[PATCH] cifs: document 'fsc' mount option

[PATCH] cifs: document 'fsc' mount option

[Suresh Jayaraman]

Document the newly added local caching feature using FS-Cache. This patch
could be queued and considered once the local caching patches gets merged
upstream.

Signed-off-by: Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org>
---
 doc/mount.cifs.8.xml |   17 +++++++++++++++++
 1 files changed, 17 insertions(+), 0 deletions(-)

diff --git a/doc/mount.cifs.8.xml b/doc/mount.cifs.8.xml
index 4937550..2200746 100644
--- a/doc/mount.cifs.8.xml
+++ b/doc/mount.cifs.8.xml
@@ -534,6 +534,23 @@ permissions in memory that can't be stored on the server. This