Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752371AbbEZOfG (ORCPT ); Tue, 26 May 2015 10:35:06 -0400 Received: from verein.lst.de ([213.95.11.211]:47605 "EHLO newverein.lst.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752268AbbEZOfD (ORCPT ); Tue, 26 May 2015 10:35:03 -0400 Date: Tue, 26 May 2015 16:34:59 +0200 From: Torsten Duwe To: Michael Ellerman Cc: Jiri Kosina , ppc-dev , Linux Kernel Mailing List Subject: Re: [PATCH] ppc64 ftrace: mark data_access callees "notrace" (pt.1) Message-ID: <20150526143459.GA24656@lst.de> References: <20150513161100.GA1619@lst.de> <1431653687.13498.1.camel@ellerman.id.au> <20150515084542.GA20453@suse.de> <20150516080534.GA27059@lst.de> <1432006027.8339.3.camel@ellerman.id.au> <20150520090325.GA11577@lst.de> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20150520090325.GA11577@lst.de> User-Agent: Mutt/1.5.17 (2007-11-01) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2670 Lines: 75 On Wed, May 20, 2015 at 11:03:25AM +0200, Torsten Duwe wrote: > On Tue, May 19, 2015 at 01:27:07PM +1000, Michael Ellerman wrote: > > On Mon, 2015-05-18 at 14:29 +0200, Jiri Kosina wrote: > > > > > > ftrace already handles recursion protection by itself (depending on the > > > per-ftrace-ops FTRACE_OPS_FL_RECURSION_SAFE flag). > > > > OK, so I wonder why that's not working for us? > > IIRC a data access fault happens just before that flag is looked at ;-) > > I'm now thinking about a hybrid solution: mark the most critical functions > "notrace", especially those directly involved with MMU loading, and add > a per-thread flag to catch the not-so-obvious cases. I realised the trace_recursion in the "current" task struct is not so far away, and it should not fault, right? So that part of the solution would look like this on top of my previous ftrace patch set. It would impact performance further so I'd stick with the "notrace" on certain hot functions, like hash_page. What do you think? Torsten diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c index 4717859..ae10752 100644 --- a/arch/powerpc/kernel/asm-offsets.c +++ b/arch/powerpc/kernel/asm-offsets.c @@ -72,6 +72,7 @@ int main(void) DEFINE(THREAD, offsetof(struct task_struct, thread)); DEFINE(MM, offsetof(struct task_struct, mm)); DEFINE(MMCONTEXTID, offsetof(struct mm_struct, context.id)); + DEFINE(TASK_TRACEREC, offsetof(struct task_struct, trace_recursion)); #ifdef CONFIG_PPC64 DEFINE(AUDITCONTEXT, offsetof(struct task_struct, audit_context)); DEFINE(SIGSEGV, SIGSEGV); diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S index a4132ef..7842092 100644 --- a/arch/powerpc/kernel/entry_64.S +++ b/arch/powerpc/kernel/entry_64.S @@ -1202,7 +1202,13 @@ _GLOBAL(ftrace_caller) SAVE_8GPRS(16,r1) SAVE_8GPRS(24,r1) - + ld r3, PACACURRENT(r13) + ld r4, TASK_TRACEREC(r3) + andi. r5, r4, 0x0010 // ( 1 << TRACE_FTRACE_BIT ) + ori r4, r4, 0x0010 + std r4, TASK_TRACEREC(r3) + bne 3f // ftrace in progress - avoid recursion! + LOAD_REG_IMMEDIATE(r3,function_trace_op) ld r5,0(r3) @@ -1224,9 +1230,14 @@ ftrace_call: bl ftrace_stub nop + ld r3, PACACURRENT(r13) + ld r4, TASK_TRACEREC(r3) + andi r4, r4, 0xffef // ~( 1 << TRACE_FTRACE_BIT ) + std r4, TASK_TRACEREC(r3) + ld r3, _NIP(r1) mtlr r3 - +3: REST_8GPRS(0,r1) REST_8GPRS(8,r1) REST_8GPRS(16,r1) -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/