| Message ID | 20240701110317.99631-1-bigeasy@linutronix.de (mailing list archive) |
|---|---|
| Headers | show |
| Series | ARM: vfp: Introduce vfp_state_hold() to acquire VFP context. | expand |
On 2024-07-01 12:56:52 [+0200], To linux-arm-kernel@lists.infradead.org wrote: > There was a bug report on PREEMPT_RT which made me look into VFP locking > on ARM. The usage of local_bh_disable() to ensure exclusive access to > the VFP unit is not working on PREEMPT_RT because the softirq context is > preemptible. > > This series introduces vfp_vfp_state_hold() which does the right thing. is this okay now? Sebastian
On 2024-08-08 11:17:36 [+0200], To linux-arm-kernel@lists.infradead.org wrote: > On 2024-07-01 12:56:52 [+0200], To linux-arm-kernel@lists.infradead.org wrote: > > There was a bug report on PREEMPT_RT which made me look into VFP locking > > on ARM. The usage of local_bh_disable() to ensure exclusive access to > > the VFP unit is not working on PREEMPT_RT because the softirq context is > > preemptible. > > > > This series introduces vfp_vfp_state_hold() which does the right thing. > > is this okay now? The patches were reviewed by Ard and I fixed the remaining parts as suggested by Russell in my v5. Arnd suggested recently to submit them to the patch system if there is nothing outstanding which I just did. 9423/1 - 9426/1. Sebastian
There was a bug report on PREEMPT_RT which made me look into VFP locking on ARM. The usage of local_bh_disable() to ensure exclusive access to the VFP unit is not working on PREEMPT_RT because the softirq context is preemptible. This series introduces vfp_vfp_state_hold() which does the right thing. v5…v6: https://lore.kernel.org/all/20240628134648.1852041-1-bigeasy@linutronix.de/ - Rename vfp_{lock/unlock}() to vfp_state_{hold/release}() as suggested by Russell King. v4…v5: - Repost v3…v4: - Repost. v2…v3: https://lore.kernel.org/all/20230926125502.1127015-1-bigeasy@linutronix.de - Repost with Ard's Reviewed-by tag. v1…v2: https://lore.kernel.org/all/20230628080516.798032-1-bigeasy@linutronix.de - Split the last patch into two patches (3 and 4 of this series). Suggested by Ard. v1: https://lore.kernel.org/all/20230615101656.1308942-1-bigeasy@linutronix.de