Horton

Re: fix pg_mkdir_p to tolerate concurrent directory creation

Andrew Dunstan <andrew@dunslane.net>

View thread View on postgr.es
From: Andrew Dunstan <andrew@dunslane.net>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2026-06-18T15:37:49Z
Lists: pgsql-hackers

Attachments

On 2026-06-18 Th 10:24 AM, Tom Lane wrote:
> Andrew Dunstan <andrew@dunslane.net> writes:
>> While working on the pytest stuff, I found this issue when making it
>> work on Windows, but the issue can exist everywhere. pg_mkdir_p can fail
>> if there is a concurrent directory creation.
> This bit:
>
> +			if (errno != EEXIST || stat(path, &sb) != 0 || !S_ISDIR(sb.st_mode))
> +			{
> +				retval = -1;
> +				break;
> +			}
>
> looks like it could corrupt the reported errno, ie stat() could
> overwrite what mkdir() reported.
>
> 			


Oh, good catch, thanks! Here's a revised patch.


cheers


andrew


--
Andrew Dunstan
EDB: https://www.enterprisedb.com

Commits

  1. Make pg_mkdir_p() tolerant of a concurrent directory creation.