
 
+		for (i = 0; i < MAX_PIPES; i++) {
+			pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i];
+			if (pipe_ctx && pipe_ctx->stream && pipe_ctx->stream->link == link)
+				link->dc->hwss.blank_stream(pipe_ctx);
+		}
+
 		for (i = 0; i < MAX_PIPES; i++) {
 			pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i];
 			if (pipe_ctx && pipe_ctx->stream && pipe_ctx->stream->link == link)
@@ -2904,6 +2910,12 @@ bool dc_link_handle_hpd_rx_irq(struct dc_link *link, union hpd_irq_data *out_hpd
 		if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST)
 			dc_link_reallocate_mst_payload(link);
 
+		for (i = 0; i < MAX_PIPES; i++) {
+			pipe_ctx = &link->dc->current_state->res_ctx.pipe_ctx[i];
+			if (pipe_ctx && pipe_ctx->stream && pipe_ctx->stream->link == link)
+				link->dc->hwss.unblank_stream(pipe_ctx, &previous_link_settings);
+		}
+
 		status = false;
 		if (out_link_loss)
 			*out_link_loss = true;
-- 
2.20.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH AUTOSEL 5.6 25/50] dma-buf: Fix SET_NAME ioctl uapi

[Sasha Levin]

From: Daniel Vetter <daniel.vetter@intel.com>

[ Upstream commit a5bff92eaac45bdf6221badf9505c26792fdf99e ]

The uapi is the same on 32 and 64 bit, but the number isn't. Everyone
who botched this please re-read:

https://www.kernel.org/doc/html/v5.4-preprc-cpu/ioctl/botching-up-ioctls.html

Also, the type argument for the ioctl macros is for the type the void
__user *arg pointer points at, which in this case would be the
variable-sized char[] of a 0 terminated string. So this was botched in
more than just the usual ways.

Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: Chenbo Feng <fengc@google.com>
Cc: Greg Hackmann <ghackmann@google.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Cc: minchan@kernel.org
Cc: surenb@google.com
Cc: jenhaochen@google.com
Cc: Martin Liu <liumartin@google.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Tested-by: Martin Liu <liumartin@google.com>
Reviewed-by: Martin Liu <liumartin@google.com>
Signed-off-by: Sumit Semwal <sumit.semwal@linaro.org>
  [sumits: updated some checkpatch fixes, corrected author email]
Link: https://patchwork.freedesktop.org/patch/msgid/20200407133002.3486387-1-daniel.vetter@ffwll.ch
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/dma-buf/dma-buf.c    | 3 ++-
 include/uapi/linux/dma-buf.h | 6 ++++++
 2 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c
index c343c7c10b4cc..e7589d91de8fb 100644
--- a/drivers/dma-buf/dma-buf.c
+++ b/drivers/dma-buf/dma-buf.c
@@ -388,7 +388,8 @@ static long dma_buf_ioctl(struct file *file,
 
 		return ret;
 
-	case DMA_BUF_SET_NAME:
+	case DMA_BUF_SET_NAME_A:
+	case DMA_BUF_SET_NAME_B:
 		return dma_buf_set_name(dmabuf, (const char __user *)arg);
 
 	default:
diff --git a/include/uapi/linux/dma-buf.h b/include/uapi/linux/dma-buf.h
index dbc7092e04b5a..7f30393b92c3b 100644
--- a/include/uapi/linux/dma-buf.h
+++ b/include/uapi/linux/dma-buf.h
@@ -39,6 +39,12 @@ struct dma_buf_sync {
 
 #define DMA_BUF_BASE		'b'
 #define DMA_BUF_IOCTL_SYNC	_IOW(DMA_BUF_BASE, 0, struct dma_buf_sync)
+
+/* 32/64bitness of this uapi was botched in android, there's no difference
+ * between them in actual uapi, they're just different numbers.
+ */
 #define DMA_BUF_SET_NAME	_IOW(DMA_BUF_BASE, 1, const char *)
+#define DMA_BUF_SET_NAME_A	_IOW(DMA_BUF_BASE, 1, u32)
+#define DMA_BUF_SET_NAME_B	_IOW(DMA_BUF_BASE, 1, u64)
 
 #endif
-- 
2.20.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH AUTOSEL 5.6 32/50] drm/amd/powerplay: avoid using pm_en before it is initialized revised

[Sasha Levin]

From: Tiecheng Zhou <Tiecheng.Zhou@amd.com>

[ Upstream commit 690ae30be163d5262feae01335b2a6f30569e5aa ]

hwmgr->pm_en is initialized at hwmgr_hw_init.

during amdgpu_device_init, there is amdgpu_asic_reset that calls to
soc15_asic_reset (for V320 usecase, Vega10 asic), in which:
1) soc15_asic_reset_method calls to pp_get_asic_baco_capability (pm_en)
2) soc15_asic_baco_reset calls to pp_set_asic_baco_state (pm_en)

pm_en is used in the above two cases while it has not yet been initialized

So avoid using pm_en in the above two functions for V320 passthrough.

Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Tiecheng Zhou <Tiecheng.Zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/powerplay/amd_powerplay.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c
index c195575366a3b..e4e5a53b2b4ea 100644
--- a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c
+++ b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c
@@ -1435,7 +1435,8 @@ static int pp_get_asic_baco_capability(void *handle, bool *cap)
 	if (!hwmgr)
 		return -EINVAL;
 
-	if (!hwmgr->pm_en || !hwmgr->hwmgr_func->get_asic_baco_capability)
+	if (!(hwmgr->not_vf && amdgpu_dpm) ||
+		!hwmgr->hwmgr_func->get_asic_baco_capability)
 		return 0;
 
 	mutex_lock(&hwmgr->smu_lock);
@@ -1469,7 +1470,8 @@ static int pp_set_asic_baco_state(void *handle, int state)
 	if (!hwmgr)
 		return -EINVAL;
 
-	if (!hwmgr->pm_en || !hwmgr->hwmgr_func->set_asic_baco_state)
+	if (!(hwmgr->not_vf && amdgpu_dpm) ||
+		!hwmgr->hwmgr_func->set_asic_baco_state)
 		return 0;
 
 	mutex_lock(&hwmgr->smu_lock);
-- 
2.20.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH AUTOSEL 5.6 33/50] drm/amdgpu: bump version for invalidate L2 before SDMA IBs

[Sasha Levin]

From: Marek Olšák <marek.olsak@amd.com>

[ Upstream commit 9017a4897a20658f010bebea825262963c10afa6 ]

This fixes GPU hangs due to cache coherency issues.
Bump the driver version. Split out from the original patch.

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 42f4febe24c6d..8d45a2b662aeb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -85,9 +85,10 @@
  * - 3.34.0 - Non-DC can flip correctly between buffers with different pitches
  * - 3.35.0 - Add drm_amdgpu_info_device::tcc_disabled_mask
  * - 3.36.0 - Allow reading more status registers on si/cik
+ * - 3.37.0 - L2 is invalidated before SDMA IBs, needed for correctness
  */
 #define KMS_DRIVER_MAJOR	3
-#define KMS_DRIVER_MINOR	36
+#define KMS_DRIVER_MINOR	37
 #define KMS_DRIVER_PATCHLEVEL	0
 
 int amdgpu_vram_limit = 0;
-- 
2.20.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH AUTOSEL 5.6 35/50] drm/amd/display: check if REFCLK_CNTL register is present

[Sasha Levin]

From: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>

[ Upstream commit 3159d41db3a04330c31ece32f8b29752fc114848 ]

Check before programming the register since it isn't present on
all IPs using this code.

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Eric Bernstein <Eric.Bernstein@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
index a444fed941849..ad422e00f9fec 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c
@@ -2306,7 +2306,8 @@ void dcn20_fpga_init_hw(struct dc *dc)
 
 	REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_REFDIV, 2);
 	REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1);
-	REG_WRITE(REFCLK_CNTL, 0);
+	if (REG(REFCLK_CNTL))
+		REG_WRITE(REFCLK_CNTL, 0);
 	//
 
 
-- 
2.20.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH AUTOSEL 5.6 36/50] drm/amd/display: Defer cursor update around VUPDATE for all ASIC

[Sasha Levin]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[ Upstream commit fdfd2a858590d318cfee483bd1c73e00f77533af ]

[Why]
Fixes the following scenario:

- Flip has been prepared sometime during the frame, update pending
- Cursor update happens right when VUPDATE would happen
- OPTC lock acquired, VUPDATE is blocked until next frame
- Flip is delayed potentially infinitely

With the igt@kms_cursor_legacy cursor-vs-flip-legacy test we can
observe nearly *13* frames of delay for some flips on Navi.

[How]
Apply the Raven workaround generically. When close enough to VUPDATE
block cursor updates from occurring from the dc_stream_set_cursor_*
helpers.

This could perhaps be a little smarter by checking if there were
pending updates or flips earlier in the frame on the HUBP side before
applying the delay, but this should be fine for now.

