I have now tested the latest version of YAFFS2 from the CVS repository and am
still getting the same bug with the file system in linux 2.6.20.4. I have
tested other filesystems, ext2, and some network shares, and I do not have any
issues with the number of files in a directory. This is exclusive to my YAFFS2
mount point. I have managed to "fix" the problem, but I'm pretty sure it's not
the appropriate solution. Here's some more details of my test and results.
> I'm having some issues with YAFFS and directories with quite a few files in them
> (110+ files). When I do an ls in the directory the system locks up and
> eventually ls dies with an out of memory error.
I'm testing the file system with the following script:
#!/bin/sh
getfiles(){
FILES=`ls | wc -l`
echo "FILES=[$FILES]"
}
generate_file(){
dd if=/dev/zero of=file$1.txt bs=128 count=1 > /dev/null 2>&1
}
NUM=100000
echo working in $1
cd $1
while true
do
FILENUM=`echo $NUM | cut -c 2-`
echo "--- Generating file $FILENUM ---"
generate_file "$FILENUM"
echo "--- Running ls ---"
getfiles
NUM=`expr $NUM + 1`
echo " "
echo " "
done
I checked out the latest version of yaffs2 and patched it into my kernel and ran
the code with the same results. Here's a trimmed up version of the log from
this script.
# sh filelimit_test.sh /mnt/nand/test/tempdir
working in /mnt/nand/test/tempdir
--- Generating file 00000 ---
--- Running ls ---
FILES=[ 1]
--- Generating file 00001 ---
--- Running ls ---
FILES=[ 2]
--- Generating file 00002 ---
--- Running ls ---
FILES=[ 3]
--- Generating file 00003 ---
--- Running ls ---
FILES=[ 4]
--- Generating file 00004 ---
--- Running ls ---
FILES=[ 5]
--- Generating file 00005 ---
--- Running ls ---
FILES=[ 6]
--- Generating file 00006 ---
--- Running ls ---
FILES=[ 7]
--- Generating file 00007 ---
--- Running ls ---
FILES=[ 8]
--- Generating file 00008 ---
--- Running ls ---
FILES=[ 9]
--- Generating file 00009 ---
--- Running ls ---
FILES=[ 10]
.
.
.
--- Generating file 00123 ---
--- Running ls ---
FILES=[ 124]
--- Generating file 00124 ---
--- Running ls ---
FILES=[ 125]
--- Generating file 00125 ---
--- Running ls ---
FILES=[ 126]
--- Generating file 00126 ---
--- Running ls ---
FILES=[ 253]
--- Generating file 00127 ---
--- Running ls ---
FILES=[ 254]
--- Generating file 00128 ---
--- Running ls ---
syslogd invoked oom-killer: gfp_mask=0x201d2, order=0, oomkilladj=0
Mem-info:
DMA per-cpu:
CPU 0: Hot: hi: 0, btch: 1 usd: 0 Cold: hi: 0, btch: 1 usd: 0
Active:4922 inactive:962 dirty:0 writeback:0 unstable:0 free:179 slab:791
mapped:96 pagetables:37
DMA free:716kB min:716kB low:892kB high:1072kB active:19688kB inactive:3848kB
present:32064kB pages_scanned:36201 all_unreclaimable? yes
lowmem_reserve[]: 0 0
DMA: 1*4kB 1*8kB 0*16kB 0*32kB 1*64kB 1*128kB 0*256kB 1*512kB 0*1024kB 0*2048kB
0*4096kB = 716kB
Free swap: 0kB
8192 pages of RAM
251 free pages
1052 reserved pages
791 slab pages
310 pages shared
0 pages swap cached
Out of memory: kill process 7455 (ls) score 222 or a child
Killed process 7455 (ls)
