lkml.org 
[lkml]   [2010]   [Dec]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH -v2 2/2] x86, acpi: Parse all SRAT cpu entries even have cpu num limitation
From
linus git + these two patches still fails on my test system with the
divide error. The failure dump is similar to what I reported here
http://lkml.indiana.edu/hypermail//linux/kernel/1012.1/03641.html
This patch description talk about new Intel systems. The test system I
am seeing failure here is an ancient Intel (2 socket P4 HT) system.
AFAICS, it does not even have an SRAT table (no "ACPI: SRAT" message
in dmesg).

Thanks,
Venki

On Thu, Dec 16, 2010 at 7:09 PM, Yinghai Lu <yinghai@kernel.org> wrote:
>
> Recent Intel new system have different order in MADT, aka will list all thread0
> at first, then all thread1.
> But SRAT table still old order, it will list cpus in one socket all together.
>
> If the user have compiled limited NR_CPUS or boot with nr_cpus=, could have missed
> to put some cpus apic id to node mapping into apicid_to_node[].
>
> for example for 4 sockets system with 64 cpus with nr_cpus=32 will get crash...
>
> [    9.106288] Total of 32 processors activated (136190.88 BogoMIPS).
> [    9.235021] divide error: 0000 [#1] SMP
> [    9.235315] last sysfs file:
> [    9.235481] CPU 1
> [    9.235592] Modules linked in:
> [    9.245398]
> [    9.245478] Pid: 2, comm: kthreadd Not tainted 2.6.37-rc1-tip-yh-01782-ge92ef79-dirty #274      /Sun Fire x4800
> [    9.265415] RIP: 0010:[<ffffffff81075a8f>]  [<ffffffff81075a8f>] select_task_rq_fair+0x4f0/0x623
> ...
> [    9.645938] RIP  [<ffffffff81075a8f>] select_task_rq_fair+0x4f0/0x623
> [    9.665356]  RSP <ffff88103f8d1c40>
> [    9.665568] ---[ end trace 2296156d35fdfc87 ]---
>
> So let just parse all cpu entries in SRAT.
>
> Also add apicid checking with MAX_LOCAL_APIC, in case We could out of boundaries of
> apicid_to_node[].
>
> it fixes following bug too.
> https://bugzilla.kernel.org/show_bug.cgi?id=22662
>
> -v2: expand to 32bit according to hpa
>   need to add MAX_LOCAL_APIC for 32bit
>
> Reported-and-Tested-by: Wu Fengguang <fengguang.wu@intel.com>
> Reported-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
> Tested-by: Myron Stowe <myron.stowe@hp.com>
> Signed-off-by: Yinghai Lu <yinghai@kernel.org>
>
> ---
>  arch/x86/kernel/acpi/boot.c |    5 +++++
>  arch/x86/mm/srat_32.c       |    1 +
>  arch/x86/mm/srat_64.c       |   10 ++++++++++
>  drivers/acpi/numa.c         |   14 ++++++++++++--
>  4 files changed, 28 insertions(+), 2 deletions(-)
>
> Index: linux-2.6/arch/x86/kernel/acpi/boot.c
> ===================================================================
> --- linux-2.6.orig/arch/x86/kernel/acpi/boot.c
> +++ linux-2.6/arch/x86/kernel/acpi/boot.c
> @@ -198,6 +198,11 @@ static void __cpuinit acpi_register_lapi
>  {
>        unsigned int ver = 0;
>
> +       if (id >= (MAX_LOCAL_APIC-1)) {
> +               printk(KERN_INFO PREFIX "skipped apicid that is too big\n");
> +               return;
> +       }
> +
>        if (!enabled) {
>                ++disabled_cpus;
>                return;
> Index: linux-2.6/arch/x86/mm/srat_64.c
> ===================================================================
> --- linux-2.6.orig/arch/x86/mm/srat_64.c
