Re: [RFC PATCH-tip v4 02/10] locking/rwsem: Stop active read lock ASAP
On Thu, Oct 06, 2016 at 11:17:18AM -0700, Davidlohr Bueso wrote: > On Thu, 18 Aug 2016, Waiman Long wrote: > > >Currently, when down_read() fails, the active read locking isn't undone > >until the rwsem_down_read_failed() function grabs the wait_lock. If the > >wait_lock is contended, it may takes a while to get the lock. During > >that period, writer lock stealing will be disabled because of the > >active read lock. > > > >This patch will release the active read lock ASAP so that writer lock > >stealing can happen sooner. The only downside is when the reader is > >the first one in the wait queue as it has to issue another atomic > >operation to update the count. > > > >On a 4-socket Haswell machine running on a 4.7-rc1 tip-based kernel, > >the fio test with multithreaded randrw and randwrite tests on the > >same file on a XFS partition on top of a NVDIMM with DAX were run, > >the aggregated bandwidths before and after the patch were as follows: > > > > Test BW before patch BW after patch % change > > --- -- > > randrw1210 MB/s 1352 MB/s +12% > > randwrite 1622 MB/s 1710 MB/s +5.4% > > Yeah, this is really a bad workload to make decisions on locking > heuristics imo - if I'm thinking of the same workload. Mainly because > concurrent buffered io to the same file isn't very realistic and you > end up pathologically pounding on i_rwsem (which used to be until > recently i_mutex until Al's parallel lookup/readdir). Obviously write > lock stealing wins in this case. Except that it's DAX, and in 4.7-rc1 that used shared locking at the XFS level and never took exclusive locks. *However*, the DAX IO path locking in XFS has changed in 4.9-rc1 to match the buffered IO single writer POSIX semantics - the test is a bad test based on the fact it exercised a path that is under heavy development and so can't be used as a regression test across multiple kernels. If you want to stress concurrent access to a single file, please use direct IO, not DAX or buffered IO. Cheers, Dave. -- Dave Chinner da...@fromorbit.com -- To unsubscribe from this list: send the line "unsubscribe linux-doc" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [RFC PATCH-tip v4 02/10] locking/rwsem: Stop active read lock ASAP
On Mon, Oct 10, 2016 at 02:34:34AM -0700, Christoph Hellwig wrote: > On Mon, Oct 10, 2016 at 05:07:45PM +1100, Dave Chinner wrote: > > > > *However*, the DAX IO path locking in XFS has changed in 4.9-rc1 to > > > > match the buffered IO single writer POSIX semantics - the test is a > > > > bad test based on the fact it exercised a path that is under heavy > > > > development and so can't be used as a regression test across > > > > multiple kernels. > > > > > > That being said - I wonder if we should allow the shared lock on DAX > > > files IFF the user is specifying O_DIRECT in the open mode.. > > > > It should do - if it doesn't then we screwed up the IO path > > selection logic in XFS and we'll need to fix it. > > Depends on your defintion of "we". The DAX code has always abused the > direct I/O path, and that abuse is ingrained in the VFS path in a way that > we can't easily undo it in XFS, e.g. take a look at io_is_direct and > iocb_flags in include/linux/fs.h, which ensure that DAX I/O will always > appear as IOCB_DIRECT to the fs. Um, I seem to have completely missed that change - when did that happen and why? Oh, it was part of the misguided "enable DAX on block devices" changes - it was supposed to fix a problem with block device access using buffered IO instead of DAX (commit 65f87ee71852 "fs, block: force direct-I/O for dax-enabled block devices")). The original block device DAX code was reverted soon after this because it was badly flawed, but this change was not removed at the same time (probably because it was forgotton about). Hence I'd suggest that DAX check in io_is_direct() should be removed ASAP; the filesystems don't need it as they check the inode DAX state directly, and the code it "fixed" is no longer in the tree. Cheers, Dave. -- Dave Chinner da...@fromorbit.com -- To unsubscribe from this list: send the line "unsubscribe linux-doc" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [RFC PATCH-tip v4 02/10] locking/rwsem: Stop active read lock ASAP
