Skip to contentSkip to conversation

s390/dasd: fixes for data corruption and DMA alignment

PATCH Applied Stefan Haberland 2 participants 4 messages Aug 12, 2024

Subsystems (3)
lore
Share & tools
lore

Stefan Haberland

author maintainer
#0 lore raw

cover s390/dasd: fixes for data corruption and DMA alignment

Hi Jens,

please apply the following two patches that fix two rather severe issues. The first reverts a DMA alignment patch leading to errors with xfs and 512b blocksize DASDs.

The second patch fixes a possible data corruption for certain storage server errors.

Eric Farman (1):
s390/dasd: Remove DMA alignment

Stefan Haberland (1):
s390/dasd: fix error recovery leading to data corruption on ESE
devices

 drivers/s390/block/dasd.c          | 36 ++++++++++++-------
 drivers/s390/block/dasd_3990_erp.c | 10 ++----
 drivers/s390/block/dasd_eckd.c     | 55 +++++++++++++-----------------
 drivers/s390/block/dasd_genhd.c    |  1 -
 drivers/s390/block/dasd_int.h      |  2 +-
 5 files changed, 50 insertions(+), 54 deletions(-)
Signature · 1 line
2.43.0
author maintainer
#1 lore raw

PATCH 1/2 s390/dasd: Remove DMA alignment

From: Eric Farman <[email protected]>

This reverts commit bc792884b76f ("s390/dasd: Establish DMA alignment").

Quoting the original commit:
linux-next commit bf8d08532bc1 ("iomap: add support for dma aligned direct-io") changes the alignment requirement to come from the block device rather than the block size, and the default alignment requirement is 512-byte boundaries. Since DASD I/O has page alignments for IDAW/TIDAW requests, let's override this value to restore the expected behavior.

I mentioned TIDAW, but that was wrong. TIDAWs have no distinct alignment requirement (per p. 15-70 of POPS SA22-7832-13):

Unless otherwise specified, TIDAWs may designate a block of main storage on any boundary and length up to 4K bytes, provided the specified block does not cross a 4 K-byte boundary.

IDAWs do, but the original commit neglected that while ECKD DASD are typically formatted in 4096-byte blocks, they don't HAVE to be. Formatting an ECKD volume with smaller blocks is permitted (dasdfmt -b xxx), and the problematic commit enforces alignment properties to such a device that will result in errors, such as:

[test@host ~]# lsdasd -l a367 | grep blksz

     blksz:				512
   [test@host ~]# mkfs.xfs -f /dev/disk/by-path/ccw-0.0.a367-part1
   meta-data=/dev/dasdc1            isize=512    agcount=4, agsize=230075 blks
            =                       sectsz=512   attr=2, projid32bit=1
            =                       crc=1        finobt=1, sparse=1, rmapbt=1
            =                       reflink=1    bigtime=1 inobtcount=1 nrext64=1
   data     =                       bsize=4096   blocks=920299, imaxpct=25
            =                       sunit=0      swidth=0 blks
   naming   =version 2              bsize=4096   ascii-ci=0, ftype=1
   log      =internal log           bsize=4096   blocks=16384, version=2
            =                       sectsz=512   sunit=0 blks, lazy-count=1
   realtime =none                   extsz=4096   blocks=0, rtextents=0

error reading existing superblock: Invalid argument mkfs.xfs: pwrite failed: Invalid argument
libxfs_bwrite: write failed on (unknown) bno 0x70565c/0x100, err=22 mkfs.xfs: Releasing dirty buffer to free list!
found dirty buffer (bulk) on free list!
mkfs.xfs: pwrite failed: Invalid argument
...snipped...

The original commit omitted the FBA discipline for just this reason, but the formatted block size of the other disciplines was overlooked. The solution to all of this is to revert to the original behavior, such that the block size can be respected. There were two commits [1] that moved this code in the interim, so a straight git-revert is not possible, but the change is straightforward.

But what of the original problem? That was manifested with a direct-io QEMU guest, where QEMU itself was changed a month or two later with commit 25474d90aa ("block: use the request length for iov alignment") such that the blamed kernel commit is unnecessary.

[1] commit 0127a47f58c6 ("dasd: move queue setup to common code") commit fde07a4d74e3 ("dasd: use the atomic queue limits API")

Fixes bc792884b76f ("s390/dasd: Establish DMA alignment")
---
 drivers/s390/block/dasd_genhd.c | 1 -
 1 file changed, 1 deletion(-)
drivers/s390/block/dasd_genhd.c +0 -1
index 1aa426b1dedd..6da47a65af61 100644
@@ -41,7 +41,6 @@ int dasd_gendisk_alloc(struct dasd_block *block)
 		 */
 		.max_segment_size = PAGE_SIZE,
 		.seg_boundary_mask = PAGE_SIZE - 1,
-		.dma_alignment = PAGE_SIZE - 1,
 		.max_segments = USHRT_MAX,
 	};
 	struct gendisk *gdp;
Signature · 1 line
2.43.0
author maintainer
#2 lore raw

PATCH 2/2 s390/dasd: fix error recovery leading to data corruption on ESE devices

In reply to Stefan Haberland

Extent Space Efficient (ESE) or thin provisioned volumes need to be formatted on demand during usual IO processing.

The dasd_ese_needs_format function checks for error codes that signal the non existence of a proper track format.

The check for incorrect length is to imprecise since other error cases leading to transport of insufficient data also have this flag set. This might lead to data corruption in certain error cases for example during a storage server warmstart.

Fix by removing the check for incorrect length and replacing by explicitly checking for invalid track format in transport mode.

Also remove the check for file protected since this is not a valid ESE handling case.

Cc <[email protected]> # 5.3+
Fixes 5e2b17e712cf ("s390/dasd: Add dynamic formatting support for ESE volumes")
---
 drivers/s390/block/dasd.c          | 36 ++++++++++++-------
 drivers/s390/block/dasd_3990_erp.c | 10 ++----
 drivers/s390/block/dasd_eckd.c     | 55 +++++++++++++-----------------
 drivers/s390/block/dasd_int.h      |  2 +-
 4 files changed, 50 insertions(+), 53 deletions(-)
drivers/s390/block/dasd.c +23 -13
index 0a97cfedd706..42a4a996defb 100644
@@ -1601,9 +1601,15 @@ static int dasd_ese_needs_format(struct dasd_block *block, struct irb *irb)
 	if (!sense)
 		return 0;
 
-	return !!(sense[1] & SNS1_NO_REC_FOUND) ||
-		!!(sense[1] & SNS1_FILE_PROTECTED) ||
-		scsw_cstat(&irb->scsw) == SCHN_STAT_INCORR_LEN;
+	if (sense[1] & SNS1_NO_REC_FOUND)
+		return 1;
+
+	if ((sense[1] & SNS1_INV_TRACK_FORMAT) &&
+	    scsw_is_tm(&irb->scsw) &&
+	    !(sense[2] & SNS2_ENV_DATA_PRESENT))
+		return 1;
+
+	return 0;
 }
 
 static int dasd_ese_oos_cond(u8 *sense)
@@ -1624,7 +1630,7 @@ void dasd_int_handler(struct ccw_device *cdev, unsigned long intparm,
 	struct dasd_device *device;
 	unsigned long now;
 	int nrf_suppressed = 0;
-	int fp_suppressed = 0;
+	int it_suppressed = 0;
 	struct request *req;
 	u8 *sense = NULL;
 	int expires;
@@ -1679,8 +1685,9 @@ void dasd_int_handler(struct ccw_device *cdev, unsigned long intparm,
 		 */
 		sense = dasd_get_sense(irb);
 		if (sense) {
-			fp_suppressed = (sense[1] & SNS1_FILE_PROTECTED) &&
-				test_bit(DASD_CQR_SUPPRESS_FP, &cqr->flags);
+			it_suppressed =	(sense[1] & SNS1_INV_TRACK_FORMAT) &&
+				!(sense[2] & SNS2_ENV_DATA_PRESENT) &&
+				test_bit(DASD_CQR_SUPPRESS_IT, &cqr->flags);
 			nrf_suppressed = (sense[1] & SNS1_NO_REC_FOUND) &&
 				test_bit(DASD_CQR_SUPPRESS_NRF, &cqr->flags);
 
@@ -1695,7 +1702,7 @@ void dasd_int_handler(struct ccw_device *cdev, unsigned long intparm,
 				return;
 			}
 		}
-		if (!(fp_suppressed || nrf_suppressed))
+		if (!(it_suppressed || nrf_suppressed))
 			device->discipline->dump_sense_dbf(device, irb, "int");
 
 		if (device->features & DASD_FEATURE_ERPLOG)
@@ -2459,14 +2466,17 @@ static int _dasd_sleep_on_queue(struct list_head *ccw_queue, int interruptible)
 	rc = 0;
 	list_for_each_entry_safe(cqr, n, ccw_queue, blocklist) {
 		/*
-		 * In some cases the 'File Protected' or 'Incorrect Length'
-		 * error might be expected and error recovery would be
-		 * unnecessary in these cases.	Check if the according suppress
-		 * bit is set.
+		 * In some cases certain errors might be expected and
+		 * error recovery would be unnecessary in these cases.
+		 * Check if the according suppress bit is set.
 		 */
 		sense = dasd_get_sense(&cqr->irb);
-		if (sense && sense[1] & SNS1_FILE_PROTECTED &&
-		    test_bit(DASD_CQR_SUPPRESS_FP, &cqr->flags))
+		if (sense && (sense[1] & SNS1_INV_TRACK_FORMAT) &&
+		    !(sense[2] & SNS2_ENV_DATA_PRESENT) &&
+		    test_bit(DASD_CQR_SUPPRESS_IT, &cqr->flags))
+			continue;
+		if (sense && (sense[1] & SNS1_NO_REC_FOUND) &&
+		    test_bit(DASD_CQR_SUPPRESS_NRF, &cqr->flags))
 			continue;
 		if (scsw_cstat(&cqr->irb.scsw) == 0x40 &&
 		    test_bit(DASD_CQR_SUPPRESS_IL, &cqr->flags))
drivers/s390/block/dasd_3990_erp.c +2 -8
index bbbacfc386f2..d0aa267462c5 100644
@@ -1386,14 +1386,8 @@ dasd_3990_erp_file_prot(struct dasd_ccw_req * erp)
 
 	struct dasd_device *device = erp->startdev;
 
-	/*
-	 * In some cases the 'File Protected' error might be expected and
-	 * log messages shouldn't be written then.
-	 * Check if the according suppress bit is set.
-	 */
-	if (!test_bit(DASD_CQR_SUPPRESS_FP, &erp->flags))
-		dev_err(&device->cdev->dev,
-			"Accessing the DASD failed because of a hardware error\n");
+	dev_err(&device->cdev->dev,
+		"Accessing the DASD failed because of a hardware error\n");
 
 	return dasd_3990_erp_cleanup(erp, DASD_CQR_FAILED);
 
drivers/s390/block/dasd_eckd.c +24 -31

90 lines not on this page · raw message

drivers/s390/block/dasd_int.h +1 -1
index e5f40536b425..81cfb5c89681 100644
@@ -196,7 +196,7 @@ struct dasd_ccw_req {
  * The following flags are used to suppress output of certain errors.
  */
 #define DASD_CQR_SUPPRESS_NRF	4	/* Suppress 'No Record Found' error */
-#define DASD_CQR_SUPPRESS_FP	5	/* Suppress 'File Protected' error*/
+#define DASD_CQR_SUPPRESS_IT	5	/* Suppress 'Invalid Track' error*/
 #define DASD_CQR_SUPPRESS_IL	6	/* Suppress 'Incorrect Length' error */
 #define DASD_CQR_SUPPRESS_CR	7	/* Suppress 'Command Reject' error */
 
Signature · 1 line
2.43.0
maintainer Applied
#3 lore raw
In reply to Stefan Haberland
Quoted 8 lines · Stefan Haberland

please apply the following two patches that fix two rather severe issues. The first reverts a DMA alignment patch leading to errors with xfs and 512b blocksize DASDs.

The second patch fixes a possible data corruption for certain storage server errors.

[...]

Applied, thanks!

[1/2] s390/dasd: Remove DMA alignment

      commit: 2a07bb64d80152701d507b1498237ed1b8d83866
[2/2] s390/dasd: fix error recovery leading to data corruption on ESE devices
      commit: 7db4042336580dfd75cb5faa82c12cd51098c90b

Best regards,

Signature · 1 line
Jens Axboe