]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/commit
Revert r350841. I didn't realize that on this chip, reading the interrupt
authorian <ian@FreeBSD.org>
Sun, 11 Aug 2019 23:19:56 +0000 (23:19 +0000)
committerian <ian@FreeBSD.org>
Sun, 11 Aug 2019 23:19:56 +0000 (23:19 +0000)
commite00ee83cbc7de7a03d2911d07d8dbffd0c90fad5
treeebb0e5fbd1b3591c9995e3e99ce505428b1adee5
parentb27c98fb1e944170d8d7ca8baa34dda103176df2
Revert r350841.  I didn't realize that on this chip, reading the interrupt
status register clears pending interrupts.  By moving that code out of the
interrupt handler into a taskqueue task, I effectively created an interrupt
storm by returning from the handler with the interrupt source still active.

We'll have to find a different solution for this driver's need to sleep
in an ithread context.
sys/arm/ti/am335x/am335x_pmic.c