Print

Print


There seems to be another issue nearby ... but it has to do with Pfc and async reads.

It can happen that async posix read handler decides to drop the outstanding reads when they are delayed beyond certain timeout and call this Close() regardless of that.

This should not really happen when in pfc mode as the outstanding async read calls will eventually invoke the pfc callbacks (either successful or failed / timed out) on a potentially deleted IO object.

It is unclear if the reference count check should be performed at the async delayed request handler before Close() is called.

XrdPosixXrootd::Close() calling Detach() before checking the actual reference count potentially send IO object into a too early shutdown.


Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <xrootd/xrootd/pull/1724/c1164983628@github.com>

[ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/xrootd/xrootd/pull/1724#issuecomment-1164983628", "url": "https://github.com/xrootd/xrootd/pull/1724#issuecomment-1164983628", "name": "View Pull Request" }, "description": "View this Pull Request on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

Use REPLY-ALL to reply to list

To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1