Skip to contentSkip to conversation

mmc: core: don't initialize block size from ext_csd if not present

PATCH Superseded Peter Collingbourne 4 participants 7 messages Jan 12, 2021 – Jan 14, 2021 MMC, SD and SDIO

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mmc: core: don't initialize block size from ext_csd if not present

If extended CSD was not available, the eMMC driver would incorrectly set the block size to 0, as the data_sector_size field of ext_csd was never initialized. This issue was exposed by commit 817046ecddbc ("block: Align max_hw_sectors to logical blocksize") which caused max_sectors and max_hw_sectors to be set to 0 after setting the block size to 0, resulting in a kernel panic in bio_split when attempting to read from the device. Fix it by only reading the block size from ext_csd if it is available.

Fixes 817046ecddbc ("block: Align max_hw_sectors to logical blocksize")
---
 drivers/mmc/core/queue.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/mmc/core/queue.c +2 -1
index de7cb0369c30..735cdbf1145c 100644
@@ -20,6 +20,7 @@
 #include "core.h"
 #include "card.h"
 #include "host.h"
+#include "mmc_ops.h"
 
 #define MMC_DMA_MAP_MERGE_SEGMENTS	512
 
@@ -384,7 +385,7 @@ static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card)
 		     "merging was advertised but not possible");
 	blk_queue_max_segments(mq->queue, mmc_get_max_segments(host));
 
-	if (mmc_card_mmc(card))
+	if (mmc_card_mmc(card) && mmc_can_ext_csd(card))
 		block_size = card->ext_csd.data_sector_size;
 
 	blk_queue_logical_block_size(mq->queue, block_size);
Signature · 1 line
2.30.0.284.gd98b1dd5eaa7-goog
Reviewed-by on v1 1/1
#1 lore raw
Quoted 30 lines · Peter Collingbourne

If extended CSD was not available, the eMMC driver would incorrectly set the block size to 0, as the data_sector_size field of ext_csd was never initialized. This issue was exposed by commit 817046ecddbc ("block: Align max_hw_sectors to logical blocksize") which caused max_sectors and max_hw_sectors to be set to 0 after setting the block size to 0, resulting in a kernel panic in bio_split when attempting to read from the device. Fix it by only reading the block size from ext_csd if it is available.

Fixes 817046ecddbc ("block: Align max_hw_sectors to logical blocksize")
---
 drivers/mmc/core/queue.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/mmc/core/queue.c +1 -0
index de7cb0369c30..735cdbf1145c 100644
@@ -20,6 +20,7 @@
 #include "core.h"
 #include "card.h"
 #include "host.h"
+#include "mmc_ops.h"
 
 #define MMC_DMA_MAP_MERGE_SEGMENTS	512
 

drivers/mmc/core/queue.c static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card) · line 391

Quoted 8 lines
drivers/mmc/core/queue.c +1 -1
@@ -384,7 +385,7 @@ static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card)
 		     "merging was advertised but not possible");
 	blk_queue_max_segments(mq->queue, mmc_get_max_segments(host));
 
-	if (mmc_card_mmc(card))
+	if (mmc_card_mmc(card) && mmc_can_ext_csd(card))
 		block_size = card->ext_csd.data_sector_size;
 
 	blk_queue_logical_block_size(mq->queue, block_size);

This looks good to me. I wonder though if we should add a WARN() in blk_queue_logical_block_size() if it is passed a 0 block size to more easily catch this kind of bug.

Signature · 2 lines
Damien Le Moal
Western Digital Research
#2 lore raw
In reply to Peter Collingbourne
Thank you for the patch! Yet something to improve:

Hi Peter,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on hch-configfs/for-next v5.11-rc3 next-20210113] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Peter-Collingbourne/mmc-core-don-t-initialize-block-size-from-ext_csd-if-not-present/20210113-054414
base:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git e609571b5ffa3528bf85292de1ceaddac342bc1c
config: x86_64-rhel (attached as .config)
compiler: gcc-9 (Debian 9.3.0-15) 9.3.0
reproduce (this is a W=1 build):
        # https://github.com/0day-ci/linux/commit/562a77dc6c88601c5129c0bbb9141119494320c7
        git remote add linux-review https://github.com/0day-ci/linux
        git fetch --no-tags linux-review Peter-Collingbourne/mmc-core-don-t-initialize-block-size-from-ext_csd-if-not-present/20210113-054414
        git checkout 562a77dc6c88601c5129c0bbb9141119494320c7
        # save the attached .config to linux build tree
        make W=1 ARCH=x86_64 

If you fix the issue, kindly add following tag as appropriate

All errors (new ones prefixed by >>, old ones prefixed by <<):

>> ERROR: modpost: "mmc_can_ext_csd" [drivers/mmc/core/mmc_block.ko] undefined!

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

#3 lore raw
In reply to Peter Collingbourne
Quoted 10 lines · Peter Collingbourne

If extended CSD was not available, the eMMC driver would incorrectly set the block size to 0, as the data_sector_size field of ext_csd was never initialized. This issue was exposed by commit 817046ecddbc ("block: Align max_hw_sectors to logical blocksize") which caused max_sectors and max_hw_sectors to be set to 0 after setting the block size to 0, resulting in a kernel panic in bio_split when attempting to read from the device. Fix it by only reading the block size from ext_csd if it is available.

Fixes 817046ecddbc ("block: Align max_hw_sectors to logical blocksize")

I would go with the original commit i.e.

Fixes a5075eb94837 ("mmc: block: Allow disabling 512B sector size emulation")
Quoted 19 lines
---
 drivers/mmc/core/queue.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/mmc/core/queue.c +1 -0
index de7cb0369c30..735cdbf1145c 100644
@@ -20,6 +20,7 @@
 #include "core.h"
 #include "card.h"
 #include "host.h"
+#include "mmc_ops.h"
 
 #define MMC_DMA_MAP_MERGE_SEGMENTS	512
 

drivers/mmc/core/queue.c static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card) · line 389

Quoted 6 lines
drivers/mmc/core/queue.c +1 -1
@@ -384,7 +385,7 @@ static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card)
 		     "merging was advertised but not possible");
 	blk_queue_max_segments(mq->queue, mmc_get_max_segments(host));
 
-	if (mmc_card_mmc(card))
+	if (mmc_card_mmc(card) && mmc_can_ext_csd(card))
 		block_size = card->ext_csd.data_sector_size;

Might as well be:

	if (mmc_card_mmc(card) && card->ext_csd.data_sector_size)

block_size = card->ext_csd.data_sector_size;

Quoted 3 lines
drivers/mmc/core/queue.c +0 -0
 	blk_queue_logical_block_size(mq->queue, block_size);
 
#4 lore raw
Quoted 11 lines · Adrian Hunter
Quoted 9 lines · Peter Collingbourne

If extended CSD was not available, the eMMC driver would incorrectly set the block size to 0, as the data_sector_size field of ext_csd was never initialized. This issue was exposed by commit 817046ecddbc ("block: Align max_hw_sectors to logical blocksize") which caused max_sectors and max_hw_sectors to be set to 0 after setting the block size to 0, resulting in a kernel panic in bio_split when attempting to read from the device. Fix it by only reading the block size from ext_csd if it is available.

Quoted 5 lines
Quoted 1 line
Fixes 817046ecddbc ("block: Align max_hw_sectors to logical blocksize")

I would go with the original commit i.e.

Fixes a5075eb94837 ("mmc: block: Allow disabling 512B sector size emulation")

Sure, makes sense.

Quoted 22 lines
Quoted 22 lines
---
 drivers/mmc/core/queue.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/mmc/core/queue.c +1 -0
index de7cb0369c30..735cdbf1145c 100644
@@ -20,6 +20,7 @@
 #include "core.h"
 #include "card.h"
 #include "host.h"
+#include "mmc_ops.h"
 
 #define MMC_DMA_MAP_MERGE_SEGMENTS   512
 
@@ -384,7 +385,7 @@ static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card)
                   "merging was advertised but not possible");
      blk_queue_max_segments(mq->queue, mmc_get_max_segments(host));
 
Quoted 8 lines

drivers/mmc/core/queue.c static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card) · line 389

Quoted 3 lines
drivers/mmc/core/queue.c +1 -1
-     if (mmc_card_mmc(card))
+     if (mmc_card_mmc(card) && mmc_can_ext_csd(card))
              block_size = card->ext_csd.data_sector_size;

Might as well be:

        if (mmc_card_mmc(card) && card->ext_csd.data_sector_size)
                block_size = card->ext_csd.data_sector_size;

Can we rely on this data structure to be zero initialized? I suppose so, provided that it was allocated with mmc_alloc_card which uses kzalloc. But it isn't entirely obvious and I figure it may be a little better to be explicit in our intent here. But either way works for me.

Peter

#5 lore raw
Quoted 48 lines · Peter Collingbourne
Quoted 11 lines · Adrian Hunter
Quoted 9 lines · Peter Collingbourne

If extended CSD was not available, the eMMC driver would incorrectly set the block size to 0, as the data_sector_size field of ext_csd was never initialized. This issue was exposed by commit 817046ecddbc ("block: Align max_hw_sectors to logical blocksize") which caused max_sectors and max_hw_sectors to be set to 0 after setting the block size to 0, resulting in a kernel panic in bio_split when attempting to read from the device. Fix it by only reading the block size from ext_csd if it is available.

Quoted 5 lines
Quoted 1 line
Fixes 817046ecddbc ("block: Align max_hw_sectors to logical blocksize")

I would go with the original commit i.e.

Fixes a5075eb94837 ("mmc: block: Allow disabling 512B sector size emulation")

Sure, makes sense.

Quoted 28 lines
Quoted 19 lines
---
 drivers/mmc/core/queue.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/mmc/core/queue.c +1 -0
index de7cb0369c30..735cdbf1145c 100644
@@ -20,6 +20,7 @@
 #include "core.h"
 #include "card.h"
 #include "host.h"
+#include "mmc_ops.h"
 
 #define MMC_DMA_MAP_MERGE_SEGMENTS   512
 

drivers/mmc/core/queue.c static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card) · line 389

Quoted 6 lines
drivers/mmc/core/queue.c +1 -1
@@ -384,7 +385,7 @@ static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card)
                   "merging was advertised but not possible");
      blk_queue_max_segments(mq->queue, mmc_get_max_segments(host));
 
-     if (mmc_card_mmc(card))
+     if (mmc_card_mmc(card) && mmc_can_ext_csd(card))
              block_size = card->ext_csd.data_sector_size;

Might as well be:

Quoted 7 lines
Quoted 2 lines
        if (mmc_card_mmc(card) && card->ext_csd.data_sector_size)
                block_size = card->ext_csd.data_sector_size;

Can we rely on this data structure to be zero initialized? I suppose so, provided that it was allocated with mmc_alloc_card which uses kzalloc. But it isn't entirely obvious and I figure it may be a little better to be explicit in our intent here. But either way works for me.

The only valid values are 512 and 4096, so you could add WARN_ON(block_size != 512 && block_size != 4096) if you want.

#6 lore raw
Quoted 53 lines · Adrian Hunter
Quoted 51 lines · Peter Collingbourne
Quoted 11 lines · Adrian Hunter
Quoted 9 lines · Peter Collingbourne

If extended CSD was not available, the eMMC driver would incorrectly set the block size to 0, as the data_sector_size field of ext_csd was never initialized. This issue was exposed by commit 817046ecddbc ("block: Align max_hw_sectors to logical blocksize") which caused max_sectors and max_hw_sectors to be set to 0 after setting the block size to 0, resulting in a kernel panic in bio_split when attempting to read from the device. Fix it by only reading the block size from ext_csd if it is available.

Quoted 5 lines
Quoted 1 line
Fixes 817046ecddbc ("block: Align max_hw_sectors to logical blocksize")

I would go with the original commit i.e.

Fixes a5075eb94837 ("mmc: block: Allow disabling 512B sector size emulation")

Sure, makes sense.

Quoted 30 lines
Quoted 19 lines
---
 drivers/mmc/core/queue.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/mmc/core/queue.c +1 -0
index de7cb0369c30..735cdbf1145c 100644
@@ -20,6 +20,7 @@
 #include "core.h"
 #include "card.h"
 #include "host.h"
+#include "mmc_ops.h"
 
 #define MMC_DMA_MAP_MERGE_SEGMENTS   512
 

drivers/mmc/core/queue.c static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card) · line 389

Quoted 6 lines
drivers/mmc/core/queue.c +1 -1
@@ -384,7 +385,7 @@ static void mmc_setup_queue(struct mmc_queue *mq, struct mmc_card *card)
                   "merging was advertised but not possible");
      blk_queue_max_segments(mq->queue, mmc_get_max_segments(host));
 
-     if (mmc_card_mmc(card))
+     if (mmc_card_mmc(card) && mmc_can_ext_csd(card))
              block_size = card->ext_csd.data_sector_size;

Might as well be:

        if (mmc_card_mmc(card) && card->ext_csd.data_sector_size)
                block_size = card->ext_csd.data_sector_size;
Quoted 7 lines
Quoted 4 lines

Can we rely on this data structure to be zero initialized? I suppose so, provided that it was allocated with mmc_alloc_card which uses kzalloc. But it isn't entirely obvious and I figure it may be a little better to be explicit in our intent here. But either way works for me.

The only valid values are 512 and 4096, so you could add WARN_ON(block_size != 512 && block_size != 4096) if you want.

Okay, I did that in v2.

Peter