On Thu, Sep 24, 2020 at 03:37:08PM +0200, mwi...@suse.com wrote:
> From: Martin Wilck
>
> Add an implementation of get_multipath_config() and put_multipath_config()
> to libmultipath. The linker's symbol lookup rules will make sure that
> applications can override these functions if they need to.
On Thu, Sep 24, 2020 at 03:36:43PM +0200, mwi...@suse.com wrote:
> From: Martin Wilck
>
> This defines the ABI of libmpathpersist in the current state.
> ---
> libmpathpersist/Makefile| 6 --
> libmpathpersist/libmpathpersist.version | 20
> 2 files chan
On Thu, Sep 24, 2020 at 03:36:33PM +0200, mwi...@suse.com wrote:
> From: Martin Wilck
>
> Hi Christophe, hi Ben,
>
> Patch 1-5 are small fixes, the first two resent from an earlier
> submission. Patch 6ff. add version scripts for the linker to
> libmultipath, libmpathpersist, and libmpathcmd.
>
For LUNs between 0 and 255 peripheral addressing is used. For LUNs higher than
255 the LUN addressing
should switch to flat according to the specification. Instead of printing out
the LUN number without regard to
the shifting of address method, display the LUN as it was intended to be the
user c
On 2020/09/25 4:14, Sudhakar Panneerselvam wrote:
>>
>> On Thu, 24 Sep 2020, Sudhakar Panneerselvam wrote:
>>
By copying it to a temporary aligned buffer and issuing I/O on this
buffer.
>>>
>>> I don't like this idea. Because, you need to allocate additional pages
>>> for the entire I/O s
On Thu, Sep 24, 2020 at 07:22:21PM +, Martin Wilck wrote:
> On Thu, 2020-09-24 at 11:30 -0500, Benjamin Marzinski wrote:
> > On Thu, Sep 24, 2020 at 08:18:02AM +, Martin Wilck wrote:
> > > On Wed, 2020-09-23 at 23:59 -0500, Benjamin Marzinski wrote:
> > > >
> > > > The second patch adds ge
On Thu, Sep 24, 2020 at 03:36:41PM +0200, mwi...@suse.com wrote:
> From: Martin Wilck
>
> The unit tests use a lot of internal symbols that we don't want
> to add to the ABI if we don't have to. As long as we don't
> have to make incompatible changes to functions, we can work around
> that by sim
On Thu, Sep 24, 2020 at 03:40:54PM +0200, mwi...@suse.com wrote:
> From: Martin Wilck
>
Git complains about the extra blank line at the end of the file
third-party/valgrind/mpath-tools.supp:33: new blank line at EOF.
-Ben
> These leaks are caused by other libraries (libsystemd, glibc,
> libgc
On Thu, Sep 24, 2020 at 03:36:39PM +0200, mwi...@suse.com wrote:
> From: Martin Wilck
>
> If compiled with -DLOAD_ALL_SHARED_LIBS, all known prioritizers
> and checkers will be loaded immediately on startup. This is useful
> for determining symbol usage (start executables with LD_BIND_NOW=1).
It
On Thu, Sep 24, 2020 at 03:36:35PM +0200, mwi...@suse.com wrote:
> From: Martin Wilck
>
> Since f0462f0 ("libmultipath: use vector for for pp->hwe and mp->hwe"),
> we've been trying to fix issues caused by paths getting freed and mpp->hwe
> dangling. This approach couldn't work because we need mp
On 9/24/20 12:51 AM, Christoph Hellwig wrote:
> Hi Jens,
>
> this series contains a bunch of different BDI cleanups. The biggest item
> is to isolate block drivers from the BDI in preparation of changing the
> lifetime of the block device BDI in a follow up series.
Applied, thanks.
--
Jens Axb
On Thu, 2020-09-24 at 11:30 -0500, Benjamin Marzinski wrote:
> On Thu, Sep 24, 2020 at 08:18:02AM +, Martin Wilck wrote:
> > On Wed, 2020-09-23 at 23:59 -0500, Benjamin Marzinski wrote:
> > >
> > > The second patch adds get_uid fallback code for dasd devices. The
> > > third
> > > patch just c
On Thu, Sep 24 2020 at 2:56pm -0400,
John Dorminy wrote:
> On Thu, Sep 24, 2020 at 1:24 PM John Dorminy wrote:
> >
> > I am impressed at how much I read wrong...
> >
> > On Thu, Sep 24, 2020 at 1:00 PM Mike Snitzer wrote:
> > >
> > > On Thu, Sep 24 2020 at 12:45pm -0400,
> > > John Dorminy wr
>
> On Thu, 24 Sep 2020, Sudhakar Panneerselvam wrote:
>
> > > By copying it to a temporary aligned buffer and issuing I/O on this
> > > buffer.
> >
> > I don't like this idea. Because, you need to allocate additional pages
> > for the entire I/O size(for the misaligned case, if you think through
On Thu, Sep 24 2020 at 2:12pm -0400,
Mikulas Patocka wrote:
>
>
> On Thu, 24 Sep 2020, Mikulas Patocka wrote:
>
> >
> >
> > On Thu, 24 Sep 2020, Mike Snitzer wrote:
> >
> > > WAIT... Could it be that raid_io_hints _really_ meant to special case
> > > raid0 and raid10 -- due to their stripi
On Thu, 2020-09-24 at 11:20 -0500, Benjamin Marzinski wrote:
> On Thu, Sep 24, 2020 at 08:06:41AM +, Martin Wilck wrote:
> > On Wed, 2020-09-23 at 23:59 -0500, Benjamin Marzinski wrote:
> > >
> > > diff --git a/libmultipath/discovery.c b/libmultipath/discovery.c
> > > index f650534f..435f2639
On Thu, Sep 24, 2020 at 1:24 PM John Dorminy wrote:
>
> I am impressed at how much I read wrong...
>
> On Thu, Sep 24, 2020 at 1:00 PM Mike Snitzer wrote:
> >
> > On Thu, Sep 24 2020 at 12:45pm -0400,
> > John Dorminy wrote:
> >
> > > I don't understand how this works...
> > >
> > > Can chunk_si
On Thu, 24 Sep 2020, Sudhakar Panneerselvam wrote:
> > By copying it to a temporary aligned buffer and issuing I/O on this
> > buffer.
>
> I don't like this idea. Because, you need to allocate additional pages
> for the entire I/O size(for the misaligned case, if you think through
You can b
On Thu, 24 Sep 2020, Mikulas Patocka wrote:
>
>
> On Thu, 24 Sep 2020, Mike Snitzer wrote:
>
> > WAIT... Could it be that raid_io_hints _really_ meant to special case
> > raid0 and raid10 -- due to their striping/splitting requirements!?
> > So, not raid1 but raid0?
> >
> > E.g.:
> >
> > d
> By copying it to a temporary aligned buffer and issuing I/O on this
> buffer.
I don't like this idea. Because, you need to allocate additional pages for the
entire I/O size(for the misaligned case, if you think through carefully, you
will know why we have to allocate a temporary buffer that is
On Thu, 24 Sep 2020, Sudhakar Panneerselvam wrote:
> Hi Mikulas,
>
> > > Windows Guest <--> Vhost-Scsi <--> LIO(scsi/target/blockio) <--> dm-crypt
> > <--> iSCSI block device
> > >
> > > One real example out of my debugging: Windows sends a I/O request with
> > > 6656 bytes to vhost-scsi inte
Hi Mikulas,
> > Windows Guest <--> Vhost-Scsi <--> LIO(scsi/target/blockio) <--> dm-crypt
> <--> iSCSI block device
> >
> > One real example out of my debugging: Windows sends a I/O request with
> > 6656 bytes to vhost-scsi interface. Vhost-scsi uses translate_desc() in
> > drivers/vhost/vhost.c
On Thu, 24 Sep 2020, Sudhakar Panneerselvam wrote:
> Hello Eric,
>
> > -Original Message-
> > From: Eric Biggers [mailto:ebigg...@kernel.org]
> > Sent: Wednesday, September 23, 2020 11:14 PM
> > To: Mike Snitzer
> > Cc: Sudhakar Panneerselvam ;
> > damien.lem...@wdc.com; ssudhak...@gm
I am impressed at how much I read wrong...
