Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933076Ab3CEDja (ORCPT ); Mon, 4 Mar 2013 22:39:30 -0500 Received: from mail-wg0-f45.google.com ([74.125.82.45]:42011 "EHLO mail-wg0-f45.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932472Ab3CEDj3 (ORCPT ); Mon, 4 Mar 2013 22:39:29 -0500 MIME-Version: 1.0 In-Reply-To: <5134EA49.6020807@gmail.com> References: <1362401764-31494-1-git-send-email-chenhc@lemote.com> <5134EA49.6020807@gmail.com> Date: Tue, 5 Mar 2013 11:39:27 +0800 X-Google-Sender-Auth: QDcKc1o-cpYPGMBiIiQRZzKl3qc Message-ID: Subject: Re: [PATCH RFC] MIPS: Build uasm-generated code only once to avoid CPU Hotplug problem From: Huacai Chen To: David Daney Cc: Ralf Baechle , linux-mips@linux-mips.org, linux-kernel@vger.kernel.org, Fuxin Zhang , Zhangjin Wu , Hongbing Hu Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 6378 Lines: 182 OK, I'm reworking... On Tue, Mar 5, 2013 at 2:39 AM, David Daney wrote: > On 03/04/2013 04:56 AM, Huacai Chen wrote: >> >> Currently, clear_page()/copy_page() are generated by Micro-assembler >> dynamically. But they are unavailable until uasm_resolve_relocs() has >> finished because jump labels are illegal before that. Since these >> functions are shared by every CPU, we only call build_clear_page()/ >> build_copy_page() on CPU#0 at boot time. Without this patch, programs >> will get random memory corruption (segmentation fault, bus error, etc.) >> while CPU Hotplug (e.g. one CPU is using clear_page() while another is >> generating it in cpu_cache_init()). >> >> For similar reasons we modify build_tlb_refill_handler()'s invocation. >> > > I like the general idea behind the patch. But would prefer to move the test > to the functions that are generating the code. That way if more call sites > are added, we don't have to copy the boiler-plate testing code. > > Also can we change the test to something like: > > static atomic_t execute_once; > . > . > . > if (!atomic_xchg(&execute_once, 1)) { > /* Generate the code */ > } > . > . > . > > That way if we take CPU-0 on and off line, the code will not be regenerated. > > Thanks, > David Daney > > > >> Signed-off-by: Huacai Chen >> Signed-off-by: Hongbing Hu >> --- >> arch/mips/mm/c-octeon.c | 6 ++++-- >> arch/mips/mm/c-r3k.c | 6 ++++-- >> arch/mips/mm/c-r4k.c | 6 ++++-- >> arch/mips/mm/c-tx39.c | 6 ++++-- >> arch/mips/mm/tlb-r3k.c | 3 ++- >> arch/mips/mm/tlb-r4k.c | 3 ++- >> arch/mips/mm/tlb-r8k.c | 3 ++- >> arch/mips/sgi-ip27/ip27-init.c | 3 ++- >> 8 files changed, 24 insertions(+), 12 deletions(-) >> >> diff --git a/arch/mips/mm/c-octeon.c b/arch/mips/mm/c-octeon.c >> index 6ec04da..181a1bc 100644 >> --- a/arch/mips/mm/c-octeon.c >> +++ b/arch/mips/mm/c-octeon.c >> @@ -280,8 +280,10 @@ void __cpuinit octeon_cache_init(void) >> >> __flush_kernel_vmap_range = octeon_flush_kernel_vmap_range; >> >> - build_clear_page(); >> - build_copy_page(); >> + if (smp_processor_id() == 0) { >> + build_clear_page(); >> + build_copy_page(); >> + } >> >> board_cache_error_setup = octeon_cache_error_setup; >> } >> diff --git a/arch/mips/mm/c-r3k.c b/arch/mips/mm/c-r3k.c >> index 031c4c2..b7b0cfd 100644 >> --- a/arch/mips/mm/c-r3k.c >> +++ b/arch/mips/mm/c-r3k.c >> @@ -342,6 +342,8 @@ void __cpuinit r3k_cache_init(void) >> printk("Primary data cache %ldkB, linesize %ld bytes.\n", >> dcache_size >> 10, dcache_lsize); >> >> - build_clear_page(); >> - build_copy_page(); >> + if (smp_processor_id() == 0) { >> + build_clear_page(); >> + build_copy_page(); >> + } >> } >> diff --git a/arch/mips/mm/c-r4k.c b/arch/mips/mm/c-r4k.c >> index 5f9171d..3f671d6 100644 >> --- a/arch/mips/mm/c-r4k.c >> +++ b/arch/mips/mm/c-r4k.c >> @@ -1548,8 +1548,10 @@ void __cpuinit r4k_cache_init(void) >> } >> #endif >> >> - build_clear_page(); >> - build_copy_page(); >> + if (smp_processor_id() == 0) { >> + build_clear_page(); >> + build_copy_page(); >> + } >> #if !defined(CONFIG_MIPS_CMP) >> local_r4k___flush_cache_all(NULL); >> #endif >> diff --git a/arch/mips/mm/c-tx39.c b/arch/mips/mm/c-tx39.c >> index 87d23ca..1e42e12 100644 >> --- a/arch/mips/mm/c-tx39.c >> +++ b/arch/mips/mm/c-tx39.c >> @@ -434,7 +434,9 @@ void __cpuinit tx39_cache_init(void) >> printk("Primary data cache %ldkB, linesize %d bytes\n", >> dcache_size >> 10, current_cpu_data.dcache.linesz); >> >> - build_clear_page(); >> - build_copy_page(); >> + if (smp_processor_id() == 0) { >> + build_clear_page(); >> + build_copy_page(); >> + } >> tx39h_flush_icache_all(); >> } >> diff --git a/arch/mips/mm/tlb-r3k.c b/arch/mips/mm/tlb-r3k.c >> index a63d1ed..86b4a79 100644 >> --- a/arch/mips/mm/tlb-r3k.c >> +++ b/arch/mips/mm/tlb-r3k.c >> @@ -280,5 +280,6 @@ void __cpuinit tlb_init(void) >> { >> local_flush_tlb_all(); >> >> - build_tlb_refill_handler(); >> + if (smp_processor_id() == 0) >> + build_tlb_refill_handler(); >> } >> diff --git a/arch/mips/mm/tlb-r4k.c b/arch/mips/mm/tlb-r4k.c >> index 0113330..db19624 100644 >> --- a/arch/mips/mm/tlb-r4k.c >> +++ b/arch/mips/mm/tlb-r4k.c >> @@ -439,5 +439,6 @@ void __cpuinit tlb_init(void) >> printk("Ignoring invalid argument ntlb=%d\n", >> ntlb); >> } >> >> - build_tlb_refill_handler(); >> + if (smp_processor_id() == 0) >> + build_tlb_refill_handler(); >> } >> diff --git a/arch/mips/mm/tlb-r8k.c b/arch/mips/mm/tlb-r8k.c >> index 91c2499..43fe634 100644 >> --- a/arch/mips/mm/tlb-r8k.c >> +++ b/arch/mips/mm/tlb-r8k.c >> @@ -244,5 +244,6 @@ void __cpuinit tlb_init(void) >> >> local_flush_tlb_all(); >> >> - build_tlb_refill_handler(); >> + if (smp_processor_id() == 0) >> + build_tlb_refill_handler(); >> } >> diff --git a/arch/mips/sgi-ip27/ip27-init.c >> b/arch/mips/sgi-ip27/ip27-init.c >> index 923c080..62c41ab 100644 >> --- a/arch/mips/sgi-ip27/ip27-init.c >> +++ b/arch/mips/sgi-ip27/ip27-init.c >> @@ -84,7 +84,8 @@ static void __cpuinit per_hub_init(cnodeid_t cnode) >> >> memcpy((void *)(CKSEG0 + 0x100), &except_vec2_generic, >> 0x80); >> memcpy((void *)(CKSEG0 + 0x180), &except_vec3_generic, >> 0x80); >> - build_tlb_refill_handler(); >> + if (smp_processor_id() == 0) >> + build_tlb_refill_handler(); >> memcpy((void *)(CKSEG0 + 0x100), (void *) CKSEG0, 0x80); >> memcpy((void *)(CKSEG0 + 0x180), &except_vec3_generic, >> 0x100); >> __flush_cache_all(); >> > > -- 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/