I want to modify spin_lock & spin_unlock API in spinlock.h of 2.6.36.4. I want to add a counter for each core so that every time a lock is taken on a core, its counter in incremented and decremented when spin_lock is called. At any point of time I can get lock_depth of each core.
I tried doing this by adding a per CPU variable.
I am modifying the lock & unlock API of 2.6.36 linux.
I want to add a per cpu counter, such that every time some thread does a spin_lock I increment this counter and decrement on spin_unlock.
I tried adding DEFINE_PER_CPU(int, counter) in /usr/src/linux-2.6**/include/linux/spinlock.h but I get errors since the percpu.h #includes spinlock.h.
I also tried creating a static array and use get_cpu(
Waiting for deletion of this thread. I posted a homework thread in the wrong place. Sorry for any inconvenience. I do not have privileges to delete this thread.
Hi All,
I have a core of an multi-threaded application. when i check all the threads it shows me that many threads are blocked for taking a mutex lock. In GDB how I can identify that thread which is holding the mutex lock.
In dbx I can do that using thread -blockedby but in GDB i am not getting how I can do that.
Hummmmm, I looked over related issue but not sure of the answers.
Here is what the book says on page 95:
To delete a message thread from the Messaging window
1. Touch and hold the message thread.
2. Touch Delete thread in the Context menu that opens.
3.
I am trying to assign a routine to the posix thread. When it is completing the routine, the calling thread is not exiting. Every time i am creating a new thread and assigning a same routine to the new thread in while loop. New thread is created only for 379 times, after that thread creation is failed with status 11 means PTHREAD_MAX_LIMIT has been reached.
Hello there
Here it says that this operation is thread safe, however I'm not quite sure about how to interpret that.
Since this particular call doesn't accept a "tag" parameter, I think it is NOT thread safe; for instance, if we have 10 process with 3 threads each and each thread calls MPI_Allreduce...
Can we be sure that the calls will not get mixed?
Anyone knows an alterna
brian_panneton wrote:I keep getting a kernel dump ( generally when using flash or downloading a file ). It may have something to do with conky since I'm pretty sure it shouldn't be able to schedule while atomic. If I could get hints on whats happening that would be great.No, this has nothing to do with conky. Conky is a userspace process which is supposed to never crash the system.
Please do not post Rom specific questions here, ask such questions in the respective ROM thread only.
Feel free to post suggestions, any changes or my mistakes ;)
Still under-construction..........
********************TW based ICS ROMS*******************AlBa Syndicate by xaoc747 - Thread
AllianceROM N7000 ICS byRobbie Hood - Thread
Cassie's XtraLiteRom ICS by thering1975 - Thread
Chra