From: Maninder Singh Date: Wed, 21 Oct 2015 22:04:02 +0000 (+1100) Subject: ipc/msg.c: msgsnd: use freezable blocking call X-Git-Tag: KARO-TX6UL-2015-11-03~14^2~1 X-Git-Url: https://git.karo-electronics.de/?a=commitdiff_plain;h=3ec31afc4592cb5c383e9fbc0f96d2f397cec08d;p=karo-tx-linux.git ipc/msg.c: msgsnd: use freezable blocking call When any task is stuck in Interruptible or Uninterruptible state then waking up of that task fails. If wakeup fails, then suspend operation fails and all process send to frezeer state at this moment also gets wakeup. Correct implementation is that if suspend fails, then kernel would retry suspend operation again after some specific timeinterval for some fixed retry count. But as changes suggested by Mr Colin Cross (https://lkml.org/lkml/2013/5/1/424), for the system calls for which issue has been faced process flag being appended with PF_FREEZER_SKIP. We are testing some scenarios in which we have to do multi suspend-resume scenario, and we faced the problem, hence the suggested changes for msgsnd and msgrcv. Avoid waking up every thread sleeping in a msgsnd call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls. This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked. Signed-off-by: Vaneet narang Signed-off-by: Maninder Singh Cc: Yogesh Gaur Cc: Manjeet Pawar Cc: Ajeet Yadav Cc: Peter Zijlstra Cc: Tejun Heo Signed-off-by: Andrew Morton --- diff --git a/ipc/msg.c b/ipc/msg.c index 1471db9a7e61..3c99fcff874c 100644 --- a/ipc/msg.c +++ b/ipc/msg.c @@ -675,7 +675,7 @@ long do_msgsnd(int msqid, long mtype, void __user *mtext, ipc_unlock_object(&msq->q_perm); rcu_read_unlock(); - schedule(); + freezable_schedule(); rcu_read_lock(); ipc_lock_object(&msq->q_perm);