On Sun, Oct 09, 2016 at 08:17:48AM -0700, Christoph Hellwig wrote: > On Fri, Oct 07, 2016 at 08:47:51AM +1100, Dave Chinner wrote: > > Except that it's DAX, and in 4.7-rc1 that used shared locking at the > > XFS level and never took exclusive locks. > > > > *However*, the DAX IO path locking in XFS has changed in 4.9-rc1 to > > match the buffered IO single writer POSIX semantics - the test is a > > bad test based on the fact it exercised a path that is under heavy > > development and so can't be used as a regression test across > > multiple kernels. > > That being said - I wonder if we should allow the shared lock on DAX > files IFF the user is specifying O_DIRECT in the open mode.. It should do - if it doesn't then we screwed up the IO path selection logic in XFS and we'll need to fix it. Cheers, Dave. -- Dave Chinner da...@fromorbit.com -- To unsubscribe from this list: send the line "unsubscribe linux-doc" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] xfs: Change URL for the project in xfs.txt
On Fri, Mar 02, 2018 at 04:08:24PM -0600, Eric Sandeen wrote: > > > On 3/2/18 3:57 PM, Dave Chinner wrote: > > On Fri, Mar 02, 2018 at 09:24:01AM -0800, Darrick J. Wong wrote: > >> On Fri, Mar 02, 2018 at 10:30:13PM +0900, Masanari Iida wrote: > >>> The oss.sgi.com doesn't exist any more. > >>> Change it to current project URL, https://xfs.wiki.kernel.org/ > >>> > >>> Signed-off-by: Masanari Iida <standby2...@gmail.com> > >>> --- > >>> Documentation/filesystems/xfs.txt | 2 +- > >>> 1 file changed, 1 insertion(+), 1 deletion(-) > >>> > >>> diff --git a/Documentation/filesystems/xfs.txt > >>> b/Documentation/filesystems/xfs.txt > >>> index 3b9b5c149f32..4d9ff0a7f8e1 100644 > >>> --- a/Documentation/filesystems/xfs.txt > >>> +++ b/Documentation/filesystems/xfs.txt > >>> @@ -9,7 +9,7 @@ variable block sizes, is extent based, and makes > >>> extensive use of > >>> Btrees (directories, extents, free space) to aid both performance > >>> and scalability. > >>> > >>> -Refer to the documentation at http://oss.sgi.com/projects/xfs/ > >>> +Refer to the documentation at https://xfs.wiki.kernel.org/ > > > > Did I miss a memo? > > About which part, the loss of oss.sgi or the addition of the kernel.org wiki? > > The kernel.org wiki is pretty bare though. OTOH xfs.org is a bit less > official. We really need to resolve this issue. Moving everything to kernel.org wiki. As I mentioned on IRC, I'd much prefer we move away from wiki's to something we can edit locally, review via email, has proper revision control and a "publish" mechanism that pushes built documentation out to the public website. Cheers, Dave. -- Dave Chinner da...@fromorbit.com -- To unsubscribe from this list: send the line "unsubscribe linux-doc" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH] xfs: Change URL for the project in xfs.txt
On Fri, Mar 02, 2018 at 09:24:01AM -0800, Darrick J. Wong wrote: > On Fri, Mar 02, 2018 at 10:30:13PM +0900, Masanari Iida wrote: > > The oss.sgi.com doesn't exist any more. > > Change it to current project URL, https://xfs.wiki.kernel.org/ > > > > Signed-off-by: Masanari Iida <standby2...@gmail.com> > > --- > > Documentation/filesystems/xfs.txt | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/Documentation/filesystems/xfs.txt > > b/Documentation/filesystems/xfs.txt > > index 3b9b5c149f32..4d9ff0a7f8e1 100644 > > --- a/Documentation/filesystems/xfs.txt > > +++ b/Documentation/filesystems/xfs.txt > > @@ -9,7 +9,7 @@ variable block sizes, is extent based, and makes extensive > > use of > > Btrees (directories, extents, free space) to aid both performance > > and scalability. > > > > -Refer to the documentation at http://oss.sgi.com/projects/xfs/ > > +Refer to the documentation at https://xfs.wiki.kernel.org/ Did I miss a memo? Cheers, Dave. -- Dave Chinner da...@fromorbit.com -- To unsubscribe from this list: send the line "unsubscribe linux-doc" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH v2 00/22] xfs-4.20: major documentation surgery
