Date: Sat, 13 Oct 2007 13:00:04 GMT From: "Adrian Punga" <adrian.punga@gmail.com> To: freebsd-net@FreeBSD.org Subject: Re: kern/109406: [ndis] Broadcom WLAN driver 4.100.15.5 doesn't work with Ndisgen Message-ID: <200710131300.l9DD04wu084074@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR kern/109406; it has been noted by GNATS.
From: "Adrian Punga" <adrian.punga@gmail.com>
To: bug-followup@freebsd.org, darkvincentdude@yahoo.com
Cc:
Subject: Re: kern/109406: [ndis] Broadcom WLAN driver 4.100.15.5 doesn't work with Ndisgen
Date: Sat, 13 Oct 2007 15:58:27 +0300
Without anymore nice ideas I got brutal and identified the call to
curthread by classic print-debug and it seems a piece of "goofy" code
in kern_ndis.c was calling it:
/*
* This may look really goofy, but apparently it is possible
* to halt a miniport too soon after it's been initialized.
* After MiniportInitialize() finishes, pause for 1 second
* to give the chip a chance to handle any short-lived timers
* that were set in motion. If we call MiniportHalt() too soon,
* some of the timers may not be cancelled, because the driver
* expects them to fire before the halt is called.
*/
tsleep(curthread->td_proc, PWAIT, "ndwait", hz);
After commenting the tsleep call which seems to affect only some cards
I got a little bit further:
ndis0: <Broadcom 802.11b/g WLAN> mem 0xf4000000-0xf4003fff irq 17 at
device 0.0 on pci16
ndis0: NDIS API version: 5.1
ndis0: Ethernet address: 00:14:a5:f0:fa:c1
Fatal trap 12: page fault while in kernel mode
fault virtual address = 0x8
fault code = supervisor read, page not present
instruction pointer = 0x20:0xc4362da5
stack pointer = 0x28:0xd53cdaf0
frame pointer = 0x28:0xd53cdaf8
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 = 2283 (kldload)
trap number = 12
panic: page fault
I'm suspecting an uninitialized pointer returned as the current thread
in some conditions.
Continuing with print-debug ;) as it seems I'm not far from a solution.
Adrian Punga
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200710131300.l9DD04wu084074>