Re: [PATCH 00/12] add syscall tracepoints V3 - s390 arch update

From: Frederic Weisbecker <fweisbec@gmail.com>
To: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>,
	Jason Baron <jbaron@redhat.com>,
	linux-kernel@vger.kernel.org, mingo@elte.hu,
	laijs@cn.fujitsu.com, rostedt@goodmis.org, peterz@infradead.org,
	mathieu.desnoyers@polymtl.ca, jiayingz@google.com,
	mbligh@google.com, lizf@cn.fujitsu.com,
	Martin Schwidefsky <schwidefsky@de.ibm.com>
Subject: Re: [PATCH 00/12] add syscall tracepoints V3 - s390 arch update
Date: Tue, 25 Aug 2009 21:52:06 +0200	[thread overview]
Message-ID: <20090825195203.GB8215@nowhere> (raw)
In-Reply-To: <20090825143951.GA28268@osiris.boeblingen.de.ibm.com>

On Tue, Aug 25, 2009 at 04:39:51PM +0200, Heiko Carstens wrote:
> On Tue, Aug 25, 2009 at 03:52:32PM +0200, Frederic Weisbecker wrote:
> > On Tue, Aug 25, 2009 at 02:31:11PM +0200, Hendrik Brueckner wrote:
> > >  		meta = find_syscall_meta((unsigned long)sys_call_table[i]);
> > >  		syscalls_metadata[i] = meta;
> > >  	}
> > We can even probably move most of this code to the core, expect the tiny parts
> > that rely on the arch syscall table.
> > 
> > BTW, perhaps a silly question: would it be hard to have a generic syscall table
> > common to every archs?
> 
> That would cause a lot of churn. Every architecture initializes the syscall
> table (two tables if CONFIG_COMPAT is enabled) differently.
> s390 also only uses 32 bit pointers in the system call table for 64 bit
> kernels, since we know that the functions are within the first 4GB.
> I don't think its worth the effort.

Ok.
Well I remembered about some projects of a unified syscall table but may be
it has been given up for such reasons.