wake_up on wait-queue called from hard-irq context in 3.10.63-rt65

2016-04-27 Thread Andreas Starzer
This Bug was already fixed in rcutiny_plugin.h with changing the wait-queue to simple-waiter. Found this bug with 3.10.63-rt65 in rcutree_plugin.h too. (It is not fixed in current 3.10-release) SVC (hard-irq-context) for scheduling tries to wake-up wait-queue waiters and therefore simple-waiter

wake_up on wait-queue called from hard-irq context in 3.10.63-rt65

2016-04-27 Thread Andreas Starzer
This Bug was already fixed in rcutiny_plugin.h with changing the wait-queue to simple-waiter. Found this bug with 3.10.63-rt65 in rcutree_plugin.h too. (It is not fixed in current 3.10-release) SVC (hard-irq-context) for scheduling tries to wake-up wait-queue waiters and therefore simple-waiter

possible recursive locking detected n_tty_write vs. process_echoes

2015-02-06 Thread Andreas Starzer
With enabled deadlock detection the following warning occurs when connected via ssh (Dropbear server v2015.67). With "while :; do dmesg -c; done" the problem was fast reproduceable. System: Kernel with RT-patch applied: Linux version 3.10.63-rt65 (gcc version 4.9.2 (iMX6) ) #1 SMP PREEMPT RT

possible recursive locking detected n_tty_write vs. process_echoes

2015-02-06 Thread Andreas Starzer
With enabled deadlock detection the following warning occurs when connected via ssh (Dropbear server v2015.67). With while :; do dmesg -c; done the problem was fast reproduceable. System: Kernel with RT-patch applied: Linux version 3.10.63-rt65 (gcc version 4.9.2 (iMX6) ) #1 SMP PREEMPT RT This

Is it possible to get deterministic system with 3.10.xx and Full-RT-preemption?

2014-02-28 Thread Andreas Starzer
Is there a possible configuration/implementation at Kernel 3.10.xx with PREEMPT_RT_FULL=y (or newer) to get an explicit thread for workqueues which do exclusive work? I want to increase the priority of my drivers workqueue (and only this) to be still functional under heavy system load like

Is it possible to get deterministic system with 3.10.xx and Full-RT-preemption?

2014-02-28 Thread Andreas Starzer
Is there a possible configuration/implementation at Kernel 3.10.xx with PREEMPT_RT_FULL=y (or newer) to get an explicit thread for workqueues which do exclusive work? I want to increase the priority of my drivers workqueue (and only this) to be still functional under heavy system load like

corrupt mm-slab-wrap-functions

2013-09-30 Thread Andreas Starzer
Hello! I just tried to use the latest RT-patch from https://rt.wiki.kernel.org (patch-3.10.10-rt7.patch) and found a corrupt patch file "mm-slab-wrap-functions.patch" For example at line 165 in the attached patch file, the line had been duplicated. So I cannot trust the patch at all. :( I had

corrupt mm-slab-wrap-functions

2013-09-30 Thread Andreas Starzer
Hello! I just tried to use the latest RT-patch from https://rt.wiki.kernel.org (patch-3.10.10-rt7.patch) and found a corrupt patch file mm-slab-wrap-functions.patch For example at line 165 in the attached patch file, the line had been duplicated. So I cannot trust the patch at all. :( I had many