This fixes the kms_cursor_legacy test.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../gpu/drm/amd/display/dc/core/dc_stream.c   | 28 +++++++++----------
 1 file changed, 14 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
index 6ddbb00ed37a5..8c20e9e907b2f 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
@@ -239,24 +239,24 @@ static void delay_cursor_until_vupdate(struct pipe_ctx *pipe_ctx, struct dc *dc)
 	struct dc_stream_state *stream = pipe_ctx->stream;
 	unsigned int us_per_line;
 
-	if (stream->ctx->asic_id.chip_family == FAMILY_RV &&
-			ASICREV_IS_RAVEN(stream->ctx->asic_id.hw_internal_rev)) {
+	if (!dc->hwss.get_vupdate_offset_from_vsync)
+		return;
 
-		vupdate_line = dc->hwss.get_vupdate_offset_from_vsync(pipe_ctx);
-		if (!dc_stream_get_crtc_position(dc, &stream, 1, &vpos, &nvpos))
-			return;
+	vupdate_line = dc->hwss.get_vupdate_offset_from_vsync(pipe_ctx);
+	if (!dc_stream_get_crtc_position(dc, &stream, 1, &vpos, &nvpos))
+		return;
 
-		if (vpos >= vupdate_line)
-			return;
+	if (vpos >= vupdate_line)
+		return;
 
-		us_per_line = stream->timing.h_total * 10000 / stream->timing.pix_clk_100hz;
-		lines_to_vupdate = vupdate_line - vpos;
-		us_to_vupdate = lines_to_vupdate * us_per_line;
+	us_per_line =
+		stream->timing.h_total * 10000 / stream->timing.pix_clk_100hz;
+	lines_to_vupdate = vupdate_line - vpos;
+	us_to_vupdate = lines_to_vupdate * us_per_line;
 
-		/* 70 us is a conservative estimate of cursor update time*/
-		if (us_to_vupdate < 70)
-			udelay(us_to_vupdate);
-	}
+	/* 70 us is a conservative estimate of cursor update time*/
+	if (us_to_vupdate < 70)
+		udelay(us_to_vupdate);
 #endif
 }
 
-- 
2.20.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH AUTOSEL 5.6 37/50] drm/amd/display: Update downspread percent to match spreadsheet for DCN2.1

[Sasha Levin]

From: Sung Lee <sung.lee@amd.com>

[ Upstream commit 668a6741f809f2d15d125cfe2b39661e8f1655ea ]

[WHY]
The downspread percentage was copied over from a previous version
of the display_mode_lib spreadsheet. This value has been updated,
and the previous value is too high to allow for such modes as
4K120hz. The new value is sufficient for such modes.

[HOW]
Update the value in dcn21_resource to match the spreadsheet.

Signed-off-by: Sung Lee <sung.lee@amd.com>
Reviewed-by: Yongqiang Sun <yongqiang.sun@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
index 33d0a176841a5..122d3e734c59a 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
@@ -250,7 +250,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn2_1_soc = {
 	.dram_channel_width_bytes = 4,
 	.fabric_datapath_to_dcn_data_return_bytes = 32,
 	.dcn_downspread_percent = 0.5,
-	.downspread_percent = 0.5,
+	.downspread_percent = 0.38,
 	.dram_page_open_time_ns = 50.0,
 	.dram_rw_turnaround_time_ns = 17.5,
 	.dram_return_buffer_per_channel_bytes = 8192,
-- 
2.20.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH AUTOSEL 5.6 39/50] drm/qxl: lost qxl_bo_kunmap_atomic_page in qxl_image_init_helper()

[Sasha Levin]

From: Vasily Averin <vvs@virtuozzo.com>

[ Upstream commit 5b5703dbafae74adfbe298a56a81694172caf5e6 ]

v2: removed TODO reminder

Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Link: http://patchwork.freedesktop.org/patch/msgid/a4e0ae09-a73c-1c62-04ef-3f990d41bea9@virtuozzo.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/qxl/qxl_image.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/qxl/qxl_image.c b/drivers/gpu/drm/qxl/qxl_image.c
index 43688ecdd8a04..60ab7151b84dc 100644
--- a/drivers/gpu/drm/qxl/qxl_image.c
+++ b/drivers/gpu/drm/qxl/qxl_image.c
@@ -212,7 +212,8 @@ qxl_image_init_helper(struct qxl_device *qdev,
 		break;
 	default:
 		DRM_ERROR("unsupported image bit depth\n");
-		return -EINVAL; /* TODO: cleanup */
+		qxl_bo_kunmap_atomic_page(qdev, image_bo, ptr);
+		return -EINVAL;
 	}
 	image->u.bitmap.flags = QXL_BITMAP_TOP_DOWN;
 	image->u.bitmap.x = width;
-- 
2.20.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH AUTOSEL 5.6 33/50] drm/amdgpu: bump version for invalidate L2 before SDMA IBs

