Messages in this thread | | | Date | Thu, 11 Mar 2010 17:13:39 +0800 | From | Yong Zhang <> | Subject | Re: [RFC][PATCH] IRQ: Fix oneshot irq race between irq_finalize_oneshot and handle_level_irq |
| |
On Thu, Mar 11, 2010 at 09:41:45AM +0100, Thomas Gleixner wrote: > No, it is _NOT_ harmless. interrupt handlers are guaranteed _NOT_ to > be reentered. That's why we check the INPROGRESS flag. > > And you totally miss the case which caused the disussion in the first > place: When the thread finishes _before_ the hard irq handler on the > other CPU has set IRQ_MASKED. So this patch solves nothing at all, it > just makes stuff worse than it was. > > Here is the solution which solves the inconsistent lock state _AND_ > the reentrancy race. > > http://git.kernel.org/?p=linux/kernel/git/tip/linux-2.6-tip.git;a=commit;h=0b1adaa031a55e44f5dd942f234bf09d28e8a0d6 > > The change log has a full explanation of the scenarios.
Yeah, I see. Thanks for your point. It will smooth the concerns.
Thanks, Yong
| |