NVE

[Kövesdán Gábor]

Maxime Henrion wrote:

>That's the plan.  I'm actually waiting for feedback to know whether this
>helps make things better or not.  I'll ask re@ for permission to MFC it
>once I know it's good (I have no nve(4) hardware myself).
>
>  
>
Well... I had an another panic but it's a bit different from the 
previous panics I've got
with nve (http://www.freebsd.org/cgi/query-pr.cgi?pr=83943),but it is 
quite similar to
an another panic I've recently had on RELENG_6. 
(http://www.freebsd.org/cgi/query-pr.cgi?pr=84133)
I took a piece of paper again and wrote down the trace output:

panic: nve_ifstart: attempted use of a free mbuf!
KDB: enter: panic
[ thread pid 555 tid 100068 ]
Stopped at   kdb_enter+0x2f:nop
db> trace
Tracing pid 555 tid 100068 td 0xffffff001ed2ea00
kdb_enter() at kdb_enter+0x2f
panic() at panic+0x1d2
nve_ifstart() at nve_ifstart+0x3a5
ether_output_frame() at ether_output_frame+0x249
ether_output() at ether_output+0x349
ip_output() at ip_output+0xb17
tcp_output() at tcp_output+0x9c2
tcp_usr_rcvd() at tcp_usr_rcvd+0x70
soreceive() at soreceive+0x1009
soo_read() at soo_read+0xc0
dofileread() at dofileread+0x94
kern_readv() at kern_readv+0x60
read() at read+0x4a
syscall() at syscall+0x4b2
Xfast_syscall() at Xfast_syscall+0xa8
--- syscall (3, FreeBSD ELF64, read), rip = 0x80095d84c, rsp = 0x6d5530, 
rbp = 0x6d55b8 ---


Look at the second PR I mentioned. It is a general protection fault, but 
the trace output is quite similar.
They must be in relation.

Cheers,

Gábor Kövesdán