On Thu, Sep 24, 2020 at 1:00 PM Mike Snitzer wrote:
>
> On Thu, Sep 24 2020 at 12:45pm -0400,
> John Dorminy wrote:
>
> > I don't understand how this works...
> >
> > Can chunk_size_bytes be 0? If not, how is discard_granularity being set to
> > 0?
>
>
Christoph,
> aoe forces a larger readahead size, but any reason to do larger I/O is
> not limited to readahead. Also set the optimal I/O size, and remove
> the local constants in favor of just using SZ_2G.
Reviewed-by: Martin K. Petersen
--
Martin K. Petersen Oracle Linux Engineering
Hello Mikulas,
> -Original Message-
> From: Mikulas Patocka [mailto:mpato...@redhat.com]
> Sent: Thursday, September 24, 2020 6:40 AM
> To: Sudhakar Panneerselvam
> Cc: a...@redhat.com; snit...@redhat.com; dm-devel@redhat.com; dm-
> cr...@saout.de; damien.lem...@wdc.com; Shirley Ma
> ; ss
On Thu, 24 Sep 2020, Mike Snitzer wrote:
> WAIT... Could it be that raid_io_hints _really_ meant to special case
> raid0 and raid10 -- due to their striping/splitting requirements!?
> So, not raid1 but raid0?
>
> E.g.:
>
> diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
> index 56b72
On Thu, Sep 24, 2020 at 12:16:24PM -0400, Mike Snitzer wrote:
> On Thu, Sep 24 2020 at 11:45am -0400,
> Eric Biggers wrote:
>
> > On Thu, Sep 24, 2020 at 09:46:49AM -0400, Mike Snitzer wrote:
> > > > > Can you help me better understand the expected consumer of this code?
> > > > > If you have som
On Thu, Sep 24 2020 at 12:45pm -0400,
John Dorminy wrote:
> I don't understand how this works...
>
> Can chunk_size_bytes be 0? If not, how is discard_granularity being set to 0?
Yeah, I had same question.. see the reply I just sent in this thread:
https://www.redhat.com/archives/dm-devel/2020-
On Thu, 24 Sep 2020, John Dorminy wrote:
> I don't understand how this works...
>
> Can chunk_size_bytes be 0? If not, how is discard_granularity being set to 0?
Yes - chunk_size_bytes is 0 here.
> I think also limits is local to the ti in question here, initialized
> by blk_set_stacking_lim
On Thu, Sep 24 2020 at 12:26pm -0400,
Mikulas Patocka wrote:
> This patch fixes a warning WARN_ON_ONCE(!q->limits.discard_granularity).
> The reason is that the function raid_io_hints overwrote
> limits->discard_granularity with zero. We need to properly stack the
> limits instead of overwriting
Hello Milan,
> -Original Message-
> From: Milan Broz [mailto:gmazyl...@gmail.com]
> Sent: Thursday, September 24, 2020 2:16 AM
> To: Eric Biggers ; Mike Snitzer
> Cc: Sudhakar Panneerselvam ;
> damien.lem...@wdc.com; ssudhak...@gmail.com; Martin Petersen
> ; dm-cr...@saout.de; dm-devel@re
Hello Eric,
> -Original Message-
> From: Eric Biggers [mailto:ebigg...@kernel.org]
> Sent: Wednesday, September 23, 2020 11:14 PM
> To: Mike Snitzer
> Cc: Sudhakar Panneerselvam ;
> damien.lem...@wdc.com; ssudhak...@gmail.com; Martin Petersen
> ; dm-cr...@saout.de; dm-devel@redhat.com;
>
I don't understand how this works...
Can chunk_size_bytes be 0? If not, how is discard_granularity being set to 0?
I think also limits is local to the ti in question here, initialized
by blk_set_stacking_limits() via dm-table.c, and therefore has only
default values and not anything to do with th
On Thu, Sep 24, 2020 at 08:18:02AM +, Martin Wilck wrote:
> On Wed, 2020-09-23 at 23:59 -0500, Benjamin Marzinski wrote:
> > The main part of the this patchset is the first patch, which adds a
> > new library interface to check whether devices are valid paths. This
> > was designed for use in t
This patch fixes a warning WARN_ON_ONCE(!q->limits.discard_granularity).
