Messages in this thread | | | Date | Tue, 8 Feb 2000 03:52:50 +0000 (GMT) | From | "imel..." <> | Subject | [BUG] write(2) hangs in 2.2.[13-14-15pre5] + OOPS in 2.3.42 |
| |
hi, rik, et. al.
after hand patching the boobytrap, on 2.2.15pre5, the problem hasn't gone away. but it did run better, usually it goes up to 20megs before it dies. with 2.2.15pre5 it went 198megs and then froze.
another thing, your boobytrap didn't show anything.
then i tried 2.3.42 with uncertainty about initrd, i got this oops.
ksymoops 2.3.3 on i686 2.2.13. Options used -v vmlinux (specified) -K (specified) -L (specified) -O (specified) -m System.map (specified)
EIP: 0010:[<c012b8a1>] Using defaults from ksymoops -t elf32-i386 -a i386 Call Trace: [<c01ba6c8>] [<c010716b>] [<c0107524>] Code: 39 46 18 74 2a 50 8b 46 18 50 68 20 01 1a c0 e8 63 7a fe ff
>>EIP; c012b8a1 <blkdev_open+19/d4> <===== Trace; c01ba6c8 <is_mii+4b4/2d9c> Trace; c010716b <init+7/108> Trace; c0107524 <kernel_thread+28/38> Code; c012b8a1 <blkdev_open+19/d4> 00000000 <_EIP>: Code; c012b8a1 <blkdev_open+19/d4> <===== 0: 39 46 18 cmpl %eax,0x18(%esi) <===== Code; c012b8a4 <blkdev_open+1c/d4> 3: 74 2a je 2f <_EIP+0x2f> c012b8d0 <blkdev_open+48/d4> Code; c012b8a6 <blkdev_open+1e/d4> 5: 50 pushl %eax Code; c012b8a7 <blkdev_open+1f/d4> 6: 8b 46 18 movl 0x18(%esi),%eax Code; c012b8aa <blkdev_open+22/d4> 9: 50 pushl %eax Code; c012b8ab <blkdev_open+23/d4> a: 68 20 01 1a c0 pushl $0xc01a0120 Code; c012b8b0 <blkdev_open+28/d4> f: e8 63 7a fe ff call fffe7a77 <_EIP+0xfffe7a77> c0113318 <printk+0/160>
imel
On Sat, 29 Jan 2000, Rik van Riel wrote: > On Fri, 28 Jan 2000, imel... wrote: > > > i was testing my nbd array with: > > dd if=/dev/urandom of=test count=4000000 > > > > on a raid5 partition. the test hangs one of the nbd client. > > the problem was on the client. doing: > > #strace nbd-server 100 /dev/hda > > > > showed that one of the machine was doing write() and it never > > returns. > > the hda is a 3gb disk. > > With this stress-test you could have hit the problem > window in __get_free_pages(). Could you try 2.2.15pre5 > + my boobytrap2 patch? That one should probably work > and if it doesn't it'll give us the error messages we > need to track down the bug for you and make sure that > the next kernel version will be stable.
- 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/
| |