Re: [BUG] cpu hotplug vs scheduler

[Avi Kivity <avi@qumranet.com>]

Heiko Carstens wrote:
>> Now it died when allocating the pda:
>>
>> [4301986.526627] Booting processor 1/6 ip 6000
>> [4301986.537195] swapper: page allocation failure. order:2, mode:0x20
>> [4301986.537195] Pid: 0, comm: swapper Not tainted 2.6.26-rc2 #728
>> [4301986.537195]
>> [4301986.537195] Call Trace:
>> [4301986.537195]  [<ffffffff802388ae>] ? __printk_ratelimit+0x9/0xb
>> [4301986.537195]  [<ffffffff80272e07>] __alloc_pages_internal+0x387/0x39f
>> [4301986.537195]  [<ffffffff80272e35>] __alloc_pages+0xb/0xd
>> [4301986.537195]  [<ffffffff80272e4c>] __get_free_pages+0x15/0x46
>> [4301986.537195]  [<ffffffff80210cdf>] pda_init+0x87/0xac
>> [4301986.537195]  [<ffffffff8042fed1>] cpu_init+0x5e/0x34a
>> [4301986.537195]  [<ffffffff8026b437>] ? rcu_needs_cpu+0x36/0x47
>> [4301986.537195]  [<ffffffff802311a7>] ? idle_task_exit+0x71/0xad
>> [4301986.537195]  [<ffffffff804331aa>] start_secondary+0xc/0x178
>> [4301986.537195]
>> [4301986.537195] Mem-info:
>> [4301986.537195] DMA per-cpu:
>> [4301986.537195] CPU    0: hi:    0, btch:   1 usd:   0
>> [4301986.537195] CPU    2: hi:    0, btch:   1 usd:   0
>> [4301986.537195] DMA32 per-cpu:
>> [4301986.537195] CPU    0: hi:  186, btch:  31 usd: 153
>> [4301986.537195] CPU    2: hi:  186, btch:  31 usd: 135
>> [4301986.537195] Active:347879 inactive:109278 dirty:107 writeback:0  
>> unstable:0
>> [4301986.537195]  free:5311 slab:10186 mapped:30496 pagetables:1587 bounce:0
>> [4301986.537195] DMA free:8012kB min:24kB low:28kB high:36kB active:60kB  
>> inactive:0kB present:8644kB pages_scanned:0 all_unreclaimable? no
>> [4301986.537195] lowmem_reserve[]: 0 1999 1999 1999
>> [4301986.537195] DMA32 free:13232kB min:5704kB low:7128kB high:8556kB  
>> active:1391456kB inactive:437112kB present:2047248kB pages_scanned:0  
>> all_unreclaimable? no
>> [4301986.537195] lowmem_reserve[]: 0 0 0 0
>> [4301986.537195] DMA: 19*4kB 16*8kB 10*16kB 11*32kB 6*64kB 4*128kB  
>> 1*256kB 2*512kB 3*1024kB 1*2048kB 0*4096kB = 8012kB
>> [4301986.537195] DMA32: 2034*4kB 570*8kB 1*16kB 0*32kB 0*64kB 0*128kB  
>> 0*256kB 1*512kB 0*1024kB 0*2048kB 0*4096kB = 13224kB
>> [4301986.537195] 212750 total pagecache pages
>> [4301986.537195] Swap cache: add 73, delete 18, find 0/0
>> [4301986.537195] Free swap  = 3210964kB
>> [4301986.537195] Total swap = 3211256kB
>>
>> There seem to be some order-2 allocations available, though.  And in any  
>> case, cpu hotplug should't crash on such failures, only fail.
>>     
>
> Hi Avi,
>
> I think you fixed this specific bug. Does cpu hotplug stresstest now work
> for you?
>   

Which bug, the order-2 allocs?  I certainly didn't fix it.

Last I tried, cpu hotplug still failed, but it was some time ago as 
patches flow.

-- 
error compiling committee.c: too many arguments to function