From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.gentoo.org (woodpecker.gentoo.org [140.211.166.183]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id B2B5E5662 for ; Tue, 26 Dec 2023 20:50:25 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gentoo.org Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=gentoo.org References: <20231215013657.1995699-1-sam@gentoo.org> <20231215013657.1995699-2-sam@gentoo.org> User-agent: mu4e 1.10.8; emacs 30.0.50 From: Sam James To: Christoph Hellwig Cc: Sam James , linux-xfs@vger.kernel.org, Felix Janda Subject: Re: [PATCH v3 2/4] io: Assert we have a sensible off_t Date: Tue, 26 Dec 2023 20:49:31 +0000 Organization: Gentoo In-reply-to: Message-ID: <87v88k1yeq.fsf@gentoo.org> Precedence: bulk X-Mailing-List: linux-xfs@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain Christoph Hellwig writes: > On Fri, Dec 15, 2023 at 01:36:41AM +0000, Sam James wrote: >> + /* We're only interested in supporting an off_t which can handle >=4GiB. */ > > This adds a < 80 character line. Also I find the wording a bit odd, the > point is that xfsprogs relies on (it or rather will with your entire > series), so maybe: > > /* > * xfsprogs relies on the LFS interfaces with a 64-bit off_t to > * actually support sensible file systems sizes. > */ > > And while I'm nitpicking, maybe a better place would be to move this to > libxfs as that's where we really care. If you use the C99 static_assert > instead of the kernel BUILD_BUG_ON this can even move outside a function > and just into a header somewhere, say include/xfs,h. Which actually > happens to have this assert in an awkware open coded way already: > > /* > * make sure that any user of the xfs headers has a 64bit off_t type > */ > extern int xfs_assert_largefile[sizeof(off_t)-8]; > > Enough of my stream of consciousness, sorry. To summarize the findings: > > - we don't really need this patch all > - but cleaning up xfs_assert_largefile to just use static_assert would > probably be nice to have anyway Thanks, I agree, but I think static_assert is C11 (and then it gets a nicer name in C23). If it's still fine for us, I can then use it. Does it change your thinking at all or should I send a v4 with it included? Thanks, sam