Re: [PATCH 1/14] move user_enable_single_step & co prototypes to linux/ptrace.h

[Christoph Hellwig <hch@lst.de>]

On Wed, Feb 03, 2010 at 03:42:05AM -0500, Mike Frysinger wrote:
> On Tue, Feb 2, 2010 at 13:57, Christoph Hellwig wrote:
> > While in theory user_enable_single_step/user_disable_single_step/
> > user_enable_blockstep could also be provided as an inline or macro there's no
> > good reason to do so, and having the prototype in one places keeps code size
> > and confusion down.
> 
> the only annoying thing here is that we currently have to enable both
> user_disable_single_step() and ptrace_disable() that do exactly the
> same thing.  i avoided this somewhat on Blackfin by cheating:
> #define user_disable_single_step(child) ptrace_disable(child)
> 
> so now there's no code bloat.  perhaps this could be moved into common
> linux/ptrace.h too ?

What is done by most architectures is ptrace_disable simply
calling user_disable_single_step.  Long-term I expect ptrace_disable to
go away entirely.  While a few architectures do more than just
user_disable_single_step in it that seems at least fishy to me, but
I'll wait with the audit until we have everyone actually using
ptrace_resume.