[PULL REQUEST] fixes for GCC 8.x, automatic RTnet interface dismantling - v2

2018-10-25 Thread Philippe Gerum
Fixed a locking imbalance in the commit dealing with extraneous UP/DOWN interface requests. The following changes since commit 04df93890befa4e5a642ecdef38eebed4660b217: rtdm: expose user-side file descriptors to drivers (2018-10-17 13:20:06 +0200) are available in the Git repository

Re: [PATCH] x86/process: ipipe: prevent false positive in testing for preemptible section

2018-10-25 Thread Alec Ari
Hahaha, that works too! Thank you for fixing this so soon, the right way! Alec

[PATCH] x86/process: ipipe: prevent false positive in testing for preemptible section

2018-10-25 Thread Philippe Gerum
We may call switch_to_bitmap() from the head stage when a co-kernel switches to a new thread context, in which case IRQs are hard disabled, although the root stage might be unstalled. Make sure the preemption debug code in refresh_tss_limit() is aware of this. --- arch/x86/include/asm/desc.h | 2

[PULL REQUEST] fixes for GCC 8.x, automatic RTnet interface dismantling

2018-10-25 Thread Philippe Gerum
This patch set addresses a couple of issues with GCC 8.x, and a current RTnet limitation too. Regarding the latter, I'd like a thorough review of 809847c58, to make sure it's not breaking some assumption wrt interface management in RTnet. The motivation for the change is to expose the necessary