linux-kernel - [PATCH 2/2] drm/radeon: repeat the same EOP packet for EOP workaround on CIK

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20240617105846.1516006-3-uwu@icenowy.me>
Date: Mon, 17 Jun 2024 18:58:46 +0800
From: Icenowy Zheng <uwu@...nowy.me>
To: Alex Deucher <alexander.deucher@....com>,
	Christian König <christian.koenig@....com>,
	Pan Xinhui <Xinhui.Pan@....com>,
	David Airlie <airlied@...il.com>,
	Daniel Vetter <daniel@...ll.ch>,
	Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@....com>,
	Huacai Chen <chenhuacai@...nel.org>,
	WANG Xuerui <kernel@...0n.name>
Cc: amd-gfx@...ts.freedesktop.org,
	dri-devel@...ts.freedesktop.org,
	linux-kernel@...r.kernel.org,
	loongarch@...ts.linux.dev,
	Icenowy Zheng <uwu@...nowy.me>
Subject: [PATCH 2/2] drm/radeon: repeat the same EOP packet for EOP workaround on CIK

Ths first EOP packet with a sequence number as seq-1 seems to confuse
some PCIe hardware (e.g. Loongson 7A PCHs).

Use the real sequence number instead.

Fixes: a9c73a0e022c ("drm/radeon: workaround for CP HW bug on CIK")
Signed-off-by: Icenowy Zheng <uwu@...nowy.me>
---
 drivers/gpu/drm/radeon/cik.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
index 8275eeba0b349..9d203054f922a 100644
--- a/drivers/gpu/drm/radeon/cik.c
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -3543,9 +3543,7 @@ void cik_fence_gfx_ring_emit(struct radeon_device *rdev,
 	struct radeon_ring *ring = &rdev->ring[fence->ring];
 	u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
 
-	/* Workaround for cache flush problems. First send a dummy EOP
-	 * event down the pipe with seq one below.
-	 */
+	/* Workaround for cache flush problems by sending the EOP event twice */
 	radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4));
 	radeon_ring_write(ring, (EOP_TCL1_ACTION_EN |
 				 EOP_TC_ACTION_EN |
@@ -3554,10 +3552,9 @@ void cik_fence_gfx_ring_emit(struct radeon_device *rdev,
 	radeon_ring_write(ring, addr & 0xfffffffc);
 	radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) |
 				DATA_SEL(1) | INT_SEL(0));
-	radeon_ring_write(ring, fence->seq - 1);
+	radeon_ring_write(ring, fence->seq);
 	radeon_ring_write(ring, 0);
 
-	/* Then send the real EOP event down the pipe. */
 	radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4));
 	radeon_ring_write(ring, (EOP_TCL1_ACTION_EN |
 				 EOP_TC_ACTION_EN |
-- 
2.45.1