linux-mips
[Top] [All Lists]

Re: [PATCH 2/3] MIPS: Preliminary vdso.

To: Manuel Lauss <manuel.lauss@googlemail.com>
Subject: Re: [PATCH 2/3] MIPS: Preliminary vdso.
From: Ralf Baechle <ralf@linux-mips.org>
Date: Tue, 23 Feb 2010 22:35:18 +0100
Cc: David Daney <ddaney@caviumnetworks.com>, linux-mips@linux-mips.org
In-reply-to: <f861ec6f1002231240l40e1b07di6e751e40a2caa110@mail.gmail.com>
Original-recipient: rfc822;linux-mips@linux-mips.org
References: <1266538385-29088-1-git-send-email-ddaney@caviumnetworks.com> <1266538385-29088-3-git-send-email-ddaney@caviumnetworks.com> <f861ec6f1002231240l40e1b07di6e751e40a2caa110@mail.gmail.com>
Sender: linux-mips-bounce@linux-mips.org
User-agent: Mutt/1.5.20 (2009-08-17)
On Tue, Feb 23, 2010 at 09:40:38PM +0100, Manuel Lauss wrote:

> > This is a preliminary patch to add a vdso to all user processes.
> > Still missing are ELF headers and .eh_frame information.  But it is
> > enough to allow us to move signal trampolines off of the stack.  Note
> > that emulation of branch delay slots in the FPU emulator still
> > requires the stack.
> >
> > We allocate a single page (the vdso) and write all possible signal
> > trampolines into it.  The stack is moved down by one page and the vdso
> > is mapped into this space.
> 
> Is there anything special required (i.e. special glibc, ..) to make use of 
> these
> fine patches?

No - the way the signal handlers return is almost entirely hidden from
userland.  Only a few applications that have the need to unwind beyond
the signal stackframe may run into problems and have to be updated and
GDB is one of those that we identified.

  Ralf

<Prev in Thread] Current Thread [Next in Thread>