fxp0 and vlan panic

[othermark]

Randy Bush wrote:
>> Just for the record, and for people not reading CVS commit logs, I
>> committed a fix for this a few days ago and I will make sure to MFC
>> it in time for 5.4-RELEASE.
> 
> might this give me some help on occasional but repeated fxp
> crashes under load?
> 
> Fatal trap 12: page fault while in kernel mode
> fault virtual address   = 0x80808517
> fault code              = supervisor read, page not present
> instruction pointer     = 0x8:0xc047d2d0
> stack pointer           = 0x10:0xd3f78c88
> frame pointer           = 0x10:0xd3f78cac
> code segment            = base 0x0, limit 0xfffff, type 0x1b
>                         = DPL 0, pres 1, def32 1, gran 1
> processor eflags        = interrupt enabled, resume, IOPL = 0
> current process         = 15 (irq5: fxp0)
> [thread pid 15 tid 100008 ]
> Stopped at      fxp_intr_body+0xd0:     cmpw    $0,0(%esi)
> db> trace
> Tracing pid 15 tid 100008 td 0xc155fb80
> fxp_intr_body(c161a000,c161a000,40,ffffffff,c0630cb6) at
> fxp_intr_body+0xd0 fxp_intr(c161a000,0,0,0,0) at fxp_intr+0x141
> ithread_loop(c1551a00,d3f78d48,0,0,0) at ithread_loop+0x1a8
> fork_exit(c04da530,c1551a00,d3f78d48) at fork_exit+0x7f
> fork_trampoline() at fork_trampoline+0x8
> --- trap 0x1, eip = 0, esp = 0xd3f78d7c, ebp = 0 ---

Hmm, can you get the panic to occur when interrupt coalescing is turned on?
'ifconfig fxp0 link0'

I've loaded my -current boxes pretty hard and have not seen this panic, but
I have interrupt coalescing turned on everywhere.

-- 
othermark
atkin901 at nospam dot yahoo dot com
(!wired)?(coffee++):(wired);