|
From: Nicholas N. <nj...@ca...> - 2002-10-30 15:38:56
|
Hi, Jeremy: I saw your patch #33 which adds a pre_mutex_lock event. I think it's a good thing to add to core. Would you mind augmenting your patch so that it adds a pre_mutex_unlock() event too? It would be good to have for completeness but I'm not sure where the hooks need to go -- my understanding of the scheduler and threads stuff is poor. Thanks. -- Nick Nethercote nj...@ca... |
|
From: Jeremy F. <je...@go...> - 2002-10-30 16:03:20
|
On Wed, 2002-10-30 at 07:38, Nicholas Nethercote wrote:
Hi,
Jeremy: I saw your patch #33 which adds a pre_mutex_lock event. I think
it's a good thing to add to core. Would you mind augmenting your patch so
that it adds a pre_mutex_unlock() event too? It would be good to have for
completeness but I'm not sure where the hooks need to go -- my
understanding of the scheduler and threads stuff is poor.
The main reason for pre_mutex_lock is that locking can block, so you may
want to do something before that happens. Unlock doesn't block, so pre_
or post_ doesn't matter much. I didn't bother with pre_mutex_unlock
because it would be redundant.
J
|
|
From: Nicholas N. <nj...@ca...> - 2002-10-30 16:07:36
|
On 30 Oct 2002, Jeremy Fitzhardinge wrote: > The main reason for pre_mutex_lock is that locking can block, so you may > want to do something before that happens. Unlock doesn't block, so pre_ > or post_ doesn't matter much. I didn't bother with pre_mutex_unlock > because it would be redundant. Ok, a brief comment explaining this in vg_skin.h would be good then :) N |