[Michel Dänzer]

On 2020-05-07 4:27 p.m., Sasha Levin wrote:
> From: Marek Olšák <marek.olsak@amd.com>
> 
> [ Upstream commit 9017a4897a20658f010bebea825262963c10afa6 ]
> 
> This fixes GPU hangs due to cache coherency issues.
> Bump the driver version. Split out from the original patch.
> 
> Signed-off-by: Marek Olšák <marek.olsak@amd.com>
> Reviewed-by: Christian König <christian.koenig@amd.com>
> Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
>  drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> index 42f4febe24c6d..8d45a2b662aeb 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> @@ -85,9 +85,10 @@
>   * - 3.34.0 - Non-DC can flip correctly between buffers with different pitches
>   * - 3.35.0 - Add drm_amdgpu_info_device::tcc_disabled_mask
>   * - 3.36.0 - Allow reading more status registers on si/cik
> + * - 3.37.0 - L2 is invalidated before SDMA IBs, needed for correctness
>   */
>  #define KMS_DRIVER_MAJOR	3
> -#define KMS_DRIVER_MINOR	36
> +#define KMS_DRIVER_MINOR	37
>  #define KMS_DRIVER_PATCHLEVEL	0
>  
>  int amdgpu_vram_limit = 0;
> 

This requires the parent commit fdf83646c0542ecfb9adc4db8f741a1f43dca058
"drm/amdgpu: invalidate L2 before SDMA IBs (v2)". KMS_DRIVER_MINOR is
bumped to signal to userspace the fix in that commit is present.


-- 
Earthling Michel Dänzer               |               https://redhat.com
Libre software enthusiast             |             Mesa and X developer
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH AUTOSEL 5.6 33/50] drm/amdgpu: bump version for invalidate L2 before SDMA IBs

[Sasha Levin]

On Thu, May 07, 2020 at 06:11:17PM +0200, Michel Dänzer wrote:
>On 2020-05-07 4:27 p.m., Sasha Levin wrote:
>> From: Marek Olšák <marek.olsak@amd.com>
>>
>> [ Upstream commit 9017a4897a20658f010bebea825262963c10afa6 ]
>>
>> This fixes GPU hangs due to cache coherency issues.
>> Bump the driver version. Split out from the original patch.
>>
>> Signed-off-by: Marek Olšák <marek.olsak@amd.com>
>> Reviewed-by: Christian König <christian.koenig@amd.com>
>> Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
>> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
>> Signed-off-by: Sasha Levin <sashal@kernel.org>
>> ---
>>  drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++-
>>  1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>> index 42f4febe24c6d..8d45a2b662aeb 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>> @@ -85,9 +85,10 @@
>>   * - 3.34.0 - Non-DC can flip correctly between buffers with different pitches
>>   * - 3.35.0 - Add drm_amdgpu_info_device::tcc_disabled_mask
>>   * - 3.36.0 - Allow reading more status registers on si/cik
>> + * - 3.37.0 - L2 is invalidated before SDMA IBs, needed for correctness
>>   */
>>  #define KMS_DRIVER_MAJOR	3
>> -#define KMS_DRIVER_MINOR	36
>> +#define KMS_DRIVER_MINOR	37
>>  #define KMS_DRIVER_PATCHLEVEL	0
>>
>>  int amdgpu_vram_limit = 0;
>>
>
>This requires the parent commit fdf83646c0542ecfb9adc4db8f741a1f43dca058
>"drm/amdgpu: invalidate L2 before SDMA IBs (v2)". KMS_DRIVER_MINOR is
>bumped to signal to userspace the fix in that commit is present.

I've grabbed the commit you've pointed out as well as ce73516d42c9
("drm/amdgpu: simplify padding calculations (v2)") to make the backport
apply cleanly, thank you!

-- 
Thanks,
Sasha
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH AUTOSEL 5.6 33/50] drm/amdgpu: bump version for invalidate L2 before SDMA IBs

[Marek Olšák]

[-- Attachment #1.1: Type: text/plain, Size: 1843 bytes --]

Hi Sasha,

I disagree with this. Bumping the driver version will have implications on
other new features, because it's like an ABI barrier exposing new
functionality.