klogd invoked oom-killer: gfp_mask=0x201d2, order=0, oomkilladj=0
Mem-info:
DMA per-cpu:
CPU 0: Hot: hi: 0, btch: 1 usd: 0 Cold: hi: 0, btch: 1 usd: 0
Active:4922 inactive:962 dirty:0 writeback:0 unstable:0 free:179 slab:791
mapped:96 pagetables:37
DMA free:716kB min:716kB low:892kB high:1072kB active:19688kB inactive:3848kB
present:32064kB pages_scanned:36201 all_unreclaimable? yes
lowmem_reserve[]: 0 0
DMA: 1*4kB 1*8kB 0*16kB 0*32kB 1*64kB 1*128kB 0*256kB 1*512kB 0*1024kB 0*2048kB
0*4096kB = 716kB
Free swap: 0kB
8192 pages of RAM
251 free pages
1052 reserved pages
791 slab pages
311 pages shared
0 pages swap cached
syslogd invoked oom-killer: gfp_mask=0x201d2, order=0, oomkilladj=0
Mem-info:
DMA per-cpu:
CPU 0: Hot: hi: 0, btch: 1 usd: 0 Cold: hi: 0, btch: 1 usd: 0
Active:4922 inactive:962 dirty:0 writeback:0 unstable:0 free:179 slab:791
mapped:96 pagetables:37
DMA free:716kB min:716kB low:892kB high:1072kB active:19688kB inactive:3848kB
present:32064kB pages_scanned:36201 all_unreclaimable? yes
lowmem_reserve[]: 0 0
DMA: 1*4kB 1*8kB 0*16kB 0*32kB 1*64kB 1*128kB 0*256kB 1*512kB 0*1024kB 0*2048kB
0*4096kB = 716kB
Free swap: 0kB
8192 pages of RAM
251 free pages
1052 reserved pages
791 slab pages
311 pages shared
0 pages swap cached
ls invoked oom-killer: gfp_mask=0x201d2, order=0, oomkilladj=0
Mem-info:
DMA per-cpu:
CPU 0: Hot: hi: 0, btch: 1 usd: 0 Cold: hi: 0, btch: 1 usd: 0
Active:4922 inactive:962 dirty:0 writeback:0 unstable:0 free:179 slab:791
mapped:96 pagetables:37
DMA free:716kB min:716kB low:892kB high:1072kB active:19688kB inactive:3848kB
present:32064kB pages_scanned:36201 all_unreclaimable? yes
lowmem_reserve[]: 0 0
DMA: 1*4kB 1*8kB 0*16kB 0*32kB 1*64kB 1*128kB 0*256kB 1*512kB 0*1024kB 0*2048kB
0*4096kB = 716kB
Free swap: 0kB
8192 pages of RAM
251 free pages
1052 reserved pages
791 slab pages
311 pages shared
0 pages swap cached
syslogd invoked oom-killer: gfp_mask=0x201d2, order=0, oomkilladj=0
Mem-info:
DMA per-cpu:
CPU 0: Hot: hi: 0, btch: 1 usd: 0 Cold: hi: 0, btch: 1 usd: 0
Active:4923 inactive:961 dirty:0 writeback:0 unstable:0 free:178 slab:791
mapped:97 pagetables:37
DMA free:712kB min:716kB low:892kB high:1072kB active:19692kB inactive:3844kB
present:32064kB pages_scanned:36201 all_unreclaimable? yes
lowmem_reserve[]: 0 0
DMA: 0*4kB 1*8kB 0*16kB 0*32kB 1*64kB 1*128kB 0*256kB 1*512kB 0*1024kB 0*2048kB
0*4096kB = 712kB
Free swap: 0kB
8192 pages of RAM
250 free pages
1052 reserved pages
791 slab pages
313 pages shared
0 pages swap cached
klogd invoked oom-killer: gfp_mask=0x201d2, order=0, oomkilladj=0
Mem-info:
DMA per-cpu:
CPU 0: Hot: hi: 0, btch: 1 usd: 0 Cold: hi: 0, btch: 1 usd: 0
Active:4923 inactive:961 dirty:0 writeback:0 unstable:0 free:178 slab:791
mapped:97 pagetables:37
DMA free:712kB min:716kB low:892kB high:1072kB active:19692kB inactive:3844kB
present:32064kB pages_scanned:36201 all_unreclaimable? yes
lowmem_reserve[]: 0 0
DMA: 0*4kB 1*8kB 0*16kB 0*32kB 1*64kB 1*128kB 0*256kB 1*512kB 0*1024kB 0*2048kB
0*4096kB = 712kB
