[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CABqD9hbsnCL=vHpsE3HNmiB-Qv0DyhWpq5KCN0USWDm8m-6=RA@mail.gmail.com>
Date: Thu, 24 May 2012 10:39:38 -0500
From: Will Drewry <wad@...omium.org>
To: Russell King - ARM Linux <linux@....linux.org.uk>
Cc: wade_farnsworth@...tor.com, stevenrwalter@...il.com,
will.deacon@....com, Alexander Viro <viro@...iv.linux.org.uk>,
Olof Johansson <olof@...om.net>,
LKML <linux-kernel@...r.kernel.org>
Subject: Re: New ARM asm/syscall.h incompatible? (commit bf2c9f9866928df60157bc4f1ab39f93a32c754e)
On Wed, May 23, 2012 at 2:14 PM, Russell King - ARM Linux
<linux@....linux.org.uk> wrote:
> On Wed, May 23, 2012 at 02:04:20PM -0500, Will Drewry wrote:
>> I'm still curious if it wouldn't make more sense to handle the
>> sys_syscall special case prior to any cross-arch (slowpath) code
>> involvement rather than truncating the 7th parameter making
>> sys_syscall a second class citizen for those cross-arch paths.
>
> It would mean making sys_syscall an explicit special case in the fast
> path of syscall entry, which we really don't want to do. It _is_ a
> standard syscall, it just happens to have 7 arguments which are
> rewritten back to what the syscall actually expects.
>
> As I say, the alternative would be to explicitly test for the syscall
> number in the fast path of system call entry and branch away to deal
> with it. Adding unnecessary instructions to this fast path for such
> a special case when there's already a perfectly reasonable alternative
> solution doesn't fill me with any joy.
I'd been picturing this as being done exclusively after the slow-path
is triggered, in __sys_trace or syscall_trace(), but perhaps I'm
missing something that makes that untenable.
thanks!
will
--
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