Date: Mon, 4 May 2015 12:51:16 +0300 From: Konstantin Belousov <kostikbel@gmail.com> To: Gleb Smirnoff <glebius@FreeBSD.org> Cc: alc@FreeBSD.org, arch@FreeBSD.org Subject: Re: more strict KPI for vm_pager_get_pages() Message-ID: <20150504095116.GF2390@kib.kiev.ua> In-Reply-To: <20150504091137.GH34544@glebius.int.ru> References: <20150430142408.GS546@nginx.com> <20150504082426.GC2390@kib.kiev.ua> <20150504091137.GH34544@glebius.int.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, May 04, 2015 at 12:11:37PM +0300, Gleb Smirnoff wrote: > On Mon, May 04, 2015 at 11:24:26AM +0300, Konstantin Belousov wrote: > K> Below is the summary of my part of the internal discussion about the changes. > > Quite short. Is it truncated? No. IMO, I pointed out the most important point about the patch. If other changes in the patch are unrelated, they must be extracted and discussed (and committed) separately. Due to the fundamental nature of the code being changed, the extra work to make it easier to bisect and detect regressions worth it. > -- > Totus tuus, Glebius. > > K> Traditionally, Unix allows the filesystems to perform the short reads. > K> Most fundamental change in the patch removes this freedom from the > K> filesystem implementation, and I think that only local filesystems could > K> be compliant with the proposed strictness. > K> > K> IMO, the response from vm_pager_haspages() is only advisory, since > K> filesystem might not control the external entities which are the source > K> of the required data. > > That's why remote filesystems use vop_stdbmap() (or similar), which > always return zeroes for "after" and "before" hints. Which precludes useful optimizations, at all, in the future.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20150504095116.GF2390>