Re: VFS panics (cvs commit: src/sys/kern vfs_bio.c)

From: Kris Kennaway <kris_at_obsecurity.org>
Date: Fri, 4 Feb 2005 18:40:15 -0800
On Fri, Feb 04, 2005 at 06:19:03PM -0800, Kris Kennaway wrote:
> On Fri, Feb 04, 2005 at 08:27:58PM -0500, Jeff Roberson wrote:
> > This should fix the buf_vlist_remove panic and may fix the dangling deps
> > panic.
> 
> I started 2 fsx processes and they immediately deadlocked:

I think this was running both fsx instances with the same target file.

(null)
0xc5dc1270: tag ufs, type VREG
    usecount 3, writecount 2, refcount 0 mountedhere 0
    flags ()
    v_object 0xc5db5738
     lock type ufs: EXCL (count 1) by thread 0xc5af1450 (pid 644) with 1 pending
        ino 2166788, on dev da0s1e (235, 8)

Looks like process 644 went to sleep holding a lock.

Kris

> db> wh 645
> Tracing pid 645 tid 100195 td 0xc5aea730
> sched_switch(c5aea730,0,1,c06fd89e,c5aea730) at sched_switch+0x2fe
> mi_switch(1,0,c070029f,1ab,c051d45b) at mi_switch+0x41e
> sleepq_switch(c5dc1334,0,c5aea730,600,f11cead0) at sleepq_switch+0x135
> sleepq_wait(c5dc1334,0,c06fd720,da,c5aea730) at sleepq_wait+0x42
> msleep(c5dc1334,c0769c24,50,c0704446,0) at msleep+0x420
> acquire(f11ceb58,1010040,600,ee,c5dc1334) at acquire+0xf3
> debuglockmgr(c5dc1334,1030002,c5dc1270,c5aea730,c0705c42,c0706a4e,c00) at debuglockmgr+0x49f
> ufs_lock(f11cebcc,c0706b7d,333,c0711ed1,c07064e3) at ufs_lock+0x66
> VOP_LOCK_AP(f11cebcc,0,c0706b7d,333,c5dc1270) at VOP_LOCK_AP+0x78
> debug_vn_lock(c5dc1270,20002,c5aea730,c0706a4e,c00) at debug_vn_lock+0x110
> ftruncate(c5aea730,f11ced14,3a6,c0719253,c5aea730) at ftruncate+0x150
> syscall(2f,2f,2f,bfbfe478,40000) at syscall+0x2c4
> Xint0x80_syscall() at Xint0x80_syscall+0x1f
> --- syscall (198, FreeBSD ELF32, nosys), eip = 0x280c6e43, esp = 0xbfbfe37c, ebp = 0xbfbfe3a8 ---
> db> wh 644
> Tracing pid 644 tid 100213 td 0xc5af1450
> sched_switch(c5af1450,0,1,c06fd89e,c5af1450) at sched_switch+0x2fe
> mi_switch(1,0,c070029f,1ab,c051d45b) at mi_switch+0x41e
> sleepq_switch(c4640d28,0,c5af1450,0,f12287bc) at sleepq_switch+0x135
> sleepq_wait(c4640d28,0,c06fd720,da,c5af1450) at sleepq_wait+0x42
> msleep(c4640d28,c07c5da0,244,c0712c27,0) at msleep+0x420
> vm_page_sleep_if_busy(c4640d28,1,c0712c27,6e4,c465fc70) at vm_page_sleep_if_busy+0xc6
> vm_object_page_remove(c5db5738,2c,0,32,0) at vm_object_page_remove+0x12d
> vnode_pager_setsize(c5dc1270,2c000,0,c0579dee,de724dcc) at vnode_pager_setsize+0xec
> ffs_write(f1228940,c06fc0a5,12b,c0711ed1,c0712dca) at ffs_write+0x2f1
> VOP_WRITE_AP(f1228940,0,c0713894,469,86b) at VOP_WRITE_AP+0x78
> vnode_pager_generic_putpages(c5dc1270,f1228ab0,8000,5,f1228a50) at vnode_pager_generic_putpages+0x215
> vop_stdputpages(f12289fc,39d,12b,c0711ed1,c071158e) at vop_stdputpages+0x30
> VOP_PUTPAGES_AP(f12289fc,f12289f8,1,419,c585e000) at VOP_PUTPAGES_AP+0x84
> vnode_pager_putpages(c5db5738,f1228ab0,8,5,f1228a50) at vnode_pager_putpages+0xb0
> vm_pageout_flush(f1228ab0,8,5,364,c0769c24) at vm_pageout_flush+0x16d
> vm_object_page_collect_flush(c5db5738,c46b2f98,586,5,10) at vm_object_page_collect_flush+0x2fc
> vm_object_page_clean(c5db5738,2a,0,32,0) at vm_object_page_clean+0x1b4
> vm_object_sync(c5db5738,2a000,0,8000,1) at vm_object_sync+0x1ff
> vm_map_sync(c5dbd834,2815b000,28163000,1,0) at vm_map_sync+0x29b
> msync(c5af1450,f1228d14,3a6,c0719253,c5af1450) at msync+0x7f
> syscall(2f,2f,2f,75b6,2a2b5) at syscall+0x2c4
> Xint0x80_syscall() at Xint0x80_syscall+0x1f
> --- syscall (65, FreeBSD ELF32, msync), eip = 0x280c76a3, esp = 0xbfbfe38c, ebp = 0xbfbfe3e8 ---
> db>
> 
> Kris



Received on Sat Feb 05 2005 - 01:40:17 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:27 UTC