The reason is that the function raid_io_hints overwrote
limits->discard_granularity with zero. We need to properly stack the
limits instead of overwriting them.
Signed-off-by: Mikulas Patocka
Cc: sta...@vger.kernel.org
---
On Thu, 24 Sep 2020, Mike Snitzer wrote:
> On Thu, Sep 24 2020 at 10:34am -0400,
> Mikulas Patocka wrote:
>
> > Hi Mike
> >
> > Could you send this to Linus before v5.9 is released?
> >
> > Mikulas
> >
> >
> >
> > From: Mikulas Patocka
> >
> > If q->limits.discard_granularity is zero,
On Thu, Sep 24, 2020 at 08:06:41AM +, Martin Wilck wrote:
> On Wed, 2020-09-23 at 23:59 -0500, Benjamin Marzinski wrote:
> > If uid_attribute is explicitly set to an empty string, multipath
> > should
> > log the uid at the default log level, since using the fallback code
> > is
> > the expecte
On Thu, Sep 24 2020 at 11:45am -0400,
Eric Biggers wrote:
> On Thu, Sep 24, 2020 at 09:46:49AM -0400, Mike Snitzer wrote:
> > > > Can you help me better understand the expected consumer of this code?
> > > > If you have something _real_ please be explicit. It makes justifying
> > > > supporting
Hello Mike,
> -Original Message-
> From: Mike Snitzer [mailto:snit...@redhat.com]
> Sent: Wednesday, September 23, 2020 7:28 PM
> To: Sudhakar Panneerselvam
> Cc: a...@redhat.com; dm-devel@redhat.com; dm-cr...@saout.de;
> mpato...@redhat.com; damien.lem...@wdc.com; Shirley Ma
> ; ssudhak.
Christoph,
> Drivers shouldn't really mess with the readahead size, as that is a VM
> concept. Instead set it based on the optimal I/O size by lifting the
> algorithm from the md driver when registering the disk. Also set
> bdi->io_pages there as well by applying the same scheme based on
> max
Christoph,
> The raid5 and raid10 drivers currently update the read-ahead size, but
> not the optimal I/O size on reshape. To prepare for deriving the
> read-ahead size from the optimal I/O size make sure it is updated as
> well.
Reviewed-by: Martin K. Petersen
--
Martin K. Petersen Or
Christoph,
> Inherit the optimal I/O size setting just like the readahead window,
> as any reason to do larger I/O does not apply to just readahead.
Looks fine.
Reviewed-by: Martin K. Petersen
--
Martin K. Petersen Oracle Linux Engineering
--
dm-devel mailing list
dm-devel@redhat.com
On Thu, Sep 24, 2020 at 09:46:49AM -0400, Mike Snitzer wrote:
> > > Can you help me better understand the expected consumer of this code?
> > > If you have something _real_ please be explicit. It makes justifying
> > > supporting niche code like this more tolerable.
> >
> > So the motivation for t
On Thu, Sep 24 2020 at 2:51am -0400,
Christoph Hellwig wrote:
> Drivers shouldn't really mess with the readahead size, as that is a VM
> concept. Instead set it based on the optimal I/O size by lifting the
> algorithm from the md driver when registering the disk. Also set
> bdi->io_pages there
On Thu, Sep 24 2020 at 10:34am -0400,
Mikulas Patocka wrote:
> Hi Mike
>
> Could you send this to Linus before v5.9 is released?
>
> Mikulas
>
>
>
> From: Mikulas Patocka
>
> If q->limits.discard_granularity is zero, the block core will warn in
> __blkdev_issue_discard. This patch sets it
On Thu 24-09-20 08:51:34, Christoph Hellwig wrote:
> Drivers shouldn't really mess with the readahead size, as that is a VM
> concept. Instead set it based on the optimal I/O size by lifting the
> algorithm from the md driver when registering the disk. Also set
> bdi->io_pages there as well by ap
Hi Mike
Could you send this to Linus before v5.9 is released?
Mikulas
From: Mikulas Patocka
If q->limits.discard_granularity is zero, the block core will warn in
__blkdev_issue_discard. This patch sets it to a minimum - 512 bytes.
Signed-off-by: Mikulas Patocka
Fixes: b35fd7422c2f ("block:
On Thu, Sep 24 2020 at 3:38am -0400,
Satya Tangirala wrote:
> On Wed, Sep 23, 2020 at 09:21:03PM -0400, Mike Snitzer wrote:
> > On Wed, Sep 09 2020 at 7:44pm -0400,
> > Satya Tangirala wrote:
> >
> > > From: Eric Biggers
> > >
> > > Update the device-mapper core to support exposing the inli
From: Martin Wilck
Add sd_notify_exit().
Signed-off-by: Martin Wilck
---
multipathd/main.c | 22 +-
1 file changed, 13 insertions(+), 9 deletions(-)
diff --git a/multipathd/main.c b/multipathd/main.c
index 44c1bd1..e742aa5 100644
--- a/multipathd/main.c
+++ b/multipathd/ma
On Thu, Sep 24 2020 at 3:17am -0400,
Satya Tangirala wrote:
> On Wed, Sep 23, 2020 at 09:14:39PM -0400, Mike Snitzer wrote:
> > On Mon, Sep 21 2020 at 8:32pm -0400,
> > Eric Biggers wrote:
> >
> > > On Wed, Sep 09, 2020 at 11:44:21PM +, Satya Tangirala wrote:
> > > > From: Eric Biggers
>
From: Martin Wilck
This will make it easer to move the stuff around later.
The only functional change is that map destuction now happens after
joining all threads, which should actually improve robustness.
Signed-off-by: Martin Wilck
---
multipathd/main.c | 64 +
From: Martin Wilck
This fixes several issues with the log_thread. First, the running
flag logq_running should be set by the thread itself, not by
log_thread_start()/_stop(). Second, the thread was both cancelled and
terminated via a flag (again, logq_running). It's sufficient,
and better, to just
From: Martin Wilck
This function prepares for calling dm_lib_exit() on program exit.
It undoes changes to libdm internals done by libmultipath.
It doesn't call dm_lib_exit(), as the caller may want to keep
libdm active.
Signed-off-by: Martin Wilck
---
libmultipath/config.c| 1 +
libmultip
From: Martin Wilck
Signed-off-by: Martin Wilck
---
multipathd/dmevents.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/multipathd/dmevents.c b/multipathd/dmevents.c
index b561cbf..f52f597 100644
--- a/multipathd/dmevents.c
+++ b/multipathd/dmevents.c
@@ -257,6 +257,8 @@ void unwatch_all
From: Martin Wilck
Also removing the comment about dlog() and dm_write_log().
dlog() can cope with get_multipath_config() returning NULL,
and dm_write_log() hasn't accessed the configuration for a while.
Signed-off-by: Martin Wilck
---
multipathd/main.c | 24 ++--
1 file ch
From: Martin Wilck
These leaks are caused by other libraries (libsystemd, glibc,
libgcrypt) and should be ignored when debugging with valgrind
Usage example:
valgrind --suppressions=mpath-tools.supp \
--leak-check=full --show-leak-kinds=all $COMMAND
Signed-off-by: Martin Wilck
---
third-
From: Martin Wilck
Signed-off-by: Martin Wilck
---
libmultipath/log_pthread.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/libmultipath/log_pthread.c b/libmultipath/log_pthread.c
index 15baef8..0c327ff 100644
--- a/libmultipath/log_pthread.c
+++ b/libmultipath/log_pthread.c
@@ -112,6
From: Martin Wilck
Signed-off-by: Martin Wilck
---
multipath/main.c | 39 ++-
1 file changed, 18 insertions(+), 21 deletions(-)
diff --git a/multipath/main.c b/multipath/main.c
index 9ae46ed..049a36f 100644
--- a/multipath/main.c
+++ b/multipath/main.c
@@ -4
From: Martin Wilck
So far we haven't called any cleanup code if child() failed.
Fix it.
Signed-off-by: Martin Wilck
---
multipathd/main.c | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/multipathd/main.c b/multipathd/main.c
index e742aa5..3c8f893 100644
--- a/mul
From: Martin Wilck
Signed-off-by: Martin Wilck
---
multipathd/main.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/multipathd/main.c b/multipathd/main.c
index 3c8f893..483b3c7 100644
--- a/multipathd/main.c
+++ b/multipathd/main.c
@@ -3074,8 +3074,10 @@ child (__attribu
From: Martin Wilck
Signed-off-by: Martin Wilck
---
libmpathpersist/mpath_persist.c | 2 --
libmultipath/config.c | 4
libmultipath/libmultipath.version | 5 +
multipathd/main.c | 3 ---
4 files changed, 5 insertions(+), 9 deletions(-)
diff --git a/libmpat
From: Martin Wilck
It seems we've been doing this only in the failure case, for ages.
Signed-off-by: Martin Wilck
---
multipathd/main.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/multipathd/main.c b/multipathd/main.c
index 3db051b..44c1bd1 100644
--- a/multipathd/ma
From: Martin Wilck
Hi Christophe, hi Ben, hi lixiaokeng,
this series was inspired by lixiaokeng's recent posting "[QUESTION] memory
leak in main (multipath)". It implements my first idea, registering
cleanup handlers with atexit(). However it turned out to be quite
complex. In particular multipa
From: Martin Wilck
Signed-off-by: Martin Wilck
---
multipathd/main.c | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/multipathd/main.c b/multipathd/main.c
index 2642570..1cdbff1 100644
--- a/multipathd/main.c
+++ b/multipathd/main.c
@@ -116,19 +116,19 @@ struc
From: Martin Wilck
cleanup_child() calls all cleanups in the right order, in an
exit handler.
Signed-off-by: Martin Wilck
---
multipathd/main.c | 48 +++
1 file changed, 28 insertions(+), 20 deletions(-)
diff --git a/multipathd/main.c b/multipathd/m
From: Martin Wilck
Signed-off-by: Martin Wilck
---
mpathpersist/main.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/mpathpersist/main.c b/mpathpersist/main.c
index 3c2e657..14245cc 100644
--- a/mpathpersist/main.c
+++ b/mpathpersist/main.c
@@ -641,11 +641,10 @@ int m
From: Martin Wilck
With libmp_dm_exit() in place, we can make sure that the
calls are made in the right order.
Signed-off-by: Martin Wilck
---
multipathd/main.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/multipathd/main.c b/multipathd/main.c
index 483b3c7..f9749e5
From: Martin Wilck
The uxlsnr wouldn't always release the client lock when cancelled,
causing a deadlock in uxsock_cleanup(). While this hasn't been
caused by commit 3d611a2, the deadlock seems to have become much
more likely after that patch. Solving this means that we have to
treat reallocation
From: Martin Wilck
Fix this leak in multipathd, reported by valgrind, that messes up
multipathd's otherwise clean leak report:
==23823== 336 bytes in 1 blocks are possibly lost in loss record 3 of 3
==23823==at 0x483AB65: calloc (in
/usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==2
From: Martin Wilck
Also, introduce booleans that indicate a certain thread has
been started successfully. Using these booleans, we can avoid
crashing by cancelling threads that have never been started.
Signed-off-by: Martin Wilck
---
multipathd/main.c | 76 +++--
From: Martin Wilck
This thread attribute can be dynamically initialized and destroyed.
No need to carry it along in multipathd. Removal of the symbol
requires a new ABI version.
Signed-off-by: Martin Wilck
---
libmultipath/io_err_stat.c| 7 +--
libmultipath/libmultipath.version |
From: Martin Wilck
There were two leaks in check_path_valid(): if path status was
successfully determined before calling store_pathvec(), free_path()
wasn't called. Also, if an error exit occured, neither cleanup
function was called.