On Thu, Oct 11, 2018 at 11:27:35AM -0600, Jonathan Corbet wrote: > On Sat, 6 Oct 2018 10:51:54 +1000 Dave Chinner > wrote: > > > Can you let us know whether the CC-by-SA 4.0 license is > > acceptible or not? That's really the only thing that we need > > clarified at this point - if it's OK I'll to pull this into the > > XFS tree for the 4.20 merge window. If not, we'll go back to the > > drawing board > > OK, I've had a long conversation with the LF lawyer, and she said > clearly that we really should not be introducing CC-SA material > into the kernel source tree. It's a pain; I really do like CC-SA > better for documentation, but it's not GPL-compatible, and that > creates a problem for the processed docs. Ok, thanks for following upon this, Jon. We'll just keep it all in the existing external repo and work out what to do from there. Cheers, Dave. -- Dave Chinner da...@fromorbit.com
Re: [PATCH v2 00/22] xfs-4.20: major documentation surgery
On Wed, Oct 03, 2018 at 09:18:11PM -0700, Darrick J. Wong wrote: > Hi all, > > This series converts the existing in-kernel xfs documentation to rst > format, links it in with the rest of the kernel's rst documetation, and > then begins pulling in the contents of the Data Structures & Algorithms > book from the xfs-documentation git tree. No changes are made to the > text during the import process except to fix things that the conversion > process (asciidoctor + pandoc) didn't do correctly. The goal of this > series is to tie together the XFS code with the on-disk format > documentation for the features supported by the code. > > I've built the docs and put them here, in case you hate reading rst: > https://djwong.org/docs/kdoc/admin-guide/xfs.html > https://djwong.org/docs/kdoc/filesystems/xfs-data-structures/index.html > > I've posted a branch here because the png import patch is huge: > https://git.kernel.org/pub/scm/linux/kernel/git/djwong/xfs-linux.git/log/?h=docs-4.20-merge > > The patchset should apply cleanly against 4.19-rc6. Comments and > questions are, as always, welcome. Jon, Can you let us know whether the CC-by-SA 4.0 license is acceptible or not? That's really the only thing that we need clarified at this point - if it's OK I'll to pull this into the XFS tree for the 4.20 merge window. If not, we'll go back to the drawing board Cheers, Dave. -- Dave Chinner da...@fromorbit.com
Re: [PATCH v2 00/22] xfs-4.20: major documentation surgery
On Fri, Oct 05, 2018 at 07:01:20PM -0600, Jonathan Corbet wrote: > On Sat, 6 Oct 2018 10:51:54 +1000 > Dave Chinner wrote: > > > Can you let us know whether the CC-by-SA 4.0 license is acceptible > > or not? That's really the only thing that we need clarified at this > > point - if it's OK I'll to pull this into the XFS tree for the 4.20 > > merge window. If not, we'll go back to the drawing board > > I remain pretty concerned about it, to tell the truth. Rather than > continue to guess, though, I've called for help, and will be talking with > the LF lawyer about this next Thursday. Before then, I can't say anything > except "I don't think this works..." > > Will let you know what I hear. Thanks for the update, Jon. I'll put this on the backburner until I hear back from you. Cheers, Dave. -- Dave Chinner da...@fromorbit.com
Re: [PATCH] docs: improve readability for people with poorer eyesight
On Thu, Oct 04, 2018 at 06:06:03PM -0700, Darrick J. Wong wrote: > Hi, > > So my eyesight still hasn't fully recovered, so in the meantime it's > been difficult to read the online documentation. Here's some stylesheet > overrides I've been using to make it easier for me to read them: > https://djwong.org/docs/kdoc/index.html > > --- > From: Darrick J. Wong > > My eyesight is not in good shape, which means that I have difficulty > reading the online Linux documentation. Specifically, body text is > oddly small compared to list items and the contrast of various text > elements is too low for me to be able to see easily. > > Therefore, alter the HTML theme overrides to make the text larger and > increase the contrast for better visibility, and trust the typeface > choices of the reader's browser. > > For the PDF output, increase the text size, use a sans-serif typeface > for sans-serif text, and use a serif typeface for "roman" serif text. > > Signed-off-by: Darrick J. Wong This fixes problems I noticed when trying to review the built html documentation - the inconsistent font sizes on my high-dpi monitor made it almost impossible to read even though I have no eyesight problems Acked-by: Dave Chinner -Dave. -- Dave Chinner da...@fromorbit.com