> +++ linux-2.6/arch/x86/mm/srat_64.c
> @@ -134,6 +134,10 @@ acpi_numa_x2apic_affinity_init(struct ac
>        }
>
>        apic_id = pa->apic_id;
> +       if (apic_id >= MAX_LOCAL_APIC) {
> +               printk(KERN_INFO "SRAT: PXM %u -> APIC 0x%04x -> Node %u skipped apicid that is too big\n", pxm, apic_id, node);
> +               return;
> +       }
>        apicid_to_node[apic_id] = node;
>        node_set(node, cpu_nodes_parsed);
>        acpi_numa = 1;
> @@ -168,6 +172,12 @@ acpi_numa_processor_affinity_init(struct
>                apic_id = (pa->apic_id << 8) | pa->local_sapic_eid;
>        else
>                apic_id = pa->apic_id;
> +
> +       if (apic_id >= MAX_LOCAL_APIC) {
> +               printk(KERN_INFO "SRAT: PXM %u -> APIC 0x%02x -> Node %u skipped apicid that is too big\n", pxm, apic_id, node);
> +               return;
> +       }
> +
>        apicid_to_node[apic_id] = node;
>        node_set(node, cpu_nodes_parsed);
>        acpi_numa = 1;
> Index: linux-2.6/drivers/acpi/numa.c
> ===================================================================
> --- linux-2.6.orig/drivers/acpi/numa.c
> +++ linux-2.6/drivers/acpi/numa.c
> @@ -275,13 +275,23 @@ acpi_table_parse_srat(enum acpi_srat_typ
>  int __init acpi_numa_init(void)
>  {
>        int ret = 0;
> +       int nr_cpu_entries = nr_cpu_ids;
> +
> +#ifdef CONFIG_X86
> +       /*
> +        * Should not limit number with cpu num that is from NR_CPUS or nr_cpus=
> +        * SRAT cpu entries could have different order with that in MADT.
> +        * So go over all cpu entries in SRAT to get apicid to node mapping.
> +        */
> +       nr_cpu_entries = MAX_LOCAL_APIC;
> +#endif
>
>        /* SRAT: Static Resource Affinity Table */
>        if (!acpi_table_parse(ACPI_SIG_SRAT, acpi_parse_srat)) {
>                acpi_table_parse_srat(ACPI_SRAT_TYPE_X2APIC_CPU_AFFINITY,
> -                                    acpi_parse_x2apic_affinity, nr_cpu_ids);
> +                                    acpi_parse_x2apic_affinity, nr_cpu_entries);
>                acpi_table_parse_srat(ACPI_SRAT_TYPE_CPU_AFFINITY,
> -                                    acpi_parse_processor_affinity, nr_cpu_ids);
> +                                    acpi_parse_processor_affinity, nr_cpu_entries);
>                ret = acpi_table_parse_srat(ACPI_SRAT_TYPE_MEMORY_AFFINITY,
>                                            acpi_parse_memory_affinity,
>                                            NR_NODE_MEMBLKS);
> Index: linux-2.6/arch/x86/mm/srat_32.c
> ===================================================================
> --- linux-2.6.orig/arch/x86/mm/srat_32.c
> +++ linux-2.6/arch/x86/mm/srat_32.c
> @@ -92,6 +92,7 @@ acpi_numa_processor_affinity_init(struct
>        /* mark this node as "seen" in node bitmap */
>        BMAP_SET(pxm_bitmap, cpu_affinity->proximity_domain_lo);
>
> +       /* don't need to check apic_id here, because it is always 8 bits */
>        apicid_to_pxm[cpu_affinity->apic_id] = cpu_affinity->proximity_domain_lo;
>
>        printk(KERN_DEBUG "CPU %02x in proximity domain %02x\n",
>
--
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/

\
 
 \ /
  Last update: 2010-12-17 20:15    [from the cache]
©2003-2011 Jasper Spaans