WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-devel

[Xen-devel] RE: [PATCH] Fix cpu online/offline bug: mce memory leak.

To: Keir Fraser <keir.xen@xxxxxxxxx>, "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>
Subject: [Xen-devel] RE: [PATCH] Fix cpu online/offline bug: mce memory leak.
From: "Liu, Jinsong" <jinsong.liu@xxxxxxxxx>
Date: Tue, 1 Mar 2011 17:27:06 +0800
Accept-language: en-US
Acceptlanguage: en-US
Cc: Haitao Shan <maillists.shan@xxxxxxxxx>
Delivery-date: Tue, 01 Mar 2011 01:27:50 -0800
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <C9926B43.13F51%keir.xen@xxxxxxxxx>
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
References: <BC00F5384FCFC9499AF06F92E8B78A9E1F9FDC4BC8@xxxxxxxxxxxxxxxxxxxxxxxxxxxx> <C9926B43.13F51%keir.xen@xxxxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
Thread-index: AcvX8FtN6tarrAIBRGq0NWDWPv4e2AAAcmG5AAAa05A=
Thread-topic: [PATCH] Fix cpu online/offline bug: mce memory leak.
Yes :)

Keir Fraser wrote:
> Brilliant! This appears to move allocations of the CPU bringup path
> and into CPU_UP_PREPARE? If so, this should not only fix a memory
> leak but also Haitao Shan's crash during CPU bringup. Cc'ing Haitao.
> 
>  -- Keir
> 
> On 01/03/2011 09:09, "Liu, Jinsong" <jinsong.liu@xxxxxxxxx> wrote:
> 
>> Fix cpu online/offline bug: mce memory leak.
>> 
>> Current Xen mce logic didn't free mcabanks. This would be a memory
>> leak when cpu offline. When repeatly do cpu online/offline, this
>> memory leak would make xenpool shrink, and at a time point, will
>> call alloc_heap_pages --> flush_area_mask, which
>> ASSERT(local_irq_is_enabled()). 
>> However, cpu online is irq disable, so it finally result in Xen
>> crash. 
>> 
>> This patch fix the memory leak bug, and tested OK over 110,000 round
>> cpu online/offline. 
>> 
>> Signed-off-by: Liu, Jinsong <jinsong.liu@xxxxxxxxx>
>> 
>> diff -r 1a364b17d66a xen/arch/x86/cpu/mcheck/mce_intel.c
>> --- a/xen/arch/x86/cpu/mcheck/mce_intel.c Fri Feb 25 01:26:01 2011
>> +0800 +++ b/xen/arch/x86/cpu/mcheck/mce_intel.c Mon Feb 28 19:19:20
>> 2011 +0800 @@ -1227,9 +1227,24 @@ static void intel_init_mce(void)
>>      mce_uhandler_num = sizeof(intel_mce_uhandlers)/sizeof(struct
>>  mca_error_handler); }
>> 
>> -static int intel_init_mca_banks(void)
>> +static void cpu_mcabank_free(unsigned int cpu)
>>  {
>> -    struct mca_banks *mb1, *mb2, * mb3;
>> +    struct mca_banks *mb1, *mb2, *mb3, *mb4;
>> +
>> +    mb1 = per_cpu(mce_clear_banks, cpu);
>> +    mb2 = per_cpu(no_cmci_banks, cpu);
>> +    mb3 = per_cpu(mce_banks_owned, cpu);
>> +    mb4 = per_cpu(poll_bankmask, cpu);
>> +
>> +    mcabanks_free(mb1);
>> +    mcabanks_free(mb2);
>> +    mcabanks_free(mb3);
>> +    mcabanks_free(mb4);
>> +}
>> +
>> +static void cpu_mcabank_alloc(unsigned int cpu)
>> +{
>> +    struct mca_banks *mb1, *mb2, *mb3;
>> 
>>      mb1 = mcabanks_alloc();
>>      mb2 = mcabanks_alloc();
>> @@ -1237,22 +1252,23 @@ static int intel_init_mca_banks(void)
>>      if (!mb1 || !mb2 || !mb3)
>>          goto out;
>> 
>> -    __get_cpu_var(mce_clear_banks) = mb1;
>> -    __get_cpu_var(no_cmci_banks) = mb2;
>> -    __get_cpu_var(mce_banks_owned) = mb3;
>> +    per_cpu(mce_clear_banks, cpu) = mb1;
>> +    per_cpu(no_cmci_banks, cpu) = mb2;
>> +    per_cpu(mce_banks_owned, cpu) = mb3;
>> +    return;
>> 
>> -    return 0;
>>  out:
>>      mcabanks_free(mb1);
>>      mcabanks_free(mb2);
>>      mcabanks_free(mb3);
>> -    return -ENOMEM;
>>  }
>> 
>>  /* p4/p6 family have similar MCA initialization process */
>>  enum mcheck_type intel_mcheck_init(struct cpuinfo_x86 *c)  {
>> -    if (intel_init_mca_banks())
>> +    if ( !this_cpu(mce_clear_banks) ||
>> +         !this_cpu(no_cmci_banks)   ||
>> +         !this_cpu(mce_banks_owned) )
>>           return mcheck_none;
>> 
>>      intel_init_mca(c);
>> @@ -1301,13 +1317,19 @@ static int cpu_callback(
>>  static int cpu_callback(
>>      struct notifier_block *nfb, unsigned long action, void *hcpu)  {
>> +    unsigned int cpu = (unsigned long)hcpu;
>> +
>>      switch ( action )
>>      {
>> +    case CPU_UP_PREPARE:
>> +        cpu_mcabank_alloc(cpu);
>> +        break;
>>      case CPU_DYING:
>>          cpu_mcheck_disable();
>>          break;
>>      case CPU_DEAD:
>>          cpu_mcheck_distribute_cmci();
>> +        cpu_mcabank_free(cpu);
>>          break;
>>      default:
>>          break;
>> @@ -1322,6 +1344,8 @@ static struct notifier_block cpu_nfb = {
>> 
>>  static int __init intel_mce_initcall(void)
>>  {
>> +    void *hcpu = (void *)(long)smp_processor_id();
>> +    cpu_callback(&cpu_nfb, CPU_UP_PREPARE, hcpu);
>>      register_cpu_notifier(&cpu_nfb);
>>      return 0;
>>  }
>> diff -r 1a364b17d66a xen/arch/x86/setup.c
>> --- a/xen/arch/x86/setup.c Fri Feb 25 01:26:01 2011 +0800
>> +++ b/xen/arch/x86/setup.c Mon Feb 28 19:19:20 2011 +0800
>> @@ -1203,6 +1203,8 @@ void __init __start_xen(unsigned long mb
>> 
>>      arch_init_memory();
>> 
>> +    do_presmp_initcalls();
>> +
>>      identify_cpu(&boot_cpu_data);
>>      if ( cpu_has_fxsr )
>>          set_in_cr4(X86_CR4_OSFXSR);
>> @@ -1235,8 +1237,6 @@ void __init __start_xen(unsigned long mb     
>> initialize_keytable(); 
>> 
>>      console_init_postirq();
>> -
>> -    do_presmp_initcalls();
>> 
>>      for_each_present_cpu ( i )
>>      {


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel