dd ENOMEM problems

2002-11-23T01:22:46Z

Reported by Claudio Martella

this is what i get:

junior:~# dd if=/dev/cdrom of=/dev/null
dd: opening `/dev/cdrom': Cannot allocate memory

junior:~# strace dd if=/dev/hdc of=/dev/null bs=1024
execve("/bin/dd", ["dd", "if=/dev/hdc", "of=/dev/null",
"bs=1024"], [/*
31 vars */]) = 0
uname({sys="Linux", node="junior", ...}) = 0
brk(0) = 0x804fcd8
access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No
such file or
directory)
open("/etc/ld.so.preload", O_RDONLY) = -1 ENOENT (No
such file or
directory)
open("/etc/ld.so.cache", O_RDONLY) = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=57835, ...}) = 0
old_mmap(NULL, 57835, PROT_READ, MAP_PRIVATE, 3, 0) =
0x40012000
close(3) = 0
open("/lib/libc.so.6", O_RDONLY) = 3
read(3,
"\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0@Z\1\000"...,
1024) = 1024
fstat64(3, {st_mode=S_IFREG|0755, st_size=1109900,
...}) = 0
old_mmap(NULL, 1122692, PROT_READ|PROT_EXEC,
MAP_PRIVATE, 3, 0) =
0x40021000
mprotect(0x4012a000, 37252, PROT_NONE) = 0
old_mmap(0x4012a000, 20480, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED,
3, 0x109000) = 0x4012a000
old_mmap(0x4012f000, 16772, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x4012f000
close(3) = 0
open("/lib/libc.so.6", O_RDONLY) = 3
read(3,
"\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0@Z\1\000"...,
1024) = 1024
fstat64(3, {st_mode=S_IFREG|0755, st_size=1109900,
...}) = 0
old_mmap(NULL, 1122692, PROT_READ|PROT_EXEC,
MAP_PRIVATE, 3, 0) =
0x40021000
mprotect(0x4012a000, 37252, PROT_NONE) = 0
old_mmap(0x4012a000, 20480, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED,
3, 0x109000) = 0x4012a000
old_mmap(0x4012f000, 16772, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x4012f000
close(3) = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_ANONYMOUS,
-1, 0) = 0x40134000
munmap(0x40012000, 57835) = 0
brk(0) = 0x804fcd8
brk(0x8050cd8) = 0x8050cd8
brk(0x8051000) = 0x8051000
close(0) = 0
open("/dev/hdc", O_RDONLY|O_LARGEFILE) = -1 ENOMEM
(Cannot allocate
memory)
write(2, "dd: ", 4dd: ) = 4
write(2, "opening `/dev/hdc\'", 18opening `/dev/hdc')
= 18
write(2, ": Cannot allocate memory", 24: Cannot
allocate memory) = 24
write(2, "\n", 1
) = 1
_exit(1) = ?
junior:~# free
total used free shared
buffers
cached
Mem: 126508 117052 9456
0
5796 55684
-/+ buffers/cache: 55572 70936
Swap: 151192 78200 72992

junior:~# cat /proc/meminfo
total: used: free: shared: buffers: cached:
Mem: 129544192 125943808 3600384 0
7217152 69705728
Swap: 154820608 78082048 76738560
MemTotal: 126508 kB
MemFree: 3516 kB
MemShared: 0 kB
Buffers: 7048 kB
Cached: 56000 kB
CCacheAlloc: 14952 kB
CCacheUsed: 4162504 kB
SwapCached: 12072 kB
Active: 64032 kB
Inactive: 33272 kB
HighTotal: 0 kB
HighFree: 0 kB
LowTotal: 126508 kB
LowFree: 3516 kB
SwapTotal: 151192 kB
SwapFree: 74940 kB

junior:~# uname -a
Linux junior 2.4.19-ck10 #2 Tue Oct 15 14:31:43 CEST
2002 i686 unknown
unknown GNU/Linux

