Imagine you have a scenario where you acquire and hold some lwlocks before
PG_TRY/PG_CATCH block, and then you try to release the locks in PG_CATCH as
part of your cleanup. Innocent enough?
However, you are likely to get this:
TRAP: failed Assert("InterruptHoldoffCount > 0")
That’s because errfinish sets it to zero, but releasing lwlocks requires the
count to be greater than zero, as you can see above.
However, this is what errfinish implementation tells us:
/*
* We do some minimal cleanup before longjmp'ing so that handlers can
* execute in a reasonably sane state.
*
* Reset InterruptHoldoffCount in case we ereport'd from inside an
* interrupt holdoff section. (We assume here that no handler will
* itself be inside a holdoff section. If necessary, such a handler
* could **save and restore InterruptHoldoffCount** for itself, but this
* should make life easier for most.)
*/
InterruptHoldoffCount = 0;So, we can perhaps use this as an invitation to save InterruptHoldoffCount
before going into PG_TRY, even though it doesn't feel perfect – if you really
need this.