[ELDK] Native compiling freezes with swap over nfs!
Bastian
ng at max01.eu
Sun Nov 29 14:59:31 CET 2009
Hello,
while native compiling a source code which makes use of libtool
on an arm target (OMAP-L138) i got the following error:
libtool: compile: g++ -DHAVE_CONFIG_H -I. -I../..
-I/usr/local/lib/sigc++-1.2/include -I/usr/local/include/sigc++-1.2
-I/usr/include/SDL -D_GNU_SOURCE=1 -D_REENTRANT -I./../physfs
-I../../include -g -O2 -Wall -I/freetype2 -I -fno-exceptions -MT
pgfilearchive.lo -MD -MP -MF .deps/pgfilearchive.Tpo -c
pgfilearchive.cpp -fPIC -DPIC -o .libs/pgfilearchive.o
cc1plus invoked oom-killer: gfp_mask=0x201da, order=0, oom_adj=0
[<c002b8f0>] (unwind_backtrace+0x0/0xe0) from [<c006dea4>]
(oom_kill_process+0x7c/0x200)
[<c006dea4>] (oom_kill_process+0x7c/0x200) from [<c006e418>]
(__out_of_memory+0x160/0x184) [...]
It looks like there is not enough memory.
So i tried to mount a swap - file over nfs and do the following steps:
insmod /lib/modules/2.6.32-rc8-06953-g7b20551/kernel/drivers/block/loop.ko
mount -t nfs -o rsize=2048,wsize=2048 192.168.2.16:/home/lumax/.swap
/mnt/swap
mkswap /mnt/swap/swapx
/*swapx is a file created with a cmd like:
dd if=/dev/zero of=swapfile bs=1M count=128*/
losetup /dev/loop0 /mnt/swap/swapx
mkswap /dev/loop0
swapon /dev/loop0
cat /proc/meminfo
MemTotal: 29328 kB
MemFree: 20872 kB
Buffers: 4 kB
Cached: 4336 kB
SwapCached: 0 kB
Active: 2400 kB
Inactive: 3032 kB
Active(anon): 212 kB
Inactive(anon): 880 kB
Active(file): 2188 kB
Inactive(file): 2152 kB
Unevictable: 0 kB
Mlocked: 0 kB
SwapTotal: 131064 kB
SwapFree: 131064 kB
Dirty: 4 kB
Writeback: 0 kB
AnonPages: 1108 kB
Mapped: 1308 kB
Shmem: 0 kB
Slab: 2040 kB
SReclaimable: 448 kB
SUnreclaim: 1592 kB
KernelStack: 216 kB
PageTables: 116 kB
NFS_Unstable: 0 kB
Bounce: 0 kB
WritebackTmp: 0 kB
CommitLimit: 145728 kB
Committed_AS: 3304 kB
VmallocTotal: 985088 kB
VmallocUsed: 152 kB
VmallocChunk: 984864 kB
There seems to be swap-memory,now!
Compiling the code again, freeze!
Once i interrupted the compiling process after a while i
got the folowing message:
^CINFO: task rpcbind:1042 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[<c022991c>] (schedule+0x2f0/0x374) from [<c0229f14>]
(io_schedule+0x48/0x7c)
[<c0229f14>] (io_schedule+0x48/0x7c) from [<c006b15c>] (sync_page+0x4c/0x5c)
[...]
For me its looks like there are still problems with mounting the swap-file
over nfs.
What's to do to stop the system blocking?
What's to do to get a running swap over nfs?
Any Ideas are very welcome!
Regards,
Bastian.
More information about the eldk
mailing list