Re: Xorg 7 on sparc64
- To: Jurij Smakov <email@example.com>
- Cc: firstname.lastname@example.org
- Subject: Re: Xorg 7 on sparc64
- From: Ludovic Courtès <email@example.com>
- Date: Wed, 21 Jun 2006 18:42:46 +0200
- Message-id: <20060621164245.GA2345@cassis>
- Mail-followup-to: Jurij Smakov <firstname.lastname@example.org>, email@example.com
- In-reply-to: <Pine.LNX.4.63.0606202237020.3911@bobcat>
- References: <20060613183847.GA2401@cassis> <f238d79b1278589987f6508479d1d6d2@localhost> <20060617135154.GA2345@cassis> <Pine.LNX.4.63.0606171335210.2932@bobcat> <20060618162346.GA2370@cassis> <20060619204949.GC2324@cassis> <Pine.LNX.4.63.0606202237020.3911@bobcat>
Today, 10 hours, 34 minutes, 9 seconds ago, Jurij Smakov wrote:
> So it looks like it just opens the module file (descriptor 8) and reads
> the first 256 bytes of it. I have no clue why read could block on a normal
> file like that.
It's not that `read ()' blocks but rather that `strace' fails to read
the memory region pointed to by the second argument to `read ()' (e.g.,
its `ptrace ()' call never returned).
> Do you have some weird filesystem?
> I think we have
> verified that the file has the correct md5sum before, so it's not
Actually, it _looks_ corrupt after X crashed. But upon reboot + fsck (+
additional reboot if `fsck' feels like it), the file is correct.
> Can you Stop-A at this point to get to the prom? If you can,
> the ctrace command might give you stack trace which can be useful in
> debugging it.
Unfortunately I can't because (i) at that point the screen is off ;-)
and (ii) I don't have any null-modem cable to get the console output
via the serial line.