Marek

On Thu, May 7, 2020 at 10:28 AM Sasha Levin <sashal@kernel.org> wrote:

> From: Marek Olšák <marek.olsak@amd.com>
>
> [ Upstream commit 9017a4897a20658f010bebea825262963c10afa6 ]
>
> This fixes GPU hangs due to cache coherency issues.
> Bump the driver version. Split out from the original patch.
>
> Signed-off-by: Marek Olšák <marek.olsak@amd.com>
> Reviewed-by: Christian König <christian.koenig@amd.com>
> Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
>  drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> index 42f4febe24c6d..8d45a2b662aeb 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> @@ -85,9 +85,10 @@
>   * - 3.34.0 - Non-DC can flip correctly between buffers with different
> pitches
>   * - 3.35.0 - Add drm_amdgpu_info_device::tcc_disabled_mask
>   * - 3.36.0 - Allow reading more status registers on si/cik
> + * - 3.37.0 - L2 is invalidated before SDMA IBs, needed for correctness
>   */
>  #define KMS_DRIVER_MAJOR       3
> -#define KMS_DRIVER_MINOR       36
> +#define KMS_DRIVER_MINOR       37
>  #define KMS_DRIVER_PATCHLEVEL  0
>
>  int amdgpu_vram_limit = 0;
> --
> 2.20.1
>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
>

[-- Attachment #1.2: Type: text/html, Size: 3287 bytes --]

[-- Attachment #2: Type: text/plain, Size: 160 bytes --]

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH AUTOSEL 5.6 33/50] drm/amdgpu: bump version for invalidate L2 before SDMA IBs

[Deucher, Alexander]

[-- Attachment #1.1: Type: text/plain, Size: 3056 bytes --]

[AMD Public Use]

Yes, please drop this patch for stable.

Alex
________________________________
From: Marek Olšák <maraeo@gmail.com>
Sent: Monday, May 18, 2020 6:35 PM
To: Sasha Levin <sashal@kernel.org>
Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>; stable@vger.kernel.org <stable@vger.kernel.org>; Pelloux-prayer, Pierre-eric <Pierre-eric.Pelloux-prayer@amd.com>; Olsak, Marek <Marek.Olsak@amd.com>; amd-gfx mailing list <amd-gfx@lists.freedesktop.org>; dri-devel <dri-devel@lists.freedesktop.org>; Deucher, Alexander <Alexander.Deucher@amd.com>; Koenig, Christian <Christian.Koenig@amd.com>
Subject: Re: [PATCH AUTOSEL 5.6 33/50] drm/amdgpu: bump version for invalidate L2 before SDMA IBs

Hi Sasha,

I disagree with this. Bumping the driver version will have implications on other new features, because it's like an ABI barrier exposing new functionality.

Marek

On Thu, May 7, 2020 at 10:28 AM Sasha Levin <sashal@kernel.org<mailto:sashal@kernel.org>> wrote:
From: Marek Olšák <marek.olsak@amd.com<mailto:marek.olsak@amd.com>>

[ Upstream commit 9017a4897a20658f010bebea825262963c10afa6 ]

This fixes GPU hangs due to cache coherency issues.
Bump the driver version. Split out from the original patch.

Signed-off-by: Marek Olšák <marek.olsak@amd.com<mailto:marek.olsak@amd.com>>
Reviewed-by: Christian König <christian.koenig@amd.com<mailto:christian.koenig@amd.com>>
Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com<mailto:pierre-eric.pelloux-prayer@amd.com>>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com<mailto:alexander.deucher@amd.com>>
Signed-off-by: Sasha Levin <sashal@kernel.org<mailto:sashal@kernel.org>>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 42f4febe24c6d..8d45a2b662aeb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -85,9 +85,10 @@
  * - 3.34.0 - Non-DC can flip correctly between buffers with different pitches
  * - 3.35.0 - Add drm_amdgpu_info_device::tcc_disabled_mask
  * - 3.36.0 - Allow reading more status registers on si/cik
+ * - 3.37.0 - L2 is invalidated before SDMA IBs, needed for correctness
  */
 #define KMS_DRIVER_MAJOR       3
-#define KMS_DRIVER_MINOR       36
+#define KMS_DRIVER_MINOR       37
 #define KMS_DRIVER_PATCHLEVEL  0

 int amdgpu_vram_limit = 0;
--
2.20.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
https://lists.freedesktop.org/mailman/listinfo/amd-gfx<https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.freedesktop.org%2Fmailman%2Flistinfo%2Famd-gfx&data=02%7C01%7Calexander.deucher%40amd.com%7C2bb4939b162b4d66d57708d7fb7bd608%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637254381604715852&sdata=uJEI0CDUeUbUGUgnhnobpKygBueEGg6UqwgzkEtIsh8%3D&reserved=0>

[-- Attachment #1.2: Type: text/html, Size: 5730 bytes --]

[-- Attachment #2: Type: text/plain, Size: 160 bytes --]

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel