Re: 2.6.24-rc2-mm1 -- mkfs failing on variety of fs types

[Andrew Morton <akpm@linux-foundation.org>]

On Wed, 14 Nov 2007 08:56:01 +0000 Andy Whitcroft <apw@shadowen.org> wrote:

> We seem to have some general problem with mkfs for all filesystems.
> I am seeing this across at least three test systems although
> most are unable to compile this kernel :(, even with the hotfix.
> Basically, all mkfs operations for any filsystem type are failing,
> ext2 reports this as "short write", various others are mentioning
> pwrite and pwrite64 returning bad things:
> 
> ext2: Could not write 8 blocks in inode table starting at 851970:
> 	Attempt to write block from filesystem resulted in short
> 	write
> 
> reiserfs: bwrite: write 4096 bytes returned -1 (block=851968,
> 	dev=3): No space left on device
> 
> xfs: mkfs.xfs: pwrite64 failed: No space left on device
> 
> Nothing is reported in dmesg at the time as far as I can tell.  From the
> ext2 log I would swear we get this error on a block number far below that
> which is reported written successfully, though I cannot say I trust mkfs.
> 
> Nothing obvious has changed pwrite or block/* to my eye, so heck knows
> where this is coming from.  2.6.24-rc2 works on these same systems as
> goes the latest 2.6.24-rc2-git5.
> 
> Full mkfs output below.
> 
> -apw
> 
> *** elm3b6, x86_64:
> 
> mke2fs 1.37 (21-Mar-2005)
> Filesystem label=
> OS type: Linux
> Block size=4096 (log=2)
> Fragment size=4096 (log=2)
> 1465920 inodes, 2929854 blocks
> 146492 blocks (5.00%) reserved for the super user
> First data block=0
> 90 block groups
> 32768 blocks per group, 32768 fragments per group
> 16288 inodes per group
> Superblock backups stored on blocks: 
> 	32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208
> 
> mkfs.ext2: Attempt to write block from filesystem resulted in short write while zeroing block 2929824 at end of filesystem
> Writing inode tables: 
> Could not write 8 blocks in inode table starting at 851970: Attempt to write block from filesystem resulted in short write
> ===
> mkfs.jfs version 1.1.7, 22-Jul-2004
> The specified disk did not finish formatting.
> ===
> mkfs.reiserfs 3.6.19 (2003 www.namesys.com)
> [...]
> Guessing about desired format.. Kernel 2.6.24-rc2-mm1-autokern1 is running.
> Format 3.6 with standard journal
> Count of blocks on the device: 2929840
> Number of blocks consumed by mkreiserfs formatting process: 8301
> Blocksize: 4096
> Hash function used to sort names: "r5"
> Journal Size 8193 blocks (first block 18)
> Journal Max transaction length 1024
> inode generation number: 0
> UUID: c759e218-681b-4891-b4c4-33466d4eb4f0
> Initializing journal - 0%....20%....40%....60%....80%....100%
> bwrite: write 4096 bytes returned -1 (block=851968, dev=3): No space left on device
> ===
> mkfs.xfs: pwrite64 failed: No space left on device
> meta-data=/dev/sdb2              isize=256    agcount=16, agsize=183115 blks
>          =                       sectsz=512  
> data     =                       bsize=4096   blocks=2929840, imaxpct=25
>          =                       sunit=0      swidth=0 blks, unwritten=1
> naming   =version 2              bsize=4096  
> log      =internal log           bsize=4096   blocks=2560, version=1
>          =                       sectsz=512   sunit=0 blks
> realtime =none                   extsz=65536  blocks=0, rtextents=0
> 
> 
> *** elm3b239, x86_64:
> 
> mke2fs 1.38 (30-Jun-2005)
> Filesystem label=
> OS type: Linux
> Block size=4096 (log=2)
> Fragment size=4096 (log=2)
> 2410624 inodes, 4819500 blocks
> 240975 blocks (5.00%) reserved for the super user
> First data block=0
> 148 block groups
> 32768 blocks per group, 32768 fragments per group
> 16288 inodes per group
> Superblock backups stored on blocks: 
> 	32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 
> 	4096000
> 
> mkfs.ext2: Attempt to write block from filesystem resulted in short write while zeroing block 4819472 at end of filesystem
> Writing inode tables: 
> Could not write 8 blocks in inode table starting at 655362: Attempt to write block from filesystem resulted in short write
> 
> 
> *** pSeries-101, ppc64
> 
> mke2fs 1.38 (30-Jun-2005)
> Filesystem label=
> OS type: Linux
> Block size=4096 (log=2)
> Fragment size=4096 (log=2)
> 1281696 inodes, 2560000 blocks
> 128000 blocks (5.00%) reserved for the super user
> First data block=0
> 79 block groups
> 32768 blocks per group, 32768 fragments per group
> 16224 inodes per group
> Superblock backups stored on blocks: 
> 	32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632
> 
> mkfs.ext2: Attempt to write block from filesystem resulted in short write while zeroing block 2559984 at end of filesystem
> Writing inode tables: 
> Could not write 8 blocks in inode table starting at 491522: Attempt to write block from filesystem resulted in short write
> 
> 
> *** gekko-lp1, ppc64
> 
> mke2fs 1.38 (30-Jun-2005)
> Filesystem label=
> OS type: Linux
> Block size=4096 (log=2)
> Fragment size=4096 (log=2)
> 1224000 inodes, 2443880 blocks
> 122194 blocks (5.00%) reserved for the super user
> First data block=0
> 75 block groups
> 32768 blocks per group, 32768 fragments per group
> 16320 inodes per group
> Superblock backups stored on blocks: 
> 	32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632
> 
> mkfs.ext2: Attempt to write block from filesystem resulted in short write while zeroing block 2443856 at end of filesystem
> Writing inode tables: 
> Could not write 8 blocks in inode table starting at 360450: Attempt to write block from filesystem resulted in short write
> ===
> mkfs.xfs: pwrite64 failed: No space left on device
> meta-data=/dev/sda7              isize=256    agcount=16, agsize=152742 blks
>          =                       sectsz=512   attr=0
> data     =                       bsize=4096   blocks=2443872, imaxpct=25
>          =                       sunit=0      swidth=0 blks, unwritten=1
> naming   =version 2              bsize=4096  
> log      =internal log           bsize=4096   blocks=2560, version=1
>          =                       sectsz=512   sunit=0 blks
> realtime =none                   extsz=65536  blocks=0, rtextents=0
> ===
> mkfs.reiserfs 3.6.19 (2003 www.namesys.com)
> [...]
> Guessing about desired format.. Kernel 2.6.24-rc2-mm1-autokern1 is running.
> Format 3.6 with standard journal
> Count of blocks on the device: 2443872
> Number of blocks consumed by mkreiserfs formatting process: 8286
> Blocksize: 4096
> Hash function used to sort names: "r5"
> Journal Size 8193 blocks (first block 18)
> Journal Max transaction length 1024
> inode generation number: 0
> UUID: e9aa2dc4-dfc3-47e8-865b-693f28eac2e5
> Initializing journal - 0%....20%....40%....60%....80%....100%
> bwrite: write 4096 bytes returned -1 (block=360448, dev=3): No space left on device

It was mm-fix-blkdev-size-calculation-in-generic_write_checks.patch.

Odd, I thought that looked OK.


Here's a revert (uploaded to hot-fixes/, too):

--- a/mm/filemap.c~revert-mm-fix-blkdev-size-calculation-in-generic_write_checks
+++ a/mm/filemap.c
@@ -1855,11 +1855,9 @@ inline int generic_write_checks(struct f
 	} else {
 #ifdef CONFIG_BLOCK
 		loff_t isize;
-		unsigned int blksize;
 		if (bdev_read_only(I_BDEV(inode)))
 			return -EPERM;
-		blksize = block_size(I_BDEV(inode));
-		isize = i_size_read(inode) & ~(blksize - 1);
+		isize = i_size_read(inode);
 		if (*pos >= isize) {
 			if (*count || *pos > isize)
 				return -ENOSPC;
_