On Tue 06 Oct 2020 at 17:33, Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote: > Brad, > > On Wed, Oct 07 2020 at 00:45, Brad Harper wrote: >> I'm happy to test anything on a range of amlogic hardware with standard >> / rt and multiple mmc devices. Ill test Jerome's patch in next 24 >> hours to report the results. > > please do not top-post and trim your replies. > >> On 6/10/2020 11:43 pm, Thomas Gleixner wrote: >>> We rather should make interrupts which need to have their primary >>> handler in hard interrupt context to set IRQF_NO_THREAD. That >>> should at the same time confirm that the primary handler is RT >>> safe. >>> >>> Let me stare at the core code and the actual usage sites some more. > > So there are a few nasties in there and I faintly remember that there > was an assumption that interrupts which are requested with both a > primary and a secondary handler should quiesce the device interrupt in > the primary handler if needed. OTOH, this also enforces that the primary > handler is RT safe, which is after a quick scan of all the usage sites > not a given and quite some of the users rely on IRQF_ONESHOT. > > The below untested patch should cure the problem and keep the interrupt > line masked if requested with IRQF_ONESHOT. > With arm64 defconfig on Khadas vim3, no obvious regression. Looks good. Tested-by: Jerome Brunet <jbrunet@xxxxxxxxxxxx> I did not test with RT. Brad, Could you let us know is Thomas's patch works for you ? Thx
