Date: Fri, 9 Feb 2007 08:36:17 -0600 From: "Scot Hetzel" <swhetzel@gmail.com> To: "Kip Macy" <kmacy@fsmware.com> Cc: freebsd-current@freebsd.org Subject: Re: Broadcom Wireless card Message-ID: <790a9fff0702090636o289fd03bia61697ef2bb8688f@mail.gmail.com> In-Reply-To: <20070208231833.F15257@demos.bsdclusters.com> References: <E1HEnbC-0000mO-C8@hetzner.co.za> <20070207201342.GA66964@twoflower.idi.ntnu.no> <20070207142711.V6583@demos.bsdclusters.com> <790a9fff0702082129o75e51b0fo8f2ace4cb623a905@mail.gmail.com> <20070208231833.F15257@demos.bsdclusters.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2/9/07, Kip Macy <kmacy@fsmware.com> wrote: > On Thu, 8 Feb 2007, Scot Hetzel wrote: > > On 2/7/07, Kip Macy <kmacy@fsmware.com> wrote: > >> > >> I can't comment on the mini-pci version, but NDIS works just fine with > >> the cardbus version. It would be interesting to figure out where ndis is > >> falling short. > >> > >> -Kip > > > > I'm also seeing a kernel panic when trying to load a newer version of > > the Broadcom driver on my HP dv8135nr system. > > > > This driver (for the HP dv8135nr) works: > > > > Broadcom BCM43xx 802.11 Network Adapter Driver (4.40.19.0) > > ftp://ftp.hp.com/pub/softpaq/sp33001-33500/sp33008.exe > > > > This driver (for the HP dv6110us) causes a kernel page fault with > > non-sleepable locks held: > > > > Broadcom BCM43xx 802.11 Network Adapter Driver (4.100.15.5) > > ftp://ftp.hp.com/pub/softpaq/sp34001-34500/sp34152.exe > > Thanks for the pointer, the "fpudna in kernel mode" may be the real > issue - it means that floating point is being used which is very > strange. If I can't reproduce, it would be helpful if you would apply > the kgdb patch I'm about to send along and give me an updated > backtrace. > I get this "fpudna in kernel mode" with the older driver, and it never caused the kernel to crash. The patch doesn't apply for my system since I'm using FreeBSD/amd64. But I did use the patch as a guide to modify trgt_amd64.c. I just had to do the following: - add intrframe to the kgdb_frame_cache structure - copy code which sets intrframe in kgdb_trgt_frame_cache() from trgt_i386.c to trgt_amd64.c - change "*addrp = cache->sp + ofs;" to "*addrp = cache->sp + ofs + (cache->intrframe ? 8 : 4);" in kgdb_trgt_trapframe_prev_register The only thing this change did on my system was for it to print frames 1 -10 (frames 11 - 22 were missing) as shown in the previous backtrace. It also displayed "Previous frame inner to this frame (corrupt stack?)" as the last line before the "(kgdb)" prompt. > > > > Note: This driver has my adapter listed in the bcmwl5.inf file. > > > > FreeBSD hp010 7.0-CURRENT FreeBSD 7.0-CURRENT #5: Thu Feb 8 17:14:40 CST > > 2007 > > swhetzel@hp010:/usr/src/7x/sys-p4/amd64/compile/GENERIC.debug amd64 > > Scot -- DISCLAIMER: No electrons were mamed while sending this message. Only slightly bruised.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?790a9fff0702090636o289fd03bia61697ef2bb8688f>