Date: Fri, 23 May 2003 17:44:33 -0700 From: Joshua Oreman <oremanj@webserver.get-linux.org> To: Mikhail Kruk <meshko@cs.brandeis.edu> Cc: hackers@freebsd.org Subject: Re: current crashes Message-ID: <20030524004433.GA46783@webserver.get-linux.org> In-Reply-To: <Pine.LNX.4.33.0305231845200.7675-100000@iole.cs.brandeis.edu> References: <20030523223751.GA41427@webserver.get-linux.org> <Pine.LNX.4.33.0305231845200.7675-100000@iole.cs.brandeis.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, May 23, 2003 at 06:47:24PM -0400 or thereabouts, Mikhail Kruk seemed to write: > > > Fatal trap 12: page fault while in kernel mode > > > fault virtual address = 0x80790ab0 > > > fault code = supervisor read, page not present > > > instruction pointer = 0x8:0xc06ea4d0 > > ^^^^^^^^^^ > > This value is important, but meaningless in its current form. > > Please see chapter 18 of the FAQ. Read it, do what it says, > > give us the symbol(s). > > The address I have there ^^^ is beyond addresses that nm -s finds in this > kernel. This is either because I screwed up the address when I copied it > from display (very likely) or because the problem happened in acpi, which > is a module, and therefore is not in /boot/kernel/kernel. Any way to > recalculate the offset into the module? Oh yeah... forgot about that :-) This is a wild guess, but... When the kernel loads, at the beginning of dmesg, it says something like: Preloaded elf module acpi.ko at 0x12345678 [if run from loader] If it's a cmdline-loaded kld, do `kldstat' and it'll tell you the load address. Now, if my intuition is correct, you should be able to subtract the value above from the load address. Do `nm -s /boot/kernel/acpi.ko' and search for this new value. But all of this is moot if you can get a crash dump :-) -- Josh
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030524004433.GA46783>