Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Yuanhan Liu
On Thu, Jan 31, 2013 at 10:18:18PM +0100, Ingo Molnar wrote: > > * Yuanhan Liu wrote: > > > On Thu, Jan 31, 2013 at 02:12:28PM +0100, Ingo Molnar wrote: > > > > > > * Yuanhan Liu wrote: > > > > > > > BTW, mind to tell a nice test case for mmap_sem? > > > > > > this one was write-hitting on m

Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Ingo Molnar
* Yuanhan Liu wrote: > On Thu, Jan 31, 2013 at 02:12:28PM +0100, Ingo Molnar wrote: > > > > * Yuanhan Liu wrote: > > > > > BTW, mind to tell a nice test case for mmap_sem? > > > > this one was write-hitting on mmap_sem pretty hard, last I > > checked: > > > > http://people.redhat.com/min

Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Yuanhan Liu
On Thu, Jan 31, 2013 at 02:12:28PM +0100, Ingo Molnar wrote: > > * Yuanhan Liu wrote: > > > BTW, mind to tell a nice test case for mmap_sem? > > this one was write-hitting on mmap_sem pretty hard, last I > checked: > > http://people.redhat.com/mingo/threaded-mmap-stresstest/ Thanks! Is th

Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Ingo Molnar
* Yuanhan Liu wrote: > BTW, mind to tell a nice test case for mmap_sem? this one was write-hitting on mmap_sem pretty hard, last I checked: http://people.redhat.com/mingo/threaded-mmap-stresstest/ Thanks, Ingo -- To unsubscribe from this list: send the line "unsubscribe linux-kern

Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Ingo Molnar
* Michel Lespinasse wrote: > On Wed, Jan 30, 2013 at 1:14 AM, Yuanhan Liu > wrote: > > We(Linux Kernel Performance project) found a regression introduced by > > commit 5a50508, which just convert all mutex lock to rwsem write lock. > > The semantics is same, but the results is quite huge in som

Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Yuanhan Liu
On Thu, Jan 31, 2013 at 03:57:51AM -0800, Michel Lespinasse wrote: > On Wed, Jan 30, 2013 at 1:14 AM, Yuanhan Liu > wrote: > > We(Linux Kernel Performance project) found a regression introduced by > > commit 5a50508, which just convert all mutex lock to rwsem write lock. > > The semantics is same,

Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Yuanhan Liu
On Thu, Jan 31, 2013 at 11:45:41AM +0100, Ingo Molnar wrote: > > * Yuanhan Liu wrote: > > > > > output with this patch: > > > > --- > > > > cpu 00: 0 0 ... 1 1 2 1 1 1 2 1 1 1 1 3 > > > > cpu 01: 0 0 ... 1 1 1 1 1 1 2 1

Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Michel Lespinasse
On Wed, Jan 30, 2013 at 1:14 AM, Yuanhan Liu wrote: > We(Linux Kernel Performance project) found a regression introduced by > commit 5a50508, which just convert all mutex lock to rwsem write lock. > The semantics is same, but the results is quite huge in some cases. > After investigation, we found

Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Ingo Molnar
* Yuanhan Liu wrote: > > > output with this patch: > > > --- > > > cpu 00: 0 0 ... 1 1 2 1 1 1 2 1 1 1 1 3 > > > cpu 01: 0 0 ... 1 1 1 1 1 1 2 1 1 1 1 3 > > > cpu 02: 0 0 ... 2 2 3 2 0 2 1

Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Yuanhan Liu
On Thu, Jan 31, 2013 at 10:39:31AM +0100, Ingo Molnar wrote: > > * Yuanhan Liu wrote: > > > We(Linux Kernel Performance project) found a regression introduced by > > commit 5a50508, which just convert all mutex lock to rwsem write lock. > > The semantics is same, but the results is quite huge in

Re: [PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-31 Thread Ingo Molnar
* Yuanhan Liu wrote: > We(Linux Kernel Performance project) found a regression introduced by > commit 5a50508, which just convert all mutex lock to rwsem write lock. > The semantics is same, but the results is quite huge in some cases. > After investigation, we found the root cause: mutex suppor

[PATCH] rwsem-spinlock: let rwsem write lock stealable

2013-01-30 Thread Yuanhan Liu
We(Linux Kernel Performance project) found a regression introduced by commit 5a50508, which just convert all mutex lock to rwsem write lock. The semantics is same, but the results is quite huge in some cases. After investigation, we found the root cause: mutex support lock stealing. Here is the lin