Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1760502AbYBSOrt (ORCPT ); Tue, 19 Feb 2008 09:47:49 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1757767AbYBSOp0 (ORCPT ); Tue, 19 Feb 2008 09:45:26 -0500 Received: from mtagate8.de.ibm.com ([195.212.29.157]:39967 "EHLO mtagate8.de.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754090AbYBSOpJ (ORCPT ); Tue, 19 Feb 2008 09:45:09 -0500 Message-Id: <20080219144504.392301782@de.ibm.com> References: <20080219144047.587547541@de.ibm.com> User-Agent: quilt/0.46-1 Date: Tue, 19 Feb 2008 15:40:54 +0100 From: Martin Schwidefsky To: linux-kernel@vger.kernel.org, linux-s390@vger.kernel.org Cc: Heiko Carstens , Martin Schwidefsky Subject: [patch 07/13] Initialize per cpu lowcores on cpu hotplug. Content-Disposition: inline; filename=007-lowcore.diff Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4368 Lines: 128 From: Heiko Carstens Just copy the first 512 read-only bytes of the current cpu lowcore if a new cpu gets onlined. The rest is zeroed out and must be explicitly initialized. Current code just copies the entire lowcore and initializes the needed fields. This should reveal bugs in future enhancements quite early. Also when the lowcore of the first cpu is replaced this is now done atomically (no interrupts, no machine checks). Signed-off-by: Heiko Carstens Signed-off-by: Martin Schwidefsky --- arch/s390/kernel/smp.c | 53 +++++++++++++++++++++++++++++++++++-------------- 1 file changed, 38 insertions(+), 15 deletions(-) Index: quilt-2.6/arch/s390/kernel/smp.c =================================================================== --- quilt-2.6.orig/arch/s390/kernel/smp.c +++ quilt-2.6/arch/s390/kernel/smp.c @@ -626,13 +626,17 @@ static int __cpuinit smp_alloc_lowcore(i if (!lowcore) return -ENOMEM; async_stack = __get_free_pages(GFP_KERNEL, ASYNC_ORDER); - if (!async_stack) - goto out_async_stack; panic_stack = __get_free_page(GFP_KERNEL); - if (!panic_stack) - goto out_panic_stack; - - *lowcore = S390_lowcore; + if (!panic_stack || !async_stack) + goto out; + /* + * Only need to copy the first 512 bytes from address 0. But since + * the compiler emits a warning if src == NULL for memcpy use copy_page + * instead. Copies more than needed but this code is not performance + * critical. + */ + copy_page(lowcore, &S390_lowcore); + memset((void *)lowcore + 512, 0, sizeof(*lowcore) - 512); lowcore->async_stack = async_stack + ASYNC_SIZE; lowcore->panic_stack = panic_stack + PAGE_SIZE; @@ -653,9 +657,8 @@ static int __cpuinit smp_alloc_lowcore(i out_save_area: free_page(panic_stack); #endif -out_panic_stack: +out: free_pages(async_stack, ASYNC_ORDER); -out_async_stack: free_pages((unsigned long) lowcore, lc_order); return -ENOMEM; } @@ -719,8 +722,8 @@ int __cpuinit __cpu_up(unsigned int cpu) cpu_lowcore->percpu_offset = __per_cpu_offset[cpu]; cpu_lowcore->current_task = (unsigned long) idle; cpu_lowcore->cpu_data.cpu_nr = cpu; - cpu_lowcore->softirq_pending = 0; - cpu_lowcore->ext_call_fast = 0; + cpu_lowcore->kernel_asce = S390_lowcore.kernel_asce; + cpu_lowcore->ipl_device = S390_lowcore.ipl_device; eieio(); while (signal_processor(cpu, sigp_restart) == sigp_busy) @@ -797,23 +800,43 @@ void cpu_die(void) void __init smp_prepare_cpus(unsigned int max_cpus) { +#ifndef CONFIG_64BIT + unsigned long save_area = 0; +#endif + unsigned long async_stack, panic_stack; + struct _lowcore *lowcore; unsigned int cpu; + int lc_order; smp_detect_cpus(); /* request the 0x1201 emergency signal external interrupt */ if (register_external_interrupt(0x1201, do_ext_call_interrupt) != 0) panic("Couldn't request external interrupt 0x1201"); - memset(lowcore_ptr, 0, sizeof(lowcore_ptr)); print_cpu_info(&S390_lowcore.cpu_data); - smp_alloc_lowcore(smp_processor_id()); + /* Reallocate current lowcore, but keep its contents. */ + lc_order = sizeof(long) == 8 ? 1 : 0; + lowcore = (void *) __get_free_pages(GFP_KERNEL | GFP_DMA, lc_order); + panic_stack = __get_free_page(GFP_KERNEL); + async_stack = __get_free_pages(GFP_KERNEL, ASYNC_ORDER); #ifndef CONFIG_64BIT if (MACHINE_HAS_IEEE) - ctl_set_bit(14, 29); /* enable extended save area */ + save_area = get_zeroed_page(GFP_KERNEL); #endif - set_prefix((u32)(unsigned long) lowcore_ptr[smp_processor_id()]); - + local_irq_disable(); + local_mcck_disable(); + lowcore_ptr[smp_processor_id()] = lowcore; + *lowcore = S390_lowcore; + lowcore->panic_stack = panic_stack + PAGE_SIZE; + lowcore->async_stack = async_stack + ASYNC_SIZE; +#ifndef CONFIG_64BIT + if (MACHINE_HAS_IEEE) + lowcore->extended_save_area_addr = (u32) save_area; +#endif + set_prefix((u32)(unsigned long) lowcore); + local_mcck_enable(); + local_irq_enable(); for_each_possible_cpu(cpu) if (cpu != smp_processor_id()) smp_create_idle(cpu); -- blue skies, Martin. "Reality continues to ruin my life." - Calvin. -- 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/