yes, please tell us how u've thought of detecting the deadlock situation. On 6/9/06, Ohad Ben-Cohen <ohadbc@xxxxxxxxx> wrote:
Hi. On 6/8/06, Vincenzo Mallozzi <vinjunior@xxxxxxxx> wrote: ... > I've to know when these threads are in a deadlock > situation. > To do this, I get the task struct of each threads and > I need to know what is their state when they are in > such a situation. ... Hmmm.. I disagree. There is no general connection between the task state and whether a process is in deadlock or not. If a task is not running it _may_ indicate it is waiting in a semaphore, yes, but it's _not_ an evidence of it being deadlocked. a deadlock is a logical state. when you wait on a lock you can be either in a deadlock or not - but either way your task state is the same. > I don't know if there's another way to get this > information. please give us the full picture of the problem. maybe more details will shed some light... cheers, Ohad. > > Vincenzo. > > --- Jinesh K J <jineshkj.newsletters@xxxxxxxxx> ha > scritto: > > > once u get the task_struct of the deadlocked > > process, u can get the > > state from the 'state' field in the structure. BTW, > > how are u going to > > find out which process has deadlocked? > > > > > > On 6/8/06, Ohad Ben-Cohen <ohadbc@xxxxxxxxx> wrote: > > > On 6/8/06, Vincenzo Mallozzi <vinjunior@xxxxxxxx> > > wrote: > > > > Is there a deterministic way to know if a > > process is in deadlock have state > > > > TASK_INTERRUPTIBLE or TASK_UNINTERRUPTIBLE? > > > > I really need this information to know if some > > threads are in such a > > > > situation. > > > > How can I resolve this problem? Is there another > > solution? > > > > > > It's a little weird question. can you please > > elaborate more on the problem ? > > > what's going on ? why do you need this ? > > > > > > if you think you have deadlocks, why not just try > > to solve them ? > > > anyway, > > > you might be interested to examine a recent kernel > > patch posted by Ingo Molnar, > > > which is called "the kernel lock validator". it > > essentially hunts down possible > > > deadlocks in the kernel. read on in > > http://lwn.net/Articles/185666/ > > > > > > Ohad. > > > > > > > Thanks. > > > > Vincenzo > > > > > > > > On Wednesday 07 June 2006 15:16, William J Beksi > > wrote: > > > > > Vincenzo Mallozzi wrote: > > > > > > > > > > >Hi, > > > > > >what is the task state of some > > processes/threads that are in a deadlock > > > > > >situation? I've heard about > > TASK_UNINTERRUPTIBLE. Is not it? > > > > > > > > > > > > > > > > > In a deadlock situation both tasks are waiting > > for a resource that is > > > > > held by another task, therefore none of the > > other tasks waiting for that > > > > > resource can proceed. In > > TASK_INTERRUPTIBLE/TASK_UNINTERRUPTIBLE the > > > > > task is in a sleeping state with the exception > > that TASK_UNINTERRUPTIBLE > > > > > cannot be woken up by a signal. So it seems a > > deadlock can occur with a > > > > > task in an interruptible or uniterruptible > > state. > > > > > > > > > > William > > > > > > > > > > > > > > > > > ___________________________________ > > > > Yahoo! Messenger with Voice: chiama da PC a > > telefono a tariffe esclusive > > > > http://it.messenger.yahoo.com > > > > > > > > > > > > -- > > > > Kernelnewbies: Help each other learn about the > > Linux kernel. > > > > Archive: > > http://mail.nl.linux.org/kernelnewbies/ > > > > FAQ: http://kernelnewbies.org/faq/ > > > > > > > > > > > > > > -- > > > Kernelnewbies: Help each other learn about the > > Linux kernel. > > > Archive: > > http://mail.nl.linux.org/kernelnewbies/ > > > FAQ: http://kernelnewbies.org/faq/ > > > > > > > > > > > Chiacchiera con i tuoi amici in tempo reale! > http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com >
-- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
