Re: HSF modem driver corrupts system

--- On Tue, 3/10/09, Linuxant support (Jonathan) <[email protected]> wrote:

> thank you for this information.
> "CONFIG_PREEMPT=y" is well supported but it is
> possible that the lack of
> "CONFIG_PREEMPT_VOLUNTARY=y" could be fatal to the
> driver. If this something that could be improved in the
> future.

I spoke too soon. It didn't crash for three or four times
after the disconnect, but then just a few hours ago:

Mar 10 21:36:50 irascible pppd[15913]: Connection terminated.
Mar 10 21:36:50 system ------------[ cut here ]------------
Mar 10 21:36:50 system kernel BUG at drivers/char/tty_ldisc.c:199!
Mar 10 21:36:50 system invalid opcode: 0000 [#1]
Mar 10 21:36:50 system Modules linked in: hsfusbcd2 hsfmc97sis hsfmc97ati hsfmc97ali hsfmc97via hsfmc97ich hsfpcibasic3 hs
fpcibasic2 hsfserial hsfengine(P) hsfosspec hsfsoar [last unloaded: hsfosspec]
Mar 10 21:36:50 system
Mar 10 21:36:50 system Pid: 15913, comm: pppd Tainted: P          (2.6.27-gentoo-r8 #1)
Mar 10 21:36:50 system EIP: 0060:[<c01e23db>] EFLAGS: 00010046 CPU: 0
Mar 10 21:36:50 system EIP is at tty_ldisc_put+0x1e/0x32
Mar 10 21:36:50 system EAX: 00000000 EBX: 00000283 ECX: f8858840 EDX: c046ea80
Mar 10 21:36:50 system ESI: ee920000 EDI: 00000000 EBP: 00000000 ESP: ee957ec4
Mar 10 21:36:50 system DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068
Mar 10 21:36:50 system Process pppd (pid: 15913, ti=ee956000 task=ee9a46c0 task.ti=ee956000)
Mar 10 21:36:50 system Stack: ee957f14 c01e28c7 00000000 00000000 c018ec0b 00000000 c024d20c 00000001
Mar 10 21:36:50 system ee9200d0 00000246 ee920000 00000002 00000000 c0113ee7 00000000 00000246
Mar 10 21:36:50 system ee920010 c01dd581 00000000 7fffffff 00000000 ee9a46c0 c01232e0 ee957f20
Mar 10 21:36:50 system Call Trace:
Mar 10 21:36:50 system [<c01e28c7>] tty_set_ldisc+0x2d5/0x326
Mar 10 21:36:50 system [<c018ec0b>] journal_dirty_metadata+0x7c/0x82
Mar 10 21:36:50 system [<c024d20c>] ppp_async_ioctl+0x8e/0x164
Mar 10 21:36:50 system [<c0113ee7>] __wake_up+0xf/0x15
Mar 10 21:36:50 system [<c01dd581>] tty_wakeup+0x40/0x44
Mar 10 21:36:50 system [<c01232e0>] autoremove_wake_function+0x0/0x2d
Mar 10 21:36:50 system [<c01df874>] tty_ioctl+0x45b/0x5ed
Mar 10 21:36:50 system [<c01df419>] tty_ioctl+0x0/0x5ed
Mar 10 21:36:50 system [<c015850e>] vfs_ioctl+0x16/0x48
Mar 10 21:36:50 system [<c015871e>] do_vfs_ioctl+0x1de/0x1f1
Mar 10 21:36:50 system [<c010eade>] do_page_fault+0x228/0x536
Mar 10 21:36:50 system [<c015875d>] sys_ioctl+0x2c/0x43
Mar 10 21:36:50 system [<c01029d9>] sysenter_do_call+0x12/0x25
Mar 10 21:36:50 system =======================
Mar 10 21:36:50 system Code: 58 5a 89 f2 89 d8 5b 5e eb 91 5b 5e c3 53 8b 40 08 83 f8 11 76 04 0f 0b eb fe 9c 5b fa 8b 14
85 c0 0e 4f c0 8b 42 48 85 c0 75 04 <0f> 0b eb fe 48 89 42 48 8b 42 44 e8 46 ae f4 ff 53 9d 5b c3 56
Mar 10 21:36:50 system EIP: [<c01e23db>] tty_ldisc_put+0x1e/0x32 SS:ESP 0068:ee957ec4
Mar 10 21:36:50 system ---[ end trace 785bf77333246b46 ]---

Crashing twenty percent of the time is a lot better than
crashing a hundred percent of the time (especially since
I don't use the modem unless, I can't avoid it), but it
is still a pain to reboot the system, and get everything
restored.

My offer to supply my config file or other additional
information still holds. :)

Tom