This patch fixes both, at the cost of using "static" for the p
From: Martin Wilck
Signed-off-by: Martin Wilck
---
multipathd/main.c | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/multipathd/main.c b/multipathd/main.c
index 4d5b40b..3db051b 100644
--- a/multipathd/main.c
+++ b/multipathd/main.c
@@ -2892,6 +2892,12 @@ set_oom_
On Wed, Sep 23, 2020 at 04:06:51PM -0400, Mike Snitzer wrote:
> Add QUEUE_FLAG_NOWAIT to allow a block device to advertise support for
> REQ_NOWAIT. Bio-based devices may set QUEUE_FLAG_NOWAIT where
> applicable.
>
> Update QUEUE_FLAG_MQ_DEFAULT to include QUEUE_FLAG_NOWAIT. Also
> update submit_
On Thu, Sep 24 2020 at 3:48am -0400,
Satya Tangirala wrote:
> On Wed, Sep 23, 2020 at 09:21:03PM -0400, Mike Snitzer wrote:
> > On Wed, Sep 09 2020 at 7:44pm -0400,
> > Satya Tangirala wrote:
> >
> > > From: Eric Biggers
> > >
> > > Update the device-mapper core to support exposing the inli
From: Martin Wilck
Reviewed-by: Benjamin Marzinski
Signed-off-by: Martin Wilck
---
multipath/main.c | 20
1 file changed, 4 insertions(+), 16 deletions(-)
diff --git a/multipath/main.c b/multipath/main.c
index dc4974b..4bbfce9 100644
--- a/multipath/main.c
+++ b/multipath
From: Martin Wilck
The logic is as follows: child() sets DAEMON_IDLE status after
DAEMON_CONFIGURE when reconfigure() has finished. The only other state change
that can race with that is DAEMON_SHUTDOWN. Other state changes will wait for
DAEMON_IDLE first (see set_config_state()). When DAEMON_CON
From: Martin Wilck
We can use the symbols from libmultipath now.
Signed-off-by: Martin Wilck
---
multipathd/main.c | 9 +++--
1 file changed, 3 insertions(+), 6 deletions(-)
diff --git a/multipathd/main.c b/multipathd/main.c
index 00b66ba..c5c374b 100644
--- a/multipathd/main.c
+++ b/mult
From: Martin Wilck
If multipathd gets a shutdown request during initial reconfigure(),
it shouldn't send "READY=1" to systemd. Ensure this by sending
"READY=1" via post_config_state().
Reviewed-by: Benjamin Marzinski
Signed-off-by: Martin Wilck
---
multipathd/main.c | 16 +++-
1 f
From: Martin Wilck
Hi Christophe, hi Ben,
this series contains a number of patches I've wanted to do for some
time. It's based on my "upstream-queue" tree
(https://github.com/openSUSE/multipath-tools/tree/upstream-queue),
plus my previously submitted "multipath-tools:add linker version scripts"
From: Martin Wilck
The purpose of dm_lib_release() is to release stacked device node
operations in libdevmapper. This is functionality we don't need and
use any more, as we rely on udev to set up device nodes and symlinks.
We always set DM_UDEV_DISABLE_LIBRARY_FALLBACK when we run dm tasks.
In t
From: Martin Wilck
Provide an alternative init function libmpathpersist_init() which
avoids allocating a new struct config, simply using libmultipath's
internal implementation.
Reviewed-by: Benjamin Marzinski
Signed-off-by: Martin Wilck
---
libmpathpersist/libmpathpersist.version | 6
l
From: Martin Wilck
Have libmpathpersist_{init,exit} do the udev initialization, too.
Signed-off-by: Martin Wilck
---
libmpathpersist/mpath_persist.c | 11 ---
libmpathpersist/mpath_persist.h | 9 ++---
2 files changed, 14 insertions(+), 6 deletions(-)
diff --git a/libmpathpersist
From: Martin Wilck
Reviewed-by: Benjamin Marzinski
Signed-off-by: Martin Wilck
---
mpathpersist/main.c | 16 ++--
1 file changed, 2 insertions(+), 14 deletions(-)
diff --git a/mpathpersist/main.c b/mpathpersist/main.c
index a6a3bcf..278e48f 100644
--- a/mpathpersist/main.c
+++ b/m
From: Martin Wilck
As one step towards bundling all possibly racy libdm init calls into a single
function, split the code for determining and checking versions of
libdm and kernel components. Provide a generic helper
libmp_get_version() that makes sure the versions are "lazily" initialized.
