From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Return-Path: Date: Sat, 29 Aug 2020 22:34:17 +0200 From: Martin Bukatovic Subject: Re: [PATCH] Add fill_quota option Message-ID: <20200829203417.GA18208@dione.lan> References: <20200811194445.GA152146@dione.lan> <20200829190131.GA8262@dione.lan> <3e8f33c3-6d57-8932-b84e-374c0afd4779@kernel.dk> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <3e8f33c3-6d57-8932-b84e-374c0afd4779@kernel.dk> To: Jens Axboe Cc: fio@vger.kernel.org List-ID: On Sat, Aug 29, 2020 at 01:44:07PM -0600, Jens Axboe wrote: > On 8/29/20 1:01 PM, Martin Bukatovic wrote: > > On Fri, Aug 28, 2020 at 05:53:34PM -0600, Jens Axboe wrote: > >> Might be easier and saner to just accept EDQUOT by default, instead of > >> adding an option for it? > > > > I proposed a new option to avoid changing behaviour of fill_device > > option. But if that is not necessary, I agree that making fill_device to > > stop on both ENOSPC and EDQUOT makes more sense. > > I don't think that's necessary, the option is really "stop writing when > we're full", and there's no way to reach ENOSPC before EDQUOT vs > the other way around. > > Hence I think it's totally fine to just include EDQUOT in the "stop > successfully" part of the fill_device behavior. Ok, I'm sending new version of the patch "Make fill_device to stop writing on EDQUOT". -- Martin Bukatovič