[issue1031] panic: hammer_mount: malloc limit exceeded
Francois Tigeot
ftigeot at wolfpond.org
Wed Jun 11 23:30:38 PDT 2008
On Wed, Jun 11, 2008 at 03:39:31PM -0700, Matthew Dillon wrote:
> ::After the 53H commit, the panic does'nt occur anymore.
> ::
> ::There's a deadlock instead: after about 5 minutes of copying the machine
> ::stopped responding. No specific message on the console, display frozen.
> ::
> ::I tried to break into the debugger with CTRL-ALT-ESC but it didn't change
> ::anything.
> ::
> ::Francois Tigeot
>
> I believe I have fixed the lockup with commit 54. I have also, I
> believe, completely fixed sequential read performance and have made
> good progress on sequential write performance.
I have been unable to reproduce the lockup.
Howewer, the original panic is back, this time after only about 5 minutes
of copying:
panic: hammer_mount: malloc limit exceeded
mp_lock = 00000001; cpuid = 1
Trace beginning at frame 0xdf1e28d4
panic(df1e28f8,0,c04026e0,ff80848c,df1e2920) at panic + 0x14d
panic(c03a5e6e,c03af5a9,c40f1948,c4772f5c,ff808000) at panic + 0x14d
kmalloc(88,c04026e0,102,0,0) at kmalloc + 0xbb
hammer_alloc_mem_record(e4a0d698,0,6e7c000,0,260a5c42) at hammer_alloc_mem_record + 0x22
hammer_ip_add_bulk(e4a0d698,6e78000,0,ccc30000,4000) at hammer_ip_add_bulk + 0x11c
hammer_vop_strategy(df1e2acc,c03f7148,d3a0db30,c1e72440,6e7b000) at hammer_vop_strategy + 0x64b
vop_strategy(d3a0db30,e2847168,c4608830) at vop_strategy + 0x2c
vn_strategy(e2847168,c4608830,c4608918,c4608918,0) at vn_strategy + 0x50
bwrite(c4608800) at bwrite + 0xfa
bawrite(c4608800) at bawrite + 0x13
hammer_vop_write(df1e2ba4,c03f70b8,d3a0db30,0,0) at hammer_vop_write + 0x3ef
vop_write(d3a0db30,e2847168,df1e2ca0,7f0001,def91948) at vop_write + 0x34
vn_write(de965fd0,df1e2ca0,def91948,0,de965fd0) at vn_write + 0x1e1
kern_pwritev(8,df1e2ca0,0,df1e2cf0) at kern_pwritev + 0xf6
sys_write(df1e2cf0,6,c6415ad,0,da413e18) at sys_write + 0x68
syscall2(df1e240) at syscall2 + 0x25f
Xint0x80_syscall() at Xint0x80_syscall + 0x36
Debugger("panic")
CPU1 stopping CPUs: 0x00000001
stopped
Stopped at Debugger + 0x34: movb $0,in_Debugger.4464
db>
--
Francois Tigeot
More information about the Bugs
mailing list