From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-pd0-x231.google.com (mail-pd0-x231.google.com [IPv6:2607:f8b0:400e:c02::231]) (using TLSv1 with cipher ECDHE-RSA-RC4-SHA (128/128 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 1699F2C008F for ; Wed, 15 Jan 2014 06:49:09 +1100 (EST) Received: by mail-pd0-f177.google.com with SMTP id q10so81338pdj.8 for ; Tue, 14 Jan 2014 11:49:06 -0800 (PST) Date: Tue, 14 Jan 2014 11:48:34 -0800 (PST) From: Hugh Dickins To: Mahesh J Salgaonkar Subject: Re: [PATCH v2] Move precessing of MCE queued event out from syscall exit path. In-Reply-To: <20140114101450.32385.65506.stgit@mars.in.ibm.com> Message-ID: References: <20140114101450.32385.65506.stgit@mars.in.ibm.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Cc: linuxppc-dev List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Tue, 14 Jan 2014, Mahesh J Salgaonkar wrote: > From: Mahesh Salgaonkar > > Huge Dickins reported an issue that b5ff4211a829 > "powerpc/book3s: Queue up and process delayed MCE events" breaks the > PowerMac G5 boot. This patch fixes it by moving the mce even processing > away from syscall exit, which was wrong to do that in first place, and > using irq work framework to delay processing of mce event. > > Reported-by: Hugh Dickins Signed-off-by: Mahesh Salgaonkar This version also boots and runs fine for me on the G5 (but of course, I'm probably not testing delayed MCE events at all). Hugh > --- > arch/powerpc/include/asm/mce.h | 1 - > arch/powerpc/kernel/entry_64.S | 5 ----- > arch/powerpc/kernel/mce.c | 13 ++++++++++--- > 3 files changed, 10 insertions(+), 9 deletions(-) > > diff --git a/arch/powerpc/include/asm/mce.h b/arch/powerpc/include/asm/mce.h > index 2257d1e..f97d8cb 100644 > --- a/arch/powerpc/include/asm/mce.h > +++ b/arch/powerpc/include/asm/mce.h > @@ -192,7 +192,6 @@ extern void save_mce_event(struct pt_regs *regs, long handled, > extern int get_mce_event(struct machine_check_event *mce, bool release); > extern void release_mce_event(void); > extern void machine_check_queue_event(void); > -extern void machine_check_process_queued_event(void); > extern void machine_check_print_event_info(struct machine_check_event *evt); > extern uint64_t get_mce_fault_addr(struct machine_check_event *evt); > > diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S > index 770d6d6..bbfb029 100644 > --- a/arch/powerpc/kernel/entry_64.S > +++ b/arch/powerpc/kernel/entry_64.S > @@ -184,11 +184,6 @@ syscall_exit: > bl .do_show_syscall_exit > ld r3,RESULT(r1) > #endif > -#ifdef CONFIG_PPC_BOOK3S_64 > -BEGIN_FTR_SECTION > - bl .machine_check_process_queued_event > -END_FTR_SECTION_IFSET(CPU_FTR_HVMODE) > -#endif > CURRENT_THREAD_INFO(r12, r1) > > ld r8,_MSR(r1) > diff --git a/arch/powerpc/kernel/mce.c b/arch/powerpc/kernel/mce.c > index d6edf2b..a7fd4cb 100644 > --- a/arch/powerpc/kernel/mce.c > +++ b/arch/powerpc/kernel/mce.c > @@ -26,6 +26,7 @@ > #include > #include > #include > +#include > #include > > static DEFINE_PER_CPU(int, mce_nest_count); > @@ -35,6 +36,11 @@ static DEFINE_PER_CPU(struct machine_check_event[MAX_MC_EVT], mce_event); > static DEFINE_PER_CPU(int, mce_queue_count); > static DEFINE_PER_CPU(struct machine_check_event[MAX_MC_EVT], mce_event_queue); > > +static void machine_check_process_queued_event(struct irq_work *work); > +struct irq_work mce_event_process_work = { > + .func = machine_check_process_queued_event, > +}; > + > static void mce_set_error_info(struct machine_check_event *mce, > struct mce_error_info *mce_err) > { > @@ -185,17 +191,19 @@ void machine_check_queue_event(void) > return; > } > __get_cpu_var(mce_event_queue[index]) = evt; > + > + /* Queue irq work to process this event later. */ > + irq_work_queue(&mce_event_process_work); > } > > /* > * process pending MCE event from the mce event queue. This function will be > * called during syscall exit. > */ > -void machine_check_process_queued_event(void) > +static void machine_check_process_queued_event(struct irq_work *work) > { > int index; > > - preempt_disable(); > /* > * For now just print it to console. > * TODO: log this error event to FSP or nvram. > @@ -206,7 +214,6 @@ void machine_check_process_queued_event(void) > &__get_cpu_var(mce_event_queue[index])); > __get_cpu_var(mce_queue_count)--; > } > - preempt_enable(); > } > > void machine_check_print_event_info(struct machine_check_event *evt) > >