Hi,

Petr Vandrovec wrote:
>> The reason is that the previous part of the macro
>> can jump to that part. So how can I divide those?
> Use real labels instead of numeric local labels.
OK, I am really sorry for asking such a crap :(

>> returning to the kernel if fails, should die()
>> anyway. Is this correct?
> If you'll first test result of lar, and then
> you'll do two independent RESTORE_REGS (second
> setup, instead of doing push + pop), you can reuse
> RESOTRE_REGS+add $4,%esp+iret return path, including
> its exception handling.
Yes, but I was wondering whether we really do
need an exception handler when returning to
kernel. There should be no fault there, and
if there is - it should be an Oops, I think.
Nevertheless, I did what you said. Perhaps
something like die_if_kernel() is somewhere
down the road anyway.

> Patch looks fine, though you could speedup it a bit as outlined above.
New patch is attached.

> Now you have to persuade others that this patch should include patch
> into the kernel.
Yes, if not for that anonymous guy, who kept posting
to me until he finally convinced me that the Ring-0
approach is not that difficult at all.
So I tried... It was much more difficult to code
up, but at the end it looks a little better
and localized to entry.S completely. OTOH it
touches the exception handlers, but not too much -
it adds only 5 insns on the fast path. And the
code is very fragile, but after I made all the
magic numbers a #define consts, it actually looks
not so bad.
I don't know which patch is really better, so
I am attaching both.
Any ideas on which one should I have to continue
working on and what improvements can be done to
either one?