Re: Interrupts vs signals - Mailing list pgsql-hackers

From Thomas Munro
Subject Re: Interrupts vs signals
Date
Msg-id CA+hUKGLAZCxzyhyoYz0BmGkqkmj-38ZfToF1GqNYhU+Br+XCLQ@mail.gmail.com
Whole thread Raw
In response to Re: Interrupts vs signals  (Thomas Munro <thomas.munro@gmail.com>)
Responses Re: Interrupts vs signals
List pgsql-hackers
Here's an updated version of this patch.

The main idea is that SendProcSignal(pid, PROCSIGNAL_XXX, procno)
becomes SendInterrupt(INTERRUPT_XXX, procno), and all the pending
interrupt global variables and pss_procsignalFlags[] go away, along
with the SIGUSR1 handler.  The interrupts are compressed into a single
bitmap.  See commit message for more details.

The patch is failing on Windows CI for reasons I haven't debugged yet,
but I wanted to share what I have so far.  Work in progress!

Here is my attempt to survey the use of signals and write down what I
think we might do about them all so far, to give the context for this
patch:

https://wiki.postgresql.org/wiki/Signals

Comments, corrections, edits very welcome.

Attachment

pgsql-hackers by date:

Previous
From: jian he
Date:
Subject: Re: SupportRequestRows support function for generate_series_timestamptz
Next
From:
Date:
Subject: RE: Doc: fix track_io_timing description to mention pg_stat_io