Re: panic: mb_dtor_pack: ref_cnt != 1

Le Friday 4 November 2005 19:30, Andre Oppermann a écrit :
>
> Please try the attached patch to see whether it makes a difference.

[SNIP]

with the same test (ls -lR /share_disk) :
the mbufs are better managed :
...
70/63680/63750 mbufs in use (current/cache/total)
64/404/468/3776 mbuf clusters in use (current/cache/total/max)
65/65095/65160 mbufs in use (current/cache/total)
64/404/468/3776 mbuf clusters in use (current/cache/total/max)

but :

panic: kmem_malloc(200704): kmem_map too small: 29900800 total allocated
cpuid = 1
KDB: enter: panic
[thread pid 730 tid 100096 ]
Stopped at      kdb_enter+0x2b: nop
db> trace
Tracing pid 730 tid 100096 td 0xc1595600
kdb_enter(c086a633) at kdb_enter+0x2b
panic(c088597c,31000,1c84000,c19afd80,31000) at panic+0x127
kmem_malloc(c10600c0,31000,102,cb63cabc,c079047b) at kmem_malloc+0x89
page_alloc(0,31000,cb63caaf,102,29332e0) at page_alloc+0x1a
uma_large_malloc(31000,102,c0636662,be,30fd1) at uma_large_malloc+0x3b
malloc(30fd1,c08cf1a0,102,c0652867,c0934a20) at malloc+0xf5
sysctl_kern_malloc_stats(c08cf520,0,0,cb63cb98,c08cf520) at 
sysctl_kern_malloc_stats+0x76
sysctl_root(0,cb63cc18,2,cb63cb98,c1595600) at sysctl_root+0x11b
userland_sysctl(c1595600,cb63cc18,2,8077000,bfbfea18) at userland_sysctl+0xf4
__sysctl(c1595600,cb63cd04,c,c1595600,cb63cd30) at __sysctl+0x77
syscall(2809003b,2809003b,bfbf003b,2,bfbfea18) at syscall+0x27e
Xint0x80_syscall() at Xint0x80_syscall+0x1f
--- syscall (202, FreeBSD ELF32, __sysctl), eip = 0x281494ef, esp = 
0xbfbfe8cc, ebp = 0xbfbfe908 ---
db> 

	TfH