On Mon, Aug 23, 2021 at 11:45:52AM -0700, Dan Williams wrote:
> On Mon, Aug 23, 2021 at 5:37 AM Christoph Hellwig wrote:
> >
> > Rename the main option text to clarify it is for file system access,
> > and add a bit of text that explains how to actually switch a nvdimm
> > to a fsdax capable state
On 23.08.21 16:05, Gerald Schaefer wrote:
On Fri, 20 Aug 2021 07:43:40 +0200
Christoph Hellwig wrote:
Hi all,
looking at the recent ZONE_DEVICE related changes we still have a
horrible maze of different code paths. I already suggested to
depend on ARCH_HAS_PTE_SPECIAL for ZONE_DEVICE there,
On Mon, Aug 23, 2021 at 02:15:47PM -0700, Dan Williams wrote:
> > +static inline bool dax_supported(struct dax_device *dax_dev,
> > + struct block_device *bdev, int blocksize, sector_t start,
> > + sector_t len)
> > +{
> > + return false;
> > +}
>
> I've started c
>
> On Mon, 2021-08-23 at 09:48 -0700, Dan Williams wrote:
> > On Sun, Aug 22, 2021 at 1:36 PM QI Fuli wrote:
> > >
> > > Import ciniparser from ccan[1].
> > > [1] https://ccodearchive.net/info/ciniparser.html
> >
> > Hi Qi,
> >
> > Vishal points out that an updated version of this code is now
>
On Mon, Aug 23, 2021 at 5:45 AM Christoph Hellwig wrote:
>
> All callers already have a dax_device obtained from fs_dax_get_by_bdev
> at hand, so just pass that to dax_supported() insted of doing another
> lookup.
Looks good, series passes regression tests:
Reviewed-by: Dan Williams
I can take
On Mon, Aug 23, 2021 at 5:44 AM Christoph Hellwig wrote:
>
I wouldn't mind adding:
"In support of a future change to kill bdev_dax_supported() first move
its usage to a helper."
...but either way:
Reviewed-by: Dan Williams
On Mon, Aug 23, 2021 at 5:43 AM Christoph Hellwig wrote:
>
> dax_supported calls into ->dax_supported which checks for fsdax support.
> Don't bother building it for !CONFIG_FS_DAX as it will always return
> false.
>
Looks good, modulo formatting question below:
Reviewed-by: Dan Williams
> diff
On Mon, Aug 23, 2021 at 5:42 AM Christoph Hellwig wrote:
>
> Just implement generic_fsdax_supported directly out of line instead of
> adding a wrapper. Given that generic_fsdax_supported is only supplied
> for CONFIG_FS_DAX builds this also allows to not provide it at all for
> !CONFIG_FS_DAX bui
On Mon, Aug 23, 2021 at 5:40 AM Christoph Hellwig wrote:
>
> Move the dax_read_lock/dax_read_unlock pair from the callers into
> dax_supported to make it a little easier to use.
Looks good:
Reviewed-by: Dan Williams
On Mon, Aug 23, 2021 at 5:39 AM Christoph Hellwig wrote:
>
> And move the code around a bit to avoid a forward declaration.
Looks good,
Reviewed-by: Dan Williams
On Mon, Aug 23, 2021 at 12:47 PM Gerald Schaefer
wrote:
>
> On Mon, 23 Aug 2021 16:05:46 +0200
> Gerald Schaefer wrote:
>
> > On Fri, 20 Aug 2021 07:43:40 +0200
> > Christoph Hellwig wrote:
> >
> > > Hi all,
> > >
> > > looking at the recent ZONE_DEVICE related changes we still have a
> > > horr
On Mon, 23 Aug 2021 16:05:46 +0200
Gerald Schaefer wrote:
> On Fri, 20 Aug 2021 07:43:40 +0200
> Christoph Hellwig wrote:
>
> > Hi all,
> >
> > looking at the recent ZONE_DEVICE related changes we still have a
> > horrible maze of different code paths. I already suggested to
> > depend on ARC
On Mon, 2021-08-23 at 09:48 -0700, Dan Williams wrote:
> On Sun, Aug 22, 2021 at 1:36 PM QI Fuli wrote:
> >
> > Import ciniparser from ccan[1].
> > [1] https://ccodearchive.net/info/ciniparser.html
>
> Hi Qi,
>
> Vishal points out that an updated version of this code is now
> available as a pro
On Mon, Aug 23, 2021 at 5:39 AM Christoph Hellwig wrote:
>
> There is no point in trying to finding the dax device if the DAX flag is
> not set on the queue as none of the users of the device mapper exported
> block devices could make use of the DAX capability.
