... reason for this asymetry? Shouldn't the acquiring functions > be implemented in the very same ... .e. is there any particular reason why we don't have something like ... just to demonstrate what I mean)? spinlocks: inline spin_lock() in non-debug cases ...

linux-kernel - Jiri Kosina - Mar 31 2008 - 20:40

... for this asymetry? Shouldn't the acquiring > > > functions be implemented in the ... also > have the CONFIG_OPTIMIZE_INLINING=y and CONFIG_CC_OPTIMIZE_FOR_SIZE=y > combination as well, ... degradation of mmap() when spinlocks are not inlined. I am going ...

linux-kernel - Jiri Kosina - Apr 1 2008 - 04:38

... % performance degradation of mmap() when > > spinlocks are not inlined. I am going to do ... is any particular reason behind that. > inlining decisions almost never have effects of ... suspicious to me. Definitely, that's why I am a little bit suspicious. ...

linux-kernel - Jiri Kosina - Apr 1 2008 - 04:44

... performance degradation of mmap() when > spinlocks are not inlined. I am going to ... have noticed the aforementioned asymetry in spinlock.h, so I > just wanted to ... is any particular reason behind that. inlining decisions almost never have effects of ...

linux-kernel - Ingo Molnar - Apr 1 2008 - 04:43

... degradation of mmap() when spinlocks are not > inlined. I am going ... the actual spin unlock operation and for x86-64 it was ... call has on the caller. Spinlocks don't clobber a lot ... the radar. Anyways without queued spinlocks that has probably changed again ...

linux-kernel - Andi Kleen - Apr 1 2008 - 07:12

... for this asymetry? Shouldn't the acquiring > > functions be implemented in the very ... be inlined? > > i.e. is there any particular reason why we don ... ll also have the CONFIG_OPTIMIZE_INLINING=y and CONFIG_CC_OPTIMIZE_FOR_SIZE=y combination as well, ...

linux-kernel - Ingo Molnar - Apr 1 2008 - 04:33

... code than the actual spin unlock operation and for x86-64 it > > was about the ... > spin unlocks seem to be properly inlined anyway, so that should be fine. ... counting register allocator effects though which are difficult to measure generally. -Andi -- ...

linux-kernel - Andi Kleen - Apr 1 2008 - 09:30

... larger in code than the actual spin unlock operation and for x86-64 it > was about the same. spin unlocks seem to be properly inlined anyway, so that should be fine. My concern here is the non-inlining of spin locks, for which I don't think ...

linux-kernel - Jiri Kosina - Apr 1 2008 - 08:40