Re: [PATCH 1/2] powerpc/interrupt: Refactor interrupt_exit_user_prepare() and syscall_exit_prepare()

From: Christophe Leroy
Date: Mon May 17 2021 - 09:58:35 EST




Le 17/05/2021 à 09:44, Nicholas Piggin a écrit :
Excerpts from Christophe Leroy's message of May 14, 2021 6:28 pm:
Last part of interrupt_exit_user_prepare() and syscall_exit_prepare()
are identical.

Create a __interrupt_exit_user_prepare() function that is called by
both.

Note that it replaces a local_irq_save(flags) by local_irq_disable().
This is similar because the flags are never used. On ppc 8xx it is
more efficient because it doesn't require reading MSR.

Can these cleanups go after my interrupt performance improvements?
I posted them for last series but were dropped due to crashes without
time to resubmit. I'm working on them again now.


Euh ... ok why not, but at the time being interrupt_exit_user_prepare() and syscall_exit_prepare() are very similar. Which makes sense because both of them are returning from kernel to user so they are to do the same preparation.

If you are doing the same changes to both of them, maybe it is worst including this refactor at the begining of your series. Or are you making them diverge with that series ?

Christophe