NaT consumption faults in current 2.5 BK kernel

OK, thanks to the latest mod from David, I have the system running on the simulator. Hurrah.

I can see a NaT consumption fault every time I cat /proc/net/dev

Is there any easy way to track down what's causing this? The offending code section is:

dev_seq_pr*+02a0 ld8 r3=[r15],8

     MII
                          mov              r34=gp
                          mov              r38=r33;;
			  ld8		   r1=[r15]
			  mov		   b7=r3
			  br.call.sptk.many b0·
			  ;;

in the middle of dev_seq_printf_stats().

I suspect a compiler or assembler bug. I can't see anything obviously wrong with the generated code.

The branch to dev_seq_printf+0x2a0 is here:

        alloc r37 = ar.pfs,3,3,8,0
        adds r12 = -96, r12
        mov r11 = r0
        adds r2 = 88, r33
        mov r36 = b0
        .body
	;;
	nop.i 0
	ld8 r15 = [r2]
        nop.f 0
        nop.b 0
        ;;
        cmp.eq p8, p9 = 0, r15
        nop.i 0
        (p9) br.cond.dpnt.few dev_seq_pr*+02a0
	.mii
	adds r15€,r11
	adds r16ˆ,r11
	adds r17–,r11
	...

cat[174]: NaT consumption 17179869216 [4]

Pid: 174, CPU 0, comm: cat
psr : 0000101008026018 ifs : 800000000000030e ip :
[<a000000100266f60>] Not tainted

ip is at dev_seq_printf_stats+0x2a0/0x2e0
unat: 0000000000000000 pfs : 0000000000000309 rsc : 0000000000000003
rnat: 0000000000000000 bsps: a00000010065b830 pr : 000000000006aa19
ldrs: 0000000000000000 ccv : 0000000000000000 fpsr: 0009804c0270033f csd : 0000000000000000 ssd : 0000000000000000
b0 : a000000100266fe0 b6 : a000000100002d70 b7 : a000000100266fa0 f6 : 1003e0000000000004000 f7 : 1003e0000000000000040 f8 : 1003e0000000000000000 f9 : 1003e0000000000000000 f10 : 000000000000000000000 f11 : 000000000000000000000 r1 : a00000010065b830 r2 : e000000007f42858 r3 : 0000000000000002 r8 : e000000007f42800 r9 : 0000000000004000 r10 : 0000000000000143 r11 : 0000000000000000 r12 : e0000000075bfdb0 r13 : e0000000075b8000

Call Trace:
 [<a000000100013a20>] show_stack+0x80/0xa0
                                spà000000075bf910
bspà000000075b9028
 [<a000000100020190>] die+0x130/0x1c0
                                spà000000075bfae0
bspà000000075b9000
 [<a000000100020f70>] ia64_fault+0x150/0xec0
                                spà000000075bfae0
bspà000000075b8fb8
 [<a00000010000d840>] ia64_leave_kernel+0x0/0x260
                                spà000000075bfbe0
bspà000000075b8fb8
 [<a000000100266f60>] dev_seq_printf_stats+0x2a0/0x2e0
                                spà000000075bfdb0
bspà000000075b8f48
 [<a000000100266fe0>] dev_seq_show+0x40/0x100
                                spà000000075bfe10
bspà000000075b8f18
 [<a0000001001041e0>] seq_read+0x4c0/0x760
                                spà000000075bfe10
bspà000000075b8e88
 [<a0000001000c6100>] vfs_read+0x200/0x2a0
                                spà000000075bfe20
bspà000000075b8e40
 [<a0000001000c65e0>] sys_read+0x60/0xc0
                                spà000000075bfe20
bspà000000075b8dc8
 [<a00000010000d6c0>] ia64_ret_from_syscall+0x0/0x20
                                spà000000075bfe30

bspà000000075b8dc8
Segmentation fault

Peter> OK, thanks to the latest mod from David, I have the system Peter> running on the simulator. Hurrah.

Peter> I can see a NaT consumption fault every time I cat Peter> /proc/net/dev

OK, the problem is in the simulated ethernet driver. Trivial Patch appended:

=== arch/ia64/hp/sim/simeth.c 1.8 vs edited ==--- 1.8/arch/ia64/hp/sim/simeth.c Wed Jun 4 20:38:47 2003
+++ edited/arch/ia64/hp/sim/simeth.c Fri Jul 18 09:36:19 2003
@@ -223,7 +223,7 @@
dev->set_multicast_list = set_multicast_list; /* no yet used */

err = register_netdev(dev);
- if (dev) {
+ if (err) {
kfree(dev);
return err;
}