Looks good,
Reviewed-by: Dan Willi
On Mon, Aug 23, 2021 at 5:37 AM Christoph Hellwig wrote:
>
> Just use the %pg format specifier instead.
>
Looks good to me:
Reviewed-by: Dan Williams
> Signed-off-by: Christoph Hellwig
> ---
> drivers/dax/super.c | 20 +++-
> 1 file changed, 7 insertions(+), 13 deletions(-)
On Mon, Aug 23, 2021 at 5:37 AM Christoph Hellwig wrote:
>
> Rename the main option text to clarify it is for file system access,
> and add a bit of text that explains how to actually switch a nvdimm
> to a fsdax capable state.
>
Looks good, nice improvement. A couple suggestions below.
> Signed
On Sun, Aug 22, 2021 at 1:36 PM QI Fuli wrote:
>
> Import ciniparser from ccan[1].
> [1] https://ccodearchive.net/info/ciniparser.html
Hi Qi,
Vishal points out that an updated version of this code is now
available as a proper library that ships in distributions called
"iniparser".
On Fri, 20 Aug 2021 07:43:40 +0200
Christoph Hellwig wrote:
> Hi all,
>
> looking at the recent ZONE_DEVICE related changes we still have a
> horrible maze of different code paths. I already suggested to
> depend on ARCH_HAS_PTE_SPECIAL for ZONE_DEVICE there, which all modern
> architectures ha
All callers already have a dax_device obtained from fs_dax_get_by_bdev
at hand, so just pass that to dax_supported() insted of doing another
lookup.
Signed-off-by: Christoph Hellwig
---
drivers/dax/super.c | 42 +-
fs/ext2/super.c | 3 ++-
fs/ext4/sup
Signed-off-by: Christoph Hellwig
---
fs/xfs/xfs_super.c | 15 +++
1 file changed, 11 insertions(+), 4 deletions(-)
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index 2c9e26a44546..5a89bf601d97 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -314,6 +314,14 @@ xfs_s
dax_supported calls into ->dax_supported which checks for fsdax support.
Don't bother building it for !CONFIG_FS_DAX as it will always return
false.
Signed-off-by: Christoph Hellwig
---
drivers/dax/super.c | 36 ++--
include/linux/dax.h | 18 ++
2
Just implement generic_fsdax_supported directly out of line instead of
adding a wrapper. Given that generic_fsdax_supported is only supplied
for CONFIG_FS_DAX builds this also allows to not provide it at all for
!CONFIG_FS_DAX builds.
Signed-off-by: Christoph Hellwig
---
drivers/dax/super.c |
Move the dax_read_lock/dax_read_unlock pair from the callers into
dax_supported to make it a little easier to use.
Signed-off-by: Christoph Hellwig
---
drivers/dax/super.c | 16 +---
drivers/md/dm-table.c | 9 ++---
2 files changed, 11 insertions(+), 14 deletions(-)
diff --gi
And move the code around a bit to avoid a forward declaration.
Signed-off-by: Christoph Hellwig
---
drivers/dax/super.c | 109 ++--
include/linux/dax.h | 5 --
2 files changed, 54 insertions(+), 60 deletions(-)
diff --git a/drivers/dax/super.c b/drivers
There is no point in trying to finding the dax device if the DAX flag is
not set on the queue as none of the users of the device mapper exported
block devices could make use of the DAX capability.
Signed-off-by: Christoph Hellwig
---
drivers/md/dm.c | 2 +-
1 file changed, 1 insertion(+), 1 dele
Just use the %pg format specifier instead.
Signed-off-by: Christoph Hellwig
---
drivers/dax/super.c | 20 +++-
1 file changed, 7 insertions(+), 13 deletions(-)
diff --git a/drivers/dax/super.c b/drivers/dax/super.c
index 44736cbd446e..3e6d7e9ee34f 100644
--- a/drivers/dax/super.
Rename the main option text to clarify it is for file system access,
and add a bit of text that explains how to actually switch a nvdimm
to a fsdax capable state.
Signed-off-by: Christoph Hellwig
---
fs/Kconfig | 17 ++---
1 file changed, 14 insertions(+), 3 deletions(-)
diff --git
Hi all,
this series first clarifies how to use fsdax in the Kconfig help a bit,
and te untangles the code path that checks if fsdax is supported.
Diffstat
drivers/dax/super.c | 191 +++---
drivers/md/dm-table.c |9 --
drivers/md/dm.c |
28 matches
Mail list logo
