linux-kernel - [PATCH AUTOSEL 6.12 374/486] drm/amdgpu: Use active umc info from discovery

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: <20250505223922.2682012-374-sashal@kernel.org>
Date: Mon,  5 May 2025 18:37:30 -0400
From: Sasha Levin <sashal@...nel.org>
To: linux-kernel@...r.kernel.org,
	stable@...r.kernel.org
Cc: Lijo Lazar <lijo.lazar@....com>,
	Hawking Zhang <Hawking.Zhang@....com>,
	Alex Deucher <alexander.deucher@....com>,
	Sasha Levin <sashal@...nel.org>,
	christian.koenig@....com,
	airlied@...il.com,
	simona@...ll.ch,
	tao.zhou1@....com,
	YiPeng.Chai@....com,
	zhigang.luo@....com,
	victor.skvortsov@....com,
	sunil.khatri@....com,
	alex.sierra@....com,
	boyuan.zhang@....com,
	le.ma@....com,
	shane.xiao@....com,
	amd-gfx@...ts.freedesktop.org,
	dri-devel@...ts.freedesktop.org
Subject: [PATCH AUTOSEL 6.12 374/486] drm/amdgpu: Use active umc info from discovery

From: Lijo Lazar <lijo.lazar@....com>

[ Upstream commit f7a594e40517fa2ab25d5ca10e7b6a158f529fb5 ]

There could be configs where some UMC instances are harvested. This
information is obtained through discovery data and populated in
umc.active_mask. Avoid reassigning this as AID mask, instead use the
mask directly while iterating through umc instances. This is to avoid
accesses to harvested UMC instances.

v2: fix warning (Alex)

Signed-off-by: Lijo Lazar <lijo.lazar@....com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@....com>
Signed-off-by: Alex Deucher <alexander.deucher@....com>
Signed-off-by: Sasha Levin <sashal@...nel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c | 42 +++++++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c   |  1 -
 2 files changed, 42 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
index bb7b9b2eaac1a..8da0bddab3d23 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
@@ -383,6 +383,45 @@ int amdgpu_umc_fill_error_record(struct ras_err_data *err_data,
 	return 0;
 }
 
+static int amdgpu_umc_loop_all_aid(struct amdgpu_device *adev, umc_func func,
+				   void *data)
+{
+	uint32_t umc_node_inst;
+	uint32_t node_inst;
+	uint32_t umc_inst;
+	uint32_t ch_inst;
+	int ret;
+
+	/*
+	 * This loop is done based on the following -
+	 * umc.active mask = mask of active umc instances across all nodes
+	 * umc.umc_inst_num = maximum number of umc instancess per node
+	 * umc.node_inst_num = maximum number of node instances
+	 * Channel instances are not assumed to be harvested.
+	 */
+	dev_dbg(adev->dev, "active umcs :%lx umc_inst per node: %d",
+		adev->umc.active_mask, adev->umc.umc_inst_num);
+	for_each_set_bit(umc_node_inst, &(adev->umc.active_mask),
+			 adev->umc.node_inst_num * adev->umc.umc_inst_num) {
+		node_inst = umc_node_inst / adev->umc.umc_inst_num;
+		umc_inst = umc_node_inst % adev->umc.umc_inst_num;
+		LOOP_UMC_CH_INST(ch_inst) {
+			dev_dbg(adev->dev,
+				"node_inst :%d umc_inst: %d ch_inst: %d",
+				node_inst, umc_inst, ch_inst);
+			ret = func(adev, node_inst, umc_inst, ch_inst, data);
+			if (ret) {
+				dev_err(adev->dev,
+					"Node %d umc %d ch %d func returns %d\n",
+					node_inst, umc_inst, ch_inst, ret);
+				return ret;
+			}
+		}
+	}
+
+	return 0;
+}
+
 int amdgpu_umc_loop_channels(struct amdgpu_device *adev,
 			umc_func func, void *data)
 {
@@ -391,6 +430,9 @@ int amdgpu_umc_loop_channels(struct amdgpu_device *adev,
 	uint32_t ch_inst         = 0;
 	int ret = 0;
 
+	if (adev->aid_mask)
+		return amdgpu_umc_loop_all_aid(adev, func, data);
+
 	if (adev->umc.node_inst_num) {
 		LOOP_UMC_EACH_NODE_INST_AND_CH(node_inst, umc_inst, ch_inst) {
 			ret = func(adev, node_inst, umc_inst, ch_inst, data);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 9a212413c6d3a..78c527b56f7c5 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -1461,7 +1461,6 @@ static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev)
 		adev->umc.umc_inst_num = UMC_V12_0_UMC_INSTANCE_NUM;
 		adev->umc.node_inst_num /= UMC_V12_0_UMC_INSTANCE_NUM;
 		adev->umc.channel_offs = UMC_V12_0_PER_CHANNEL_OFFSET;
-		adev->umc.active_mask = adev->aid_mask;
 		adev->umc.retire_unit = UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL;
 		if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu)
 			adev->umc.ras = &umc_v12_0_ras;
-- 
2.39.5