* [GIT PULL] dax-locking for 4.7
@ 2016-05-24 20:52 Ross Zwisler
2016-05-26 19:00 ` Vishal Verma
0 siblings, 1 reply; 2+ messages in thread
From: Ross Zwisler @ 2016-05-24 20:52 UTC (permalink / raw)
To: Linus Torvalds; +Cc: linux-fsdevel, linux-kernel, linux-nvdimm
Hi Linus, please pull from
git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm tags/dax-locking-for-4.7
... to receive filesystem DAX locking changes for v4.7.
These current version of these changes have been under review and test for a
few weeks, and have appeared in the two most recent linux-next releases
(next-20160520 and next-20160524). I've reviewed all of the DAX changes in
these patches, and Neil Brown has reviewed the patch that introduced the actual
locking and wait queues.
This pull request is dependent on Vishal Verma's "dax-misc for 4.7" pull
request.
The following changes since commit 40543f62cbdce42633e3fe10923099feee272e1f:
dax: fix a comment in dax_zero_page_range and dax_truncate_page (2016-05-18 12:16:58 -0600)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm tags/dax-locking-for-4.7
for you to fetch changes up to 4d9a2c8746671efbb0c27d3ae28c7474597a7aad:
dax: Remove i_mmap_lock protection (2016-05-19 15:28:40 -0600)
----------------------------------------------------------------
Filesystem DAX locking for 4.7
- We use a bit in an exceptional radix tree entry as a lock bit and use it
similarly to how page lock is used for normal faults. This fixes races
between hole instantiation and read faults of the same index.
- Filesystem DAX PMD faults are disabled, and will be re-enabled when PMD
locking is implemented.
----------------------------------------------------------------
Jan Kara (7):
dax: Fix condition for filling of PMD holes
dax: Make huge page handling depend of CONFIG_BROKEN
dax: Define DAX lock bit for radix tree exceptional entry
dax: Allow DAX code to replace exceptional entries
dax: New fault locking
dax: Use radix tree entry lock to protect cow faults
dax: Remove i_mmap_lock protection
fs/Kconfig | 1 +
fs/dax.c | 592 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++--------------------------
include/linux/dax.h | 16 ++-
include/linux/mm.h | 7 ++
mm/filemap.c | 30 +++--
mm/memory.c | 40 +++----
mm/truncate.c | 62 +++++-----
7 files changed, 513 insertions(+), 235 deletions(-)
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [GIT PULL] dax-locking for 4.7
2016-05-24 20:52 [GIT PULL] dax-locking for 4.7 Ross Zwisler
@ 2016-05-26 19:00 ` Vishal Verma
0 siblings, 0 replies; 2+ messages in thread
From: Vishal Verma @ 2016-05-26 19:00 UTC (permalink / raw)
To: Linus Torvalds, Ross Zwisler
Cc: Linus Torvalds, linux-fsdevel, linux-kernel, linux-nvdimm
On Tue, May 24, 2016 at 02:52:19PM -0600, Ross Zwisler wrote:
> Hi Linus, please pull from
>
> git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm tags/dax-locking-for-4.7
>
> ... to receive filesystem DAX locking changes for v4.7.
>
> These current version of these changes have been under review and test for a
> few weeks, and have appeared in the two most recent linux-next releases
> (next-20160520 and next-20160524). I've reviewed all of the DAX changes in
> these patches, and Neil Brown has reviewed the patch that introduced the actual
> locking and wait queues.
>
> This pull request is dependent on Vishal Verma's "dax-misc for 4.7" pull
> request.
>
> The following changes since commit 40543f62cbdce42633e3fe10923099feee272e1f:
>
> dax: fix a comment in dax_zero_page_range and dax_truncate_page (2016-05-18 12:16:58 -0600)
>
> are available in the git repository at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm tags/dax-locking-for-4.7
>
> for you to fetch changes up to 4d9a2c8746671efbb0c27d3ae28c7474597a7aad:
>
> dax: Remove i_mmap_lock protection (2016-05-19 15:28:40 -0600)
>
> ----------------------------------------------------------------
> Filesystem DAX locking for 4.7
>
> - We use a bit in an exceptional radix tree entry as a lock bit and use it
> similarly to how page lock is used for normal faults. This fixes races
> between hole instantiation and read faults of the same index.
>
> - Filesystem DAX PMD faults are disabled, and will be re-enabled when PMD
> locking is implemented.
>
> ----------------------------------------------------------------
Hi Linus,
Both this pull request, and mine ([GIT PULL] dax-misc for 4.7) have
minor merge conflicts with your tree. These were caught by linux-next
too (in addition to the one I mentioned in my pull request), but I
neglected to mention this one [1].
[1] https://lkml.org/lkml/2016/5/19/707
I've updated the branch at:
git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm libnvdimm-for-4.7-merge
with the expected resolution of merging both dax-misc-for-4.7 and
dax-locking-for-4.7
Let me know if you run into something odd, and apologies for not giving
an earlier heads up about the conflict.
-Vishal
> Jan Kara (7):
> dax: Fix condition for filling of PMD holes
> dax: Make huge page handling depend of CONFIG_BROKEN
> dax: Define DAX lock bit for radix tree exceptional entry
> dax: Allow DAX code to replace exceptional entries
> dax: New fault locking
> dax: Use radix tree entry lock to protect cow faults
> dax: Remove i_mmap_lock protection
>
> fs/Kconfig | 1 +
> fs/dax.c | 592 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++--------------------------
> include/linux/dax.h | 16 ++-
> include/linux/mm.h | 7 ++
> mm/filemap.c | 30 +++--
> mm/memory.c | 40 +++----
> mm/truncate.c | 62 +++++-----
> 7 files changed, 513 insertions(+), 235 deletions(-)
>
> _______________________________________________
> Linux-nvdimm mailing list
> Linux-nvdimm@lists.01.org
> https://lists.01.org/mailman/listinfo/linux-nvdimm
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2016-05-26 19:00 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-05-24 20:52 [GIT PULL] dax-locking for 4.7 Ross Zwisler
2016-05-26 19:00 ` Vishal Verma
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).