I wonder if we should also disable the prioritization APIs. In the normal,
default case, any reprioritization of an IRQ introduces a fatal, mysterious
bug. That is because nested interrupts will occur and may not be
supported.Sent from my Galaxy
-------- Original message --------From: Nathan Hartman
<hartman.nat...@gmail.com> Date: 12/9/24 10:11 PM (GMT-06:00) To:
dev@nuttx.apache.org Subject: Re: SERIOUS BUG: Zero Latency Interrupts are
broken! On Mon, Dec 9, 2024 at 9:46 PM Xiang Xiao <xiaoxiang781...@gmail.com>
wrote:>> On Tue, Dec 10, 2024 at 5:53 AM Nathan Hartman
<hartman.nat...@gmail.com>> wrote:>> > Hi all,> >> > Unfortunately I missed the
PR before it was merged, but PR-15073 has> > broken High Priority, Zero Latency
Interrupts! Fortunately I caught it> > now. It was merged 17 hours ago.> >> >>
Zero Latency Interrupts doesn't remove from PR-15073.>>> > This PR removed a
very important Kconfig: config ARMV7M_USEBASEPRI,> > and much of the logic
related to it. This config is *required* when> > using
CONFIG_ARCH_HIPRI_INTERRUPT.> >> >> PR-15073 doesn't remove ARMV7M_USEBASEPRI,
but makes it enabled> unconditionally, that's why this patch removes this
config.Thank you.I studied the changes of #15073 and #15086 and they seem
correct, toenable BASEPRI always in armv7m and armv8m. So, the Zero
LatencyInterrupts should work. I'll test it in the morning to be sure.tl;dr:
Please don't revert those PRs.I am still studying the #14881 changes. This one
is more complexbecause it changes how the recordkeeping is done.Unfortunately
it is very late now so I need to continue in the morning.>
https://github.com/apache/nuttx/pull/15102 removes ARMV7M_USEBASEPRI from>
documentation to match the new code base.Thanks for the PR. I will review
soon.Cheers,Nathan