Free swap: 0kB
8192 pages of RAM
250 free pages
1052 reserved pages
791 slab pages
313 pages shared
0 pages swap cached
syslogd invoked oom-killer: gfp_mask=0x201d2, order=0, oomkilladj=0
Mem-info:
DMA per-cpu:
CPU 0: Hot: hi: 0, btch: 1 usd: 0 Cold: hi: 0, btch: 1 usd: 0
Active:4923 inactive:961 dirty:0 writeback:0 unstable:0 free:178 slab:791
mapped:97 pagetables:37
DMA free:712kB min:716kB low:892kB high:1072kB active:19692kB inactive:3844kB
present:32064kB pages_scanned:36201 all_unreclaimable? yes
lowmem_reserve[]: 0 0
DMA: 0*4kB 1*8kB 0*16kB 0*32kB 1*64kB 1*128kB 0*256kB 1*512kB 0*1024kB 0*2048kB
0*4096kB = 712kB
Free swap: 0kB
8192 pages of RAM
250 free pages
1052 reserved pages
791 slab pages
313 pages shared
0 pages swap cached
Killed
FILES=[ 0]
You'll notice that the file count goes funny for two exections of ls where it
doubly lists all the files, then it gets stuck in an infinite loop that
eventually crashes ls.
> It seems the action of rewinding the directory if the file pointer and inode
> pointer versions differ is not the correct solution. I have added trace code
> and confirmed that when the following if statement is entered the directory
> listing starts from the beginning.
>
> if (f->f_version != inode->i_version) {
> offset = 2;
> f->f_pos = offset;
> f->f_version = inode->i_version;
> }
>
If I change the previous code to not rewind the directory when f->f_version is
not equal to inode->i_version:
if (f->f_version != inode->i_version) {
//offset = 2;
f->f_pos = offset;
f->f_version = inode->i_version;
}
Then I don't have the problem with ls
sh filelimit_test.sh /mnt/nand/test/tempdir
working in /mnt/nand/test/tempdir
--- Generating file 00000 ---
--- Running ls ---
FILES=[ 1]
--- Generating file 00001 ---
--- Running ls ---
FILES=[ 2]
--- Generating file 00002 ---
--- Running ls ---
FILES=[ 3]
--- Generating file 00003 ---
--- Running ls ---
FILES=[ 4]
--- Generating file 00004 ---
--- Running ls ---
FILES=[ 5]
--- Generating file 00005 ---
--- Running ls ---
FILES=[ 6]
--- Generating file 00006 ---
--- Running ls ---
FILES=[ 7]
--- Generating file 00007 ---
--- Running ls ---
FILES=[ 8]
--- Generating file 00008 ---
--- Running ls ---
FILES=[ 9]
--- Generating file 00009 ---
--- Running ls ---
FILES=[ 10]
.
.
.
--- Generating file 00494 ---
--- Running ls ---
FILES=[ 495]
--- Generating file 00495 ---
--- Running ls ---
FILES=[ 496]
--- Generating file 00496 ---
--- Running ls ---
FILES=[ 497]
--- Generating file 00497 ---
--- Running ls ---
FILES=[ 498]
--- Generating file 00498 ---
--- Running ls ---
FILES=[ 499]
--- Generating file 00499 ---
--- Running ls ---
FILES=[ 500]
I'm hoping the added information will help understanding of this problem. Has
anyone seen anything like this with YAFFS2 and the Linux kernel? I'm going to
test my change with my other software developers and see if they're going to
have any issues with the fix I did. We're using YAFFS2 in an application where
we get 100+ files in a directory, and it has been having problems.