[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20091102160410.GF27911@redhat.com>
Date: Mon, 2 Nov 2009 18:04:10 +0200
From: Gleb Natapov <gleb@...hat.com>
To: Ingo Molnar <mingo@...e.hu>
Cc: kvm@...r.kernel.org, linux-mm@...ck.org,
linux-kernel@...r.kernel.org, "H. Peter Anvin" <hpa@...or.com>,
Thomas Gleixner <tglx@...utronix.de>
Subject: Re: [PATCH 02/11] Add "handle page fault" PV helper.
On Mon, Nov 02, 2009 at 10:22:14AM +0100, Ingo Molnar wrote:
>
> * Gleb Natapov <gleb@...hat.com> wrote:
>
> > diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
> > index f4cee90..14707dc 100644
> > --- a/arch/x86/mm/fault.c
> > +++ b/arch/x86/mm/fault.c
> > @@ -952,6 +952,9 @@ do_page_fault(struct pt_regs *regs, unsigned long error_code)
> > int write;
> > int fault;
> >
> > + if (arch_handle_page_fault(regs, error_code))
> > + return;
> > +
>
> This patch is not acceptable unless it's done cleaner. Currently we
> already have 3 callbacks in do_page_fault() (kmemcheck, mmiotrace,
> notifier), and this adds a fourth one. Please consolidate them into a
> single callback site, this is a hotpath on x86.
>
This call is patched out by paravirt patching mechanism so overhead should be
zero for non paravirt cases. What do you want to achieve by consolidate
them into single callback? I mean the code will still exist and will have to be
executed on every #PF. Is the goal to move them out of line?
> And please always Cc: the x86 maintainers to patches that touch x86
> code.
>
Will do.
--
Gleb.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists