as someone told me i use now linux 2.1.125 with one nfs client and one server.
I use gcc 2.7.2.1 for PowerPC at user level and egcs-2.91.52 for the kernel.
I use the glibc : ls -l /lib/libc* :
-rwxr-xr-x 1 root root 838471 Jan 12 1998 /lib/libc-1.99.so
lrwxrwxrwx 1 root root 12 Sep 2 16:29 /lib/libc.so.6 -> libc-1.99.so
I installed the knfsd-981014 package, compiled the programs, and even run the
testlk that seem to be ok.
I now have the radiance imaging system running on the two machines, and an image
computing can be splited on the two machines.
When i run only one process on the server, the image is computed entierly
but when i try to run the same program on the NFS client, it never ends
and some processes are blocked :
ps axu :
USER PID %CPU %MEM SIZE RSS TTY STAT START TIME COMMAND
delchini 281 0.0 1.1 2604 700 p1 S 16:36 0:00 sh ./stand.fast
delchini 282 0.0 1.4 1856 900 p1 S 16:36 0:00 rpiece -F stand.sync -af stand.illum -t 60 -vp 1260.45 -1262.271 120.0 -v
delchini 283 14.3 4.2 3988 2676 p1 S N 16:36 1:27 rpict -S 1 -af stand.illum -t 60 -ps 6 -pt .08 -dp 512 -ar 16 -ms 75 -ds
delchini 287 0.0 1.4 1856 900 p1 D 16:37 0:00 rpiece -F stand.sync -af stand.illum -t 60 -vp 1260.45 -1262.271 120.0 -v
delchini 288 0.0 1.4 1856 900 p1 D 16:37 0:00 rpiece -F stand.sync -af stand.illum -t 60 -vp 1260.45 -1262.271 120.0 -v
delchini 292 0.0 1.4 1856 900 p1 D 16:38 0:00 rpiece -F stand.sync -af stand.illum -t 60 -vp 1260.45 -1262.271 120.0 -v
see the ones with the "D" status.
I applied the nfsd-2.1.125 patch to the kernel, what could i do more ?
Here is the syscall trace for the "rpict" process that actually compute the
image :
.....
mmap(0x7ffff368, 0, PROT_NONE, 0 /* MAP_??? */, 0, 0) = 0x3000e000
munmap(0x3000e000, 4096) = 0
mmap(0x7ffff368, 0, PROT_NONE, 0 /* MAP_??? */, 0, 0) = 0x3004f000
mprotect(0x3006c000, 285880, PROT_NONE) = 0
mmap(0, 0, PROT_NONE, 0 /* MAP_??? */, 0, 0) = 0x3009f000
close(3) = 0
open("/lib/libc.so.6", O_RDONLY) = 3
mmap(0x7ffff368, 0, PROT_NONE, 0 /* MAP_??? */, 0, 0) = 0x3000e000
munmap(0x3000e000, 4096) = 0
mmap(0x7ffff368, 0, PROT_NONE, 0 /* MAP_??? */, 0, 0) = 0x300b2000
mprotect(0x30149000, 368340, PROT_NONE) = 0
mmap(0, 0, PROT_NONE, 0 /* MAP_??? */, 0, 0) = 0x30182000
mmap(0, 0, PROT_NONE, 0 /* MAP_??? */, 0, 0) = 0x30195000
close(3) = 0
mprotect(0x3004f000, 118784, PROT_READ|PROT_WRITE) = 0
mprotect(0x3004f000, 118784, PROT_READ|PROT_EXEC) = 0
time(NULL) = 908807979
brk(0) = 0x18c8318
brk(0) = 0x18c8318
brk(0x18c9000) = 0x18c9000
brk(0) = 0x18c9000
brk(0x18cb000) = 0x18cb000
sigaction(SIGINT, {0x1803208, [], SA_RESTART}, {SIG_DFL}) = 0
sigaction(SIGHUP, {0x1803208, [], SA_RESTART}, {SIG_DFL}) = 0
sigaction(SIGTERM, {0x1803208, [], SA_RESTART}, {SIG_DFL}) = 0
sigaction(SIGPIPE, {0x1803208, [], SA_RESTART}, {SIG_DFL}) = 0
sigaction(SIGALRM, {0x1803208, [], SA_RESTART}, {SIG_DFL}) = 0
sigaction(SIGXCPU, {0x1803208, [], SA_RESTART}, {SIG_DFL}) = 0
sigaction(SIGXFSZ, {0x1803208, [], SA_RESTART}, {SIG_DFL}) = 0
nice(6) = 0
fstat(0, {st_mode=S_IFCHR|0620, st_rdev=makedev(3, 2), ...}) = 0
mmap(0x7ffff368, 0, PROT_NONE, 0 /* MAP_??? */, 0, 0) = 0x3000e000
ioctl(0, 0x402c7413, 0x7ffff328) = 0
read(0, <unfinished ...>
For the others (rpiece with "D" status), the syscall trace gives nothing, i mean
that strace -p PID prints nothing.
I saw that a patch for the statd was released, i tried it but nothing changed.
Thanx for any help.
Ugo.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/