Rusty Russell wrote:
> In message <3991D10F.5E8B4F8C@baldauf.org> you write:
> > while true; do insmod ip_tables; insmod ip_conntrack; insmod iptable_nat;
> > insmod ipt_MASQUERADE; rmmod ipt_MASQUERADE; rmmod iptable_nat; rmmod
> > ip_conntrack; rmmod ip_tables; done
>
> [snip]
>
> > Using /lib/modules/2.4.0-test5/ipv4/ipt_MASQUERADE.o
> > Using /lib/modules/2.4.0-test5/ipv4/ip_tables.o
> > insmod: module : QM_INFO: Invalid argument
> > Using /lib/modules/2.4.0-test5/ipv4/ip_conntrack.o
> > insmod: module : QM_INFO: Invalid argument
> > Using /lib/modules/2.4.0-test5/ipv4/iptable_nat.o
> > insmod: module : QM_INFO: Invalid argument
> > Using /lib/modules/2.4.0-test5/ipv4/ipt_MASQUERADE.o
> > insmod: module : QM_INFO: Invalid argument
> > rmmod: module : QM_INFO: Invalid argument
> > rmmod: module ipt_MASQUERADE is not loaded
> > rmmod: module : QM_INFO: Invalid argument
> > rmmod: module : QM_INFO: Invalid argument
> > rmmod: module iptable_nat is not loaded
> >
> > ... the complete module subsystem seems to be destroyed ...
>
> Surprise surprise, there's a race in the module subsystem. Given how
> disgusting the code is[*], that's not hugely surprising.
>
> Philipp Rumpf wrote a patch for the module subsystem a month or so
> back; it included the `freeze all CPUs to unload' patch though, which
> Alan disliked. But it's attached below, for testing. If this works,
> I'll lobby for its inclusion (well, a variant with a timeout and
> explicit bind-to-processor).
>
> Rusty.
> [*] Partially, that's because the whole `module' idea is pretty
> disgusting...
>
> diff -urN linux/arch/i386/mm/extable.c linux-prumpf/arch/i386/mm/extable.c
> --- linux/arch/i386/mm/extable.c Wed Nov 10 08:31:37 1999
> +++ linux-prumpf/arch/i386/mm/extable.c Wed Jun 28 13:43:10 2000
> @@ -30,26 +30,30 @@
> return 0;
> }
>
... for which kernel is this patch? It does not seem to work for linux-2.4.0-test5:
router|13:40:44|/usr/src> patch <module.race.patch -p0
patching file linux/arch/i386/mm/extable.c
Hunk #1 FAILED at 30.
1 out of 1 hunk FAILED -- saving rejects to file linux/arch/i386/mm/extable.c.rej
patching file linux/include/linux/smp.h
Hunk #1 succeeded at 70 with fuzz 1 (offset -6 lines).
Hunk #2 FAILED at 88.
1 out of 2 hunks FAILED -- saving rejects to file linux/include/linux/smp.h.rej
patching file linux/kernel/module.c
Hunk #1 succeeded at 46 with fuzz 1.
Hunk #2 FAILED at 119.
Hunk #3 FAILED at 127.
Hunk #4 FAILED at 180.
Hunk #5 FAILED at 359.
Hunk #6 FAILED at 367.
Hunk #7 FAILED at 384.
Hunk #8 FAILED at 409.
Hunk #9 FAILED at 440.
Hunk #10 FAILED at 456.
Hunk #11 FAILED at 679.
Hunk #12 FAILED at 725.
Hunk #13 FAILED at 746.
Hunk #14 FAILED at 789.
Hunk #15 FAILED at 870.
Hunk #16 FAILED at 938.
Hunk #17 FAILED at 957.
Hunk #18 FAILED at 992.
Hunk #19 FAILED at 1037.
Hunk #20 FAILED at 1048.
19 out of 20 hunks FAILED -- saving rejects to file linux/kernel/module.c.rej
patching file linux/lib/Makefile
Hunk #1 succeeded at 14 with fuzz 2 (offset 4 lines).
patching file linux/lib/freeze.c
patch unexpectedly ends in middle of line
patch: **** malformed patch at line 367:
router|13:40:46|/usr/src>
Xuân.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Aug 15 2000 - 21:00:23 EST