xtensa: gen_zsr: select software interrupt at <= EXCM_LEVEL

[williamtcdns]

The current logic selects a software interrupt even if it is at > EXCM_LEVEL.
However the arch_irq_lock() function only locks out interrupts upto EXCM_LEVEL.

In include/zephyr/arch/xtensa/irq.h -

/** Implementation of @ref arch_irq_lock. */
static ALWAYS_INLINE unsigned int arch_irq_lock(void) {
unsigned int key;
asm volatile("rsil %0, %1"
: "=r"(key) : "i"(XCHAL_EXCM_LEVEL) : "memory");
return key;
}

[williamtcdns]

Feedback on this PR ?

[williamtcdns]

@andyross @ceolin @nashif just one more review needed.

[andyross]

This sort of seems like an angels-on-a-pinhead argument? It's a software interrupt, it's triggered synchronously under control of the code being "interrupted". Is it really wrong that it works when interrupts are masked?

The purpose of this feature is to find a software interrupt we can use for testing purposes, e.g. for nested interrupts. And the test code doesn't mask interrupts around it currently, since it obviously works. So the effect of this patch is to disallow that on some number (maybe zero?) of hardware platforms, and therefore to reduce test coverage on those platforms.

Seems like a bad trade to me. Tentative -1, but I'm willing to be convinced if there's a bug somewhere that needs this as a fix.

[github-actions]

This pull request has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this pull request will automatically be closed in 14 days. Note, that you can always re-open a closed pull request at any time.