Note
From: Martin Wilck
It's unlikely but not impossible that other threads change the state
while we're waiting, and if we grab the lock again, it's still not
what we wanted. We need to continue waiting until either the condition
is met, or time timeout expired.
Moreover, generalize this code so tha
From: Martin Wilck
Add an implementation of get_multipath_config() and put_multipath_config()
to libmultipath. The linker's symbol lookup rules will make sure that
applications can override these functions if they need to. Defining
these functions in libmultipath avoids the need to provide stubs
From: Martin Wilck
Use __post_config_state() and __wait_for_state_change(). This
way __post_config_state() is the only place where running_state
is ever changed, and we avoid code duplication.
Reviewed-by: Benjamin Marzinski
Signed-off-by: Martin Wilck
---
multipathd/main.c | 23 +
From: Martin Wilck
With these symbols added, applications using libmultipath don't
need to define global variables "udev" and "logsink" any more.
This comes at the cost of having to call an init function.
Currently, libmultipath_init() does nothing but initialize
"udev".
The linker's symbol look
From: Martin Wilck
We can use libmultipath's symbols now.
Reviewed-by: Benjamin Marzinski
Signed-off-by: Martin Wilck
---
multipath/main.c | 7 ++-
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/multipath/main.c b/multipath/main.c
index 4bbfce9..9ae46ed 100644
--- a/multipa
From: Martin Wilck
dm_udev_wait() and dm_task_run() may access global / static state
in libdevmapper. They need to be protected by a lock in in our
multithreaded library.
Cc: lixiaok...@huawei.com
Reviewed-by: Benjamin Marzinski
Signed-off-by: Martin Wilck
---
libmultipath/devmapper.c
From: Martin Wilck
reconfigure() can be a long-running operation; both initial path
discovery and initial map setup can take a long time. Allow
the main program to indicate that the process should be
interrupted if a shutdown signal was received.
We take advantage of the dynamic linker's symbol
From: Martin Wilck
Cancel the other threads before taking vecs->lock. This avoids
delays during shutdown caused e.g. by the checker thread holding
the vecs lock.
Note: this makes it possible that cancelled threads leak memory,
because they can now be cancelled before having released the vecs
loc
From: Martin Wilck
We can use libmultipath's internal symbols now. The libmultipath
initialization is taken care of by libmpathpersist_init().
Signed-off-by: Martin Wilck
---
mpathpersist/main.c | 6 --
1 file changed, 6 deletions(-)
diff --git a/mpathpersist/main.c b/mpathpersist/main.c
From: Martin Wilck
While we access running_state only under the config_lock,
we sometimes do in a loop. Use the volatile qualifier to make
sure compilers can't optimize away the loads.
Reviewed-by: Benjamin Marzinski
Signed-off-by: Martin Wilck
---
multipathd/main.c | 2 +-
1 file changed, 1
From: Martin Wilck
Inform systemd that the daemon is shutting down. See sd_notify(3).
Reviewed-by: Benjamin Marzinski
Signed-off-by: Martin Wilck
---
multipathd/main.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/multipathd/main.c b/multipathd/main.c
index 53a22a4..c
From: Martin Wilck
Reviewed-by: Benjamin Marzinski
Signed-off-by: Martin Wilck
---
libmultipath/config.c | 3 +--
libmultipath/config.h | 2 +-
libmultipath/parser.c | 9 +
libmultipath/parser.h | 2 +-
4 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/libmultipath/config.
From: Martin Wilck
The unit tests use a lot of internal symbols that we don't want
to add to the ABI if we don't have to. As long as we don't
have to make incompatible changes to functions, we can work around
that by simply using a non-versioned library for the unit tests.
Therefore we add a sepe
From: Martin Wilck
Hi Christophe, hi Ben,
Patch 1-5 are small fixes, the first two resent from an earlier
submission. Patch 6ff. add version scripts for the linker to
libmultipath, libmpathpersist, and libmpathcmd.
Is it useful to do this for libmultipath? We have always said that this is
not a
1 - 100 of 139 matches
Mail list logo
