Hi !
while tracking my bugs down, I've noticed that the
function "show_registers" used to dump regs contents
when an Oops occurs doesn't check the EIP nor ESP
before dumping stack/code. In my case, this resulted
in a double Oops because EIP was NULL when
show_registers was entered.
Attached is a little patch which allows show_registers
to display "<INVALID EIP>" if EIP is 0 or -1, and
"<INVALID ESP>" is ESP is 0 or not 32bits-aligned.
This may avoid some cases of continuous looping of
wrong messages sometimes, IMHO. It may be enhanced
later. BTW, I wonder if it would be correct to
panic() after a crash from inside the kernel. In other
words, are there known circumstances under which such
an Oops can be recovered ? If not, I'd prefer to call
my messages dumper to get the messages on a floppy.
The patch is designed to 2.2.15pre17 but I think it
can apply to nearly all 2.2 kernels. Since it doesn't
affect the normal behaviour, I think it's safe to
include it into the final 2.2.15 release, and may be
especially usefull when the 2.2.15 comes out and lots
of people begin to test it and report bugs they would
have seen in 2.2.15pre if they had accepted to test
it.
Cheers,
Willy
PS: my boot problems posted yesterday concerning
Oopses
at partition detection are related to an
incompatibility between raid-2.2.14-B1 and one
other patch. Still investigating ...
___________________________________________________________
Do You Yahoo!?
Achetez, vendez! À votre prix! Sur http://encheres.yahoo.fr
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Apr 07 2000 - 21:00:11 EST