junior:/usr/src# cat /proc/comp_cache_*
compressed cache - fragmentation histogram (free space
x fragmented space)
total <1000 -2000 -3000 -4000 -5000
-6000 -7000 -8000 8192
1 - 400: 1639 1522 91 22 4 0 0
0 0 0
401 - 800: 118 111 7 0 0 0 0
0 0 0
801 - 1200: 64 62 2 0 0 0 0
0 0 0
1201 - 1600: 30 29 0 0 0 1 0
0 0 0
1601 - 2000: 19 19 0 0 0 0 0
0 0 0
2001 - 2400: 9 9 0 0 0 0 0
0 0 0
2401 - 2800: 3 3 0 0 0 0 0
0 0 0
2801 - 3200: 0 0 0 0 0 0 0
0 0 0
3201 - 3600: 1 1 0 0 0 0 0
0 0 0
3601 - 4000: 0 0 0 0 0 0 0
0 0 0
4001 - 4400: 0 0 0 0 0 0 0
0 0 0
4401 - 4800: 0 0 0 0 0 0 0
0 0 0
4801 - 5200: 0 0 0 0 0 0 0
0 0 0
5201 - 5600: 0 0 0 0 0 0 0
0 0 0
5601 - 6000: 0 0 0 0 0 0 0
0 0 0
6001 - 6400: 0 0 0 0 0 0 0
0 0 0
6401 - 6800: 0 0 0 0 0 0 0
0 0 0
6801 - 7200: 0 0 0 0 0 0 0
0 0 0
7201 - 7600: 0 0 0 0 0 0 0
0 0 0
7601 - 8000: 0 0 0 0 0 0 0
0 0 0
8001 - 8192: 0 0 0 0 0 0 0
0 0 0
compressed cache - free space histogram (free space x
number of fragments)
total 0f 1f 2f 3f 4f
5f 6f more
total 9149 act 3743 pages 3774
0: 4 0 0 1 0 2
1 0 0
1 - 400: 1639 0 0 2 210 575
404 235 213
401 - 800: 118 0 0 1 28 34
27 14 14
801 - 1200: 64 0 0 3 19 19
14 3 6
1201 - 1600: 30 0 0 3 7 13
3 3 1
1601 - 2000: 19 0 0 1 6 7
3 2 0
2001 - 2400: 9 0 0 0 3 6
0 0 0
2401 - 2800: 3 0 0 0 1 1
1 0 0
2801 - 3200: 0 0 0 0 0 0
0 0 0
3201 - 3600: 1 0 0 0 1 0
0 0 0
3601 - 4000: 0 0 0 0 0 0
0 0 0
4001 - 4400: 0 0 0 0 0 0
0 0 0
4401 - 4800: 0 0 0 0 0 0
0 0 0
4801 - 5200: 0 0 0 0 0 0
0 0 0
5201 - 5600: 0 0 0 0 0 0
0 0 0
5601 - 6000: 0 0 0 0 0 0
0 0 0
6001 - 6400: 0 0 0 0 0 0
0 0 0
6401 - 6800: 0 0 0 0 0 0
0 0 0
6801 - 7200: 0 0 0 0 0 0
0 0 0
7201 - 7600: 0 0 0 0 0 0
0 0 0
7601 - 8000: 0 0 0 0 0 0
0 0 0
8001 - 8192: 0 0 0 0 0 0
0 0 0
compressed cache - statistics
general
- (S) swap cache support enabled
- (P) page cache support enabled
- compressed swap support enabled
- maximum used size: 16384 KiB
- comp page size: 8 KiB
- failed allocations: 742
algorithm LZO
- (C) compressed pages: 4134129 (S: 3% P: 96%)
- (D) decompressed pages: 1059944 (S: 8% P: 91%)
D/C: 25%
- (R) read pages: 1012456 (S: 2% P: 97%)
R/C: 24%
- (W) written pages: 55168 (S: 100% P: 0%)
W/C: 1%
compression ratio: 51% (S: 32% P: 51%)
junior:/usr/src# vmstat
procs memory swap
io system cpu
r b w swpd free buff cache si so bi
bo in cs us sy id
0 0 0 76240 3388 7816 53404 0 0 5
17 16 20 1 0 99

junior:/usr/src# uptime
12:10:33 up 20 days, 22:18, 4 users, load average:
0.10, 0.10, 0.09

after a reboot everything was fine.

Over allocates ram it doesnt have

2002-11-22T13:36:32Z

Repeatedly allocating ram (eg with the mem_load utility
from contest) makes cc allocate more ram than is
available. Swap levels out at 70% usage for a while
then shoots up to 100% and the kernel oopses.

Freeing swap buffer with locked buffers

2002-08-07T18:25:11Z

There is a potential bug in refill_swp_buffer() that
can free a swap buffer pages that has buffers.

Corruption after a failed writepage()

2002-08-07T18:00:43Z

After a failed writepage(), the swap buffer would not
set the fragment->swp_buffer to NULL, so it might end
having its fragment field nulled in the future,
possibly when set to a fragment that is being written.
If that happens, the fragment would not be freed and
never freed by the writeout code, since it had been
removed from the LRU queue.

Error decompressing fragment

2002-08-07T17:57:55Z

Once, compiling the linux kernel on a host kernel with
compressed cache, we hit a bug from a error happened
when decompressing a fragment (in the case, the error
was -4 in LZO).

Deadlock in Kernel compilation

2002-08-07T17:55:28Z

A deadlock happens when compiling the linux kernel on a
host kernel with compressed cache. It usually happens
when used make -j 2/4, but also happens with make -j 1.
From the stack traces, a process hangs waiting for a
swap buffer page in comp_cache_free_locked() and all
the other processes hang trying to lock the
fragment->comp_page which is locked waiting for the
swap buffer.

Oops in ext2_check_page

2002-07-29T15:27:35Z

We get a page->mapping == NULL, and since it is
accessed without any check in the first line of this
function, it oopses. It looks like the page has been
removed from page cache while flushing a compressed
cache entry.

kernel BUG at inode.c:518

2002-07-29T12:51:34Z

Page Cache
Double Page Size
Preempt

The bug happens when clearing the inode in the exit
path. The fragment have the ToBeFreed bit set and also
an null counter.

RAMFs won't work with compressed cache

2002-07-26T21:08:46Z

RAMFS will not work properly with compressed cache
since pages from this filesystem might end up being
freed even if they have not made their way to the disk.
In this case, of course there is no disk, but since we
don't check if it has been written (what will never
happen to RAMFS), pages from this FS will be freed anyway.

Freeze when page cache is enabled

2002-07-26T11:28:28Z

When such a test like fillmem, mmap001 and dbench are
run many times concurrently, the system freezes. That
only happens when page cache support is enabled (it
doesn't matter if on demand resizing or double page
size are enabled).

Recent changes to bugs