> Did I say it was a problem? I'm actually pleased to find the only
> reason 2.3 is bigger is debugging. (Makes me want to go hack up ld
> though...)
>
> What's going on with bdflush?
Original problem: when it returned to non-lazy mode it failed to reload
cr3 and had a fair chance to kfree the PGD under itself (if it was unlucky
enough to get that timeslice from the exiting process). With obvious
results. That was fixed in -pre1. And there still *is* something odd.
BTW, for x86 gurus: what will happen if in ring 0 you change %cr3
(register, that is), leave the value in TSS unmodified and then return to
ring 3? Is there any chance that processor will reload the old value
(undefined/some processors are buggy and know to do that/it's documented
behaviour/never happens on known processors)?
-
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/