Date: Thu, 11 Jul 2024 14:56:44 -0500
From: Nathan Bossart <nathandbossart@gmail.com>
To: Sami Imseih <samimseih@gmail.com>
Cc: Bertrand Drouvot <bertranddrouvot.pg@gmail.com>,
	Tom Lane <tgl@sss.pgh.pa.us>, Robert Haas <robertmhaas@gmail.com>,
	"pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Subject: Re: Restart pg_usleep when interrupted
Message-ID: <ZpA4_Pw14FCP6025@nathan>
References: <7D50DC5B-80C6-47B5-8DA8-A6C68A115EE5@gmail.com>
 <2100439.1719610468@sss.pgh.pa.us>
 <AC328D21-21D6-4FFC-8F6B-BFE9AD940524@gmail.com>
 <ZoPC5IeP4k7sZpki@ip-10-97-1-34.eu-west-3.compute.internal>
 <373D199B-2312-431D-8018-F2808913F979@gmail.com>
 <Zo0UdeE3i9d0Wt5E@ip-10-97-1-34.eu-west-3.compute.internal>
 <3712E898-03DA-4E03-AF51-D3E24DE68216@gmail.com>
 <C18017A1-EDFD-4F2F-BCA1-0572D4CCC92B@gmail.com>
 <Zo_7blmO166ZDGbW@nathan>
 <2F781744-69B8-4BC2-A6EF-3998DCCDB634@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <2F781744-69B8-4BC2-A6EF-3998DCCDB634@gmail.com>
Archived-At: <https://www.postgresql.org/message-id/ZpA4_Pw14FCP6025%40nathan>
Precedence: bulk

On Thu, Jul 11, 2024 at 01:10:25PM -0500, Sami Imseih wrote:
>> I'm curious why we wouldn't just subtract "elapsed_time" from "delay" at
>> the bottom of the while loop to avoid needing this extra check.  
> 
> Can you elaborate further? I am not sure how this will work since delay is a timespec 
> and elapsed time is an instr_time. 
> 
> Also, in every iteration of the loop, the delay must be set to the remaining time. The
> purpose of the elapsed_time is to make sure that we don´t surpass requested time
> delay as an additional safeguard.

I'm imagining something like this:

    struct timespec delay;
    TimestampTz end_time;

    end_time = TimestampTzPlusMilliseconds(GetCurrentTimestamp(), msec);

    do
    {
        long        secs;
        int         microsecs;

        TimestampDifference(GetCurrentTimestamp(), end_time,
                            &secs, &microsecs);

        delay.tv_sec = secs;
        delay.tv_nsec = microsecs * 1000;

    } while (nanosleep(&delay, NULL) == -1 && errno == EINTR);

-- 
nathan