Re: Non-reproducible AIO failure

On 10/06/2025 8:41 pm, Andres Freund wrote:
> I was able to reproduce it with gcc, too.
> I've reproduced it without that bitfield, unfortunately :(.

But also only at MacOS?

I wonder if it is possible to set hardware watchpoint fro program itself
(not using gdb)? I.e. using ptrace?
Looks like it is not possible to debug yourself:
https://stackoverflow.com/questions/64321402/adding-a-watchpoint-in-the-current-process-not-in-gdb-not-for-debugging

but it is possible to fork process.
In theory it certainly should be possible - gdb is normal process, so at
least we can implement our mini-gdb.
But not sure how complex it will be.

It will be nice to set watchpoint on for example "op" in
`pgaio_io_reclaim` and disable it in `pgaio_io_before_start`.
It may have less overhead (both memory and CPU) than using `mprotect` to
detect illegal access.

But I afraid that if mempory corruption happens in kernal mode then
memory protection or watchpoints will not help.