drm/amdgpu: add range check for RAS bad page address
Subsystems (2)
PATCH 1/2 drm/amdgpu: add range check for RAS bad page address
Exclude invalid bad pages.
--- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 58 ++++++++++++------------- 1 file changed, 28 insertions(+), 30 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +28 -30
@@ -136,9 +136,9 @@ enum amdgpu_ras_retire_page_reservation {atomic_t amdgpu_ras_in_intr =ATOMIC_INIT (0 ); -static bool amdgpu_ras_check_bad_page_unlock (struct amdgpu_ras *con, +static int amdgpu_ras_check_bad_page_unlock (struct amdgpu_ras *con,uint64_t addr); -static bool amdgpu_ras_check_bad_page (struct amdgpu_device *adev, +static int amdgpu_ras_check_bad_page (struct amdgpu_device *adev,uint64_t addr);#ifdef CONFIG_X86_MCE_AMD static void amdgpu_register_bad_pages_mca_notifier (struct amdgpu_device *adev);
@@ -169,18 +169,16 @@ static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t addrestruct eeprom_table_record err_rec;int ret; -if ((address >= adev->gmc.mc_vram_size) || - (address >=RAS_UMC_INJECT_ADDR_LIMIT )) { + ret =amdgpu_ras_check_bad_page (adev, address); +if (ret == -EINVAL ) {dev_warn (adev->dev, -"RAS WARN: input address 0x%llx is invalid.\n" , - address); +"RAS WARN: input address 0x%llx is invalid.\n" , + address);return -EINVAL ; - } - -if (amdgpu_ras_check_bad_page (adev, address)) { + }else if (ret ==1 ) {dev_warn (adev->dev, -"RAS WARN: 0x%llx has already been marked as bad page!\n" , - address); +"RAS WARN: 0x%llx has already been marked as bad page!\n" , + address);return 0 ; }
@@ -513,22 +511,16 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f, ret =amdgpu_ras_feature_enable (adev, &data.head,1 );break ;case 2 : -if ((data.inject.address >= adev->gmc.mc_vram_size && - adev->gmc.mc_vram_size) || - (data.inject.address >=RAS_UMC_INJECT_ADDR_LIMIT )) { -dev_warn (adev->dev,"RAS WARN: input address " -"0x%llx is invalid." , +/* umc ce/ue error injection for a bad page is not allowed */ +if (data.head.block ==AMDGPU_RAS_BLOCK__UMC ) + ret =amdgpu_ras_check_bad_page (adev, data.inject.address); +if (ret == -EINVAL ) { +dev_warn (adev->dev,"RAS WARN: input address 0x%llx is invalid." , data.inject.address); - ret = -EINVAL ;break ; - } - -/* umc ce/ue error injection for a bad page is not allowed */ -if ((data.head.block ==AMDGPU_RAS_BLOCK__UMC ) && -amdgpu_ras_check_bad_page (adev, data.inject.address)) { -dev_warn (adev->dev,"RAS WARN: inject: 0x%llx has " -"already been marked as bad!\n" , - data.inject.address); + }else if (ret ==1 ) { +dev_warn (adev->dev,"RAS WARN: inject: 0x%llx has already been marked as bad!\n" , + data.inject.address);break ; }
@@ -3122,18 +3114,24 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev)return ret; } -static bool amdgpu_ras_check_bad_page_unlock (struct amdgpu_ras *con, +static int amdgpu_ras_check_bad_page_unlock (struct amdgpu_ras *con,uint64_t addr) {struct ras_err_handler_data *data = con->eh_data; +struct amdgpu_device *adev = con->adev;int i; +if ((addr >= adev->gmc.mc_vram_size && + adev->gmc.mc_vram_size) || + (addr >=RAS_UMC_INJECT_ADDR_LIMIT )) +return -EINVAL ; + addr >>=AMDGPU_GPU_PAGE_SHIFT ;for (i =0 ; i < data->count; i++)if (addr == data->bps[i].retired_page) -return true; +return 1 ; -return false; +return 0 ; }/*
@@ -3141,11 +3139,11 @@ static bool amdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con,* * Note: this check is only for umc block */ -static bool amdgpu_ras_check_bad_page (struct amdgpu_device *adev, +static int amdgpu_ras_check_bad_page (struct amdgpu_device *adev,uint64_t addr) {struct amdgpu_ras *con =amdgpu_ras_get_context (adev); -bool ret = false; +int ret =0 ;if (!con || !con->eh_data)return ret;
Signature · 1 line
2.34.1
PATCH 2/2 drm/amdgpu: adjust the update of RAS bad page number
One eeprom record may not map to unit number of bad pages, the accurate bad page number is gotten after bad page address check.
--- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 43 +++++++++++-------- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h | 3 ++ .../gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c | 6 +-- 3 files changed, 31 insertions(+), 21 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +26 -17
@@ -2568,6 +2568,9 @@ static int amdgpu_ras_badpages_read(struct amdgpu_device *adev, }for (; i < data->count; i++) { +if (!data->bps[i].ts) +continue ; + (*bps)[i] = (struct ras_badpage ){ .bp = data->bps[i].retired_page, .size =AMDGPU_GPU_PAGE_SIZE ,
@@ -2581,7 +2584,7 @@ static int amdgpu_ras_badpages_read(struct amdgpu_device *adev, (*bps)[i].flags =AMDGPU_RAS_RETIRE_PAGE_FAULT ; } - *count = data->count; + *count = con->bad_page_num; out:mutex_unlock (&con->recovery_lock);return ret;
@@ -2809,8 +2812,11 @@ static int __amdgpu_ras_restore_bad_pages(struct amdgpu_device *adev,for (j =0 ; j < count; j++) {if (amdgpu_ras_check_bad_page_unlock (con, - bps[j].retired_page <<AMDGPU_GPU_PAGE_SHIFT )) + bps[j].retired_page <<AMDGPU_GPU_PAGE_SHIFT )) { + data->count++; + data->space_left--;continue ; + }if (!data->space_left &&amdgpu_ras_realloc_eh_data_space (adev, data,256 )) {
@@ -2823,6 +2829,7 @@ static int __amdgpu_ras_restore_bad_pages(struct amdgpu_device *adev, sizeof(struct eeprom_table_record)); data->count++; data->space_left--; + con->bad_page_num++; }return 0 ;
@@ -2954,7 +2961,7 @@ int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev, ret =__amdgpu_ras_convert_rec_array_from_rom (adev, &bps[i], &err_data, nps);if (ret) - control->ras_num_bad_pages -= adev->umc.retire_unit; + con->bad_page_num -= adev->umc.retire_unit; i += (adev->umc.retire_unit -1 ); }else {break ;
@@ -2968,8 +2975,10 @@ int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev, ret =__amdgpu_ras_convert_rec_from_rom (adev, &bps[i], &err_data, nps);if (ret) - control->ras_num_bad_pages -= adev->umc.retire_unit; + con->bad_page_num -= adev->umc.retire_unit; } + + con->eh_data->count_saved = con->eh_data->count; }else { ret =__amdgpu_ras_restore_bad_pages (adev, bps, pages); }
@@ -2992,7 +3001,7 @@ int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev,struct amdgpu_ras *con =amdgpu_ras_get_context (adev);struct ras_err_handler_data *data;struct amdgpu_ras_eeprom_control *control; -int save_count, unit_num, bad_page_num, i; +int unit_num, i;if (!con || !con->eh_data) {if (new_cnt)
@@ -3013,27 +3022,25 @@ int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev,mutex_lock (&con->recovery_lock); control = &con->eeprom_control; data = con->eh_data; - bad_page_num = control->ras_num_bad_pages; - save_count = data->count - bad_page_num; + unit_num = data->count / adev->umc.retire_unit - control->ras_num_recs;mutex_unlock (&con->recovery_lock); - unit_num = save_count / adev->umc.retire_unit;if (new_cnt) *new_cnt = unit_num;/* only new entries are saved */ -if (save_count >0 ) { +if (unit_num >0 ) {/*old asics only save pa to eeprom like before*/ if (IP_VERSION_MAJ (amdgpu_ip_version (adev,UMC_HWIP ,0 )) <12 ) {if (amdgpu_ras_eeprom_append (control, - &data->bps[bad_page_num], save_count)) { + &data->bps[data->count_saved], unit_num)) {dev_err (adev->dev,"Failed to save EEPROM table data!" );return -EIO ; } }else {for (i =0 ; i < unit_num; i++) {if (amdgpu_ras_eeprom_append (control, - &data->bps[bad_page_num + + &data->bps[data->count_saved + i * adev->umc.retire_unit],1 )) {dev_err (adev->dev,"Failed to save EEPROM table data!" );return -EIO ;
@@ -3041,7 +3048,9 @@ int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev, } } -dev_info (adev->dev,"Saved %d pages to EEPROM table.\n" , save_count); +dev_info (adev->dev,"Saved %d pages to EEPROM table.\n" , + con->bad_page_num - control->ras_num_bad_pages); + data->count_saved = data->count; }return 0 ;
@@ -3096,17 +3105,17 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev) } } + ret =amdgpu_ras_add_bad_pages (adev, bps, control->ras_num_recs, true); +if (ret) + goto out; + ret =amdgpu_ras_eeprom_check (control);if (ret) goto out;/* HW not usable */ -if (amdgpu_ras_is_rma (adev)) { +if (amdgpu_ras_is_rma (adev)) ret = -EHWPOISON ; - goto out; - } - - ret =amdgpu_ras_add_bad_pages (adev, bps, control->ras_num_recs, true); } out:
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h +3 -0
@@ -570,6 +570,8 @@ struct amdgpu_ras {struct ras_event_manager *event_mgr;uint64_t reserved_pages_in_bytes; + +int bad_page_num; };struct ras_fs_data {
@@ -608,6 +610,7 @@ struct ras_err_handler_data {struct eeprom_table_record *bps;/* the count of entries */ int count; +int count_saved;/* the space can place new entries */ int space_left; };
drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c +2 -4
@@ -743,8 +743,7 @@ amdgpu_ras_eeprom_append_table(struct amdgpu_ras_eeprom_control *control,else control->ras_num_mca_recs += num; - control->ras_num_bad_pages = control->ras_num_pa_recs + - control->ras_num_mca_recs * adev->umc.retire_unit; + control->ras_num_bad_pages = con->bad_page_num; Out:kfree (buf);return res;
@@ -1457,8 +1456,7 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control)if (!__get_eeprom_i2c_addr (adev, control))return -EINVAL ; - control->ras_num_bad_pages = control->ras_num_pa_recs + - control->ras_num_mca_recs * adev->umc.retire_unit; + control->ras_num_bad_pages = ras->bad_page_num;if (hdr->header ==RAS_TABLE_HDR_VAL ) { dev_dbg(adev->dev,
Signature · 1 line
2.34.1
[AMD Official Use Only - AMD Internal Distribution Only]
Ping for the series...
Quoted 141 lines
-----Original Message-----
From: Zhou1, Tao <[email protected]>
Sent: Friday, July 11, 2025 5:06 PM
To: [email protected]
Cc: Zhou1, Tao <[email protected]>
Subject: [PATCH 1/2] drm/amdgpu: add range check for RAS bad page addressExclude invalid bad pages.
Signed-off-by Tao Zhou <[email protected]>--- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 58 ++++++++++++------------- 1 file changed, 28 insertions(+), 30 deletions(-)drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +2
-2@@ -136,9 +136,9 @@ enum amdgpu_ras_retire_page_reservation {atomic_t amdgpu_ras_in_intr =ATOMIC_INIT (0 ); -static bool amdgpu_ras_check_bad_page_unlock (struct amdgpu_ras *con, +static int amdgpu_ras_check_bad_page_unlock (struct amdgpu_ras *con,uint64_t addr); -static bool amdgpu_ras_check_bad_page (struct amdgpu_device *adev, +static int amdgpu_ras_check_bad_page (struct amdgpu_device *adev,uint64_t addr);#ifdef CONFIG_X86_MCE_AMD static void amdgpu_register_bad_pages_mca_notifier (struct amdgpu_device *adev);@@ -169,18 +169,16 @@ static int amdgpu_reserve_page_direct(struct
amdgpu_device*adev,uint64_taddrestruct eeprom_table_record err_rec; int ret;drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +5
-8-if ((address >= adev->gmc.mc_vram_size) || - (address >=RAS_UMC_INJECT_ADDR_LIMIT )) { + ret =amdgpu_ras_check_bad_page (adev, address); +if (ret == -EINVAL ) {dev_warn (adev->dev, -"RAS WARN: input address 0x%llx is invalid.\n" , - address); +"RAS WARN: input address 0x%llx is invalid.\n" , + address);return -EINVAL ; - } - -if (amdgpu_ras_check_bad_page (adev, address)) { + }else if (ret ==1 ) { dev_warn(adev->dev, - "RAS WARN: 0x%llx has already been marked as badpage!\n",
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +1
-1- address); + "RAS WARN: 0x%llx has already been marked as badpage!\n",
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +8
-14+ address);return 0 ; }@@ -513,22 +511,16 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f, ret =amdgpu_ras_feature_enable (adev, &data.head,1 );break ;case 2 : -if ((data.inject.address >= adev->gmc.mc_vram_size && - adev->gmc.mc_vram_size) || - (data.inject.address >=RAS_UMC_INJECT_ADDR_LIMIT )) { -dev_warn (adev->dev,"RAS WARN: input address " -"0x%llx is invalid." , +/* umc ce/ue error injection for a bad page is not allowed */ +if (data.head.block ==AMDGPU_RAS_BLOCK__UMC ) + ret =amdgpu_ras_check_bad_page (adev, data.inject.address); +if (ret == -EINVAL ) { + dev_warn(adev->dev, "RAS WARN: input address 0x%llx is invalid.", data.inject.address); - ret = -EINVAL ;break ; - } - -/* umc ce/ue error injection for a bad page is not allowed */ -if ((data.head.block ==AMDGPU_RAS_BLOCK__UMC ) && -amdgpu_ras_check_bad_page (adev, data.inject.address)) { -dev_warn (adev->dev,"RAS WARN: inject: 0x%llx has " -"already been marked as bad!\n" , - data.inject.address); + }else if (ret ==1 ) { + dev_warn(adev->dev, "RAS WARN: inject: 0x%llx has alreadybeen marked as bad!\n",
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +10
-3+ data.inject.address);break ; }@@ -3122,18 +3114,24 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev)return ret; } -static bool amdgpu_ras_check_bad_page_unlock (struct amdgpu_ras *con, +static int amdgpu_ras_check_bad_page_unlock (struct amdgpu_ras *con,uint64_t addr) {struct ras_err_handler_data *data = con->eh_data; +struct amdgpu_device *adev = con->adev;int i; +if ((addr >= adev->gmc.mc_vram_size && + adev->gmc.mc_vram_size) || + (addr >=RAS_UMC_INJECT_ADDR_LIMIT )) +return -EINVAL ; + addr >>=AMDGPU_GPU_PAGE_SHIFT ;for (i =0 ; i < data->count; i++)if (addr == data->bps[i].retired_page) -return true; +return 1 ; -return false; +return 0 ; }/*
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +2
-2@@ -3141,11 +3139,11 @@ static boolamdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con, * * Note: this check is only for umc block */ -static bool amdgpu_ras_check_bad_page (struct amdgpu_device *adev, +static int amdgpu_ras_check_bad_page (struct amdgpu_device *adev,uint64_t addr) {struct amdgpu_ras *con =amdgpu_ras_get_context (adev); -bool ret = false; +int ret =0 ;if (!con || !con->eh_data)return ret;--
2.34.1
[AMD Official Use Only - AMD Internal Distribution Only]
Series is
Regards,
Hawking
-----Original Message-----
From: amd-gfx <[email protected]> On Behalf Of Tao Zhou
Sent: Friday, July 11, 2025 17:06
To: [email protected]
Cc: Zhou1, Tao <[email protected]>
Subject: [PATCH 1/2] drm/amdgpu: add range check for RAS bad page address
Exclude invalid bad pages.
--- drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c | 58 ++++++++++++------------- 1 file changed, 28 insertions(+), 30 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +2 -2
@@ -136,9 +136,9 @@ enum amdgpu_ras_retire_page_reservation {atomic_t amdgpu_ras_in_intr =ATOMIC_INIT (0 ); -static bool amdgpu_ras_check_bad_page_unlock (struct amdgpu_ras *con, +static int amdgpu_ras_check_bad_page_unlock (struct amdgpu_ras *con,uint64_t addr); -static bool amdgpu_ras_check_bad_page (struct amdgpu_device *adev, +static int amdgpu_ras_check_bad_page (struct amdgpu_device *adev,uint64_t addr);#ifdef CONFIG_X86_MCE_AMD static void amdgpu_register_bad_pages_mca_notifier(struct amdgpu_device *adev); @@ -169,18 +169,16 @@ static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t addre
struct eeprom_table_record err_rec;
int ret;- if ((address >= adev->gmc.mc_vram_size) ||
- (address >= RAS_UMC_INJECT_ADDR_LIMIT)) {
+ ret = amdgpu_ras_check_bad_page(adev, address);
+ if (ret == -EINVAL) {
dev_warn(adev->dev,
- "RAS WARN: input address 0x%llx is invalid.\n",
- address);
+ "RAS WARN: input address 0x%llx is invalid.\n",
+ address);
return -EINVAL;
- }
-
- if (amdgpu_ras_check_bad_page(adev, address)) {
+ } else if (ret == 1) {
dev_warn(adev->dev,
- "RAS WARN: 0x%llx has already been marked as bad page!\n",
- address);
+ "RAS WARN: 0x%llx has already been marked as bad page!\n",
+ address);
return 0;
}@@ -513,22 +511,16 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f,
ret = amdgpu_ras_feature_enable(adev, &data.head, 1);
break;
case 2:- if ((data.inject.address >= adev->gmc.mc_vram_size &&
- adev->gmc.mc_vram_size) ||
- (data.inject.address >= RAS_UMC_INJECT_ADDR_LIMIT)) {
- dev_warn(adev->dev, "RAS WARN: input address "
- "0x%llx is invalid.",
+ /* umc ce/ue error injection for a bad page is not allowed */
+ if (data.head.block == AMDGPU_RAS_BLOCK__UMC)
+ ret = amdgpu_ras_check_bad_page(adev, data.inject.address);
+ if (ret == -EINVAL) {
+ dev_warn(adev->dev, "RAS WARN: input address 0x%llx is invalid.",
data.inject.address);
- ret = -EINVAL;
break;- }
-
- /* umc ce/ue error injection for a bad page is not allowed */
- if ((data.head.block == AMDGPU_RAS_BLOCK__UMC) &&
- amdgpu_ras_check_bad_page(adev, data.inject.address)) {
- dev_warn(adev->dev, "RAS WARN: inject: 0x%llx has "
- "already been marked as bad!\n",
- data.inject.address);
+ } else if (ret == 1) {
+ dev_warn(adev->dev, "RAS WARN: inject: 0x%llx has already been marked as bad!\n",
+ data.inject.address);
break;
}@@ -3122,18 +3114,24 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev)
return ret;
}
-static bool amdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con, +static int amdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con,
uint64_t addr)
{
struct ras_err_handler_data *data = con->eh_data;
+ struct amdgpu_device *adev = con->adev;
int i;+ if ((addr >= adev->gmc.mc_vram_size &&
+ adev->gmc.mc_vram_size) ||
+ (addr >= RAS_UMC_INJECT_ADDR_LIMIT))
+ return -EINVAL;
+
addr >>= AMDGPU_GPU_PAGE_SHIFT;
for (i = 0; i < data->count; i++)
if (addr == data->bps[i].retired_page)- return true;
+ return 1;
- return false;
+ return 0;
}
/*
@@ -3141,11 +3139,11 @@ static bool amdgpu_ras_check_bad_page_unlock(struct amdgpu_ras *con,
*
* Note: this check is only for umc block
*/
-static bool amdgpu_ras_check_bad_page(struct amdgpu_device *adev,
+static int amdgpu_ras_check_bad_page(struct amdgpu_device *adev,
uint64_t addr)
{
struct amdgpu_ras *con = amdgpu_ras_get_context(adev);- bool ret = false;
+ int ret = 0;
if (!con || !con->eh_data)
return ret;--
2.34.1