Remove the legacy bad page retirement handling for UMC v12_0
Signed-off-by: Ce Sun <[email protected]>
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 16 +++----
drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c | 31 -------------
drivers/gpu/drm/amd/amdgpu/umc_v12_0.c | 61 -------------------------
3 files changed, 6 insertions(+), 102 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 1fe4f191c76f..e7156b85e727 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -243,16 +243,12 @@ static int amdgpu_check_address_validity(struct
amdgpu_device *adev,
(address >= RAS_UMC_INJECT_ADDR_LIMIT))
return -EFAULT;
- if (amdgpu_uniras_enabled(adev)) {
- if (amdgpu_sriov_vf(adev))
- count = amdgpu_virt_ras_convert_retired_address(adev,
address,
- page_pfns, ARRAY_SIZE(page_pfns));
- else
- count = amdgpu_ras_mgr_lookup_bad_pages_in_a_row(adev,
address,
- page_pfns, ARRAY_SIZE(page_pfns));
- } else
- count = amdgpu_umc_lookup_bad_pages_in_a_row(adev,
- address, page_pfns, ARRAY_SIZE(page_pfns));
+ if (amdgpu_sriov_vf(adev))
+ count = amdgpu_virt_ras_convert_retired_address(adev, address,
+ page_pfns, ARRAY_SIZE(page_pfns));
+ else
+ count = amdgpu_ras_mgr_lookup_bad_pages_in_a_row(adev, address,
+ page_pfns, ARRAY_SIZE(page_pfns));
if (count <= 0)
return -EPERM;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
index 0a4a8394489a..516ae7a5cfeb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
@@ -519,37 +519,6 @@ int amdgpu_umc_pages_in_a_row(struct amdgpu_device *adev,
return -EINVAL;
}
-int amdgpu_umc_lookup_bad_pages_in_a_row(struct amdgpu_device *adev,
- uint64_t pa_addr, uint64_t *pfns, int len)
-{
- int i, ret;
- struct ras_err_data err_data;
-
- err_data.err_addr = kcalloc(adev->umc.retire_unit,
- sizeof(struct eeprom_table_record), GFP_KERNEL);
- if (!err_data.err_addr) {
- dev_warn(adev->dev, "Failed to alloc memory in bad page
lookup!\n");
- return 0;
- }
-
- ret = amdgpu_umc_pages_in_a_row(adev, &err_data, pa_addr);
- if (ret)
- goto out;
-
- for (i = 0; i < adev->umc.retire_unit; i++) {
- if (i >= len)
- goto out;
-
- pfns[i] = err_data.err_addr[i].retired_page;
- }
- ret = i;
- adev->umc.err_addr_cnt = err_data.err_addr_cnt;
-
-out:
- kfree(err_data.err_addr);
- return ret;
-}
-
int amdgpu_umc_pa2mca(struct amdgpu_device *adev,
uint64_t pa, uint64_t *mca, enum amdgpu_memory_partition nps)
{
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
index ebceb933481e..e1d900818a81 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
@@ -28,8 +28,6 @@
#include "umc/umc_12_0_0_sh_mask.h"
#include "mp/mp_13_0_6_sh_mask.h"
-#define MAX_ECC_NUM_PER_RETIREMENT 32
-
bool umc_v12_0_is_deferred_error(struct amdgpu_device *adev, uint64_t
mc_umc_status)
{
dev_dbg(adev->dev,
@@ -291,64 +289,6 @@ static bool umc_v12_0_check_ecc_err_status(struct
amdgpu_device *adev,
return false;
}
-static int umc_v12_0_fill_error_record(struct amdgpu_device *adev,
- struct ras_ecc_err *ecc_err, void
*ras_error_status)
-{
- struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
- uint64_t page_pfn[UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL];
- int ret, i, count;
-
- if (!err_data || !ecc_err)
- return -EINVAL;
-
- memset(page_pfn, 0, sizeof(page_pfn));
- count = amdgpu_umc_lookup_bad_pages_in_a_row(adev,
- ecc_err->pa_pfn << AMDGPU_GPU_PAGE_SHIFT,
- page_pfn, ARRAY_SIZE(page_pfn));
-
- for (i = 0; i < count; i++) {
- ret = amdgpu_umc_fill_error_record(err_data,
- ecc_err->addr,
- page_pfn[i] << AMDGPU_GPU_PAGE_SHIFT,
- ecc_err->channel_idx,
- MCA_IPID_2_UMC_INST(ecc_err->ipid));
- if (ret)
- break;
- }
-
- err_data->de_count++;
-
- return ret;
-}
-
-static void umc_v12_0_query_ras_ecc_err_addr(struct amdgpu_device *adev,
- void *ras_error_status)
-{
- struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
- struct ras_ecc_err *entries[MAX_ECC_NUM_PER_RETIREMENT];
- struct radix_tree_root *ecc_tree;
- int new_detected, ret, i;
-
- ecc_tree = &con->umc_ecc_log.de_page_tree;
-
- mutex_lock(&con->umc_ecc_log.lock);
- new_detected = radix_tree_gang_lookup_tag(ecc_tree, (void **)entries,
- 0, ARRAY_SIZE(entries), UMC_ECC_NEW_DETECTED_TAG);
- for (i = 0; i < new_detected; i++) {
- if (!entries[i])
- continue;
-
- ret = umc_v12_0_fill_error_record(adev, entries[i],
ras_error_status);
- if (ret) {
- dev_err(adev->dev, "Fail to fill umc error record,
ret:%d\n", ret);
- break;
- }
- radix_tree_tag_clear(ecc_tree,
- entries[i]->pa_pfn, UMC_ECC_NEW_DETECTED_TAG);
- }
- mutex_unlock(&con->umc_ecc_log.lock);
-}
-
static uint32_t umc_v12_0_get_die_id(struct amdgpu_device *adev,
uint64_t mca_addr, uint64_t retired_page)
{
@@ -391,7 +331,6 @@ struct amdgpu_umc_ras umc_v12_0_ras = {
.ras_block = {
.hw_ops = NULL,
},
- .ecc_info_query_ras_error_address = umc_v12_0_query_ras_ecc_err_addr,
.check_ecc_err_status = umc_v12_0_check_ecc_err_status,
.convert_ras_err_addr = umc_v12_0_convert_error_address,
.get_die_id_from_pa = umc_v12_0_get_die_id,
--
2.34.1
