From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Stancek Date: Wed, 18 Dec 2019 05:07:35 -0500 (EST) Subject: [LTP] [PATCH v2] Use real FS block size in fallocate05 In-Reply-To: <4f49ce5f-188e-2e68-d99c-5fb89acc30db@suse.cz> References: <20191213134002.GE20795@rei.lan> <20191217131703.16935-1-mdoucha@suse.cz> <1253407522.17207718.1576616558113.JavaMail.zimbra@redhat.com> <8e10a190-128f-cb5b-eeaa-716598411e45@suse.cz> <4f49ce5f-188e-2e68-d99c-5fb89acc30db@suse.cz> Message-ID: <461386048.17251961.1576663655809.JavaMail.zimbra@redhat.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ltp@lists.linux.it ----- Original Message ----- > On 12/18/19 10:09 AM, Martin Doucha wrote: > > On 12/17/19 10:02 PM, Jan Stancek wrote: > >> Is there an upstream thread link for that btrfs bug? > > > > Not yet. I've reported the bug to SUSE Bugzilla for now so that our > > kernel devs can take a closer look. But they may still decide that this > > is expected behavior. > > Update: Our kernel devs just said that this is expected behavior. Btrfs > will only release disk space when you deallocate a whole file extent > that can be up to 128MB in size. I'll have to add an exception for Btrfs > to fallocate05 then. Or would you guys prefer to always deallocate the > whole file before the last write() check? Since it's expected to work on other file systems, I'd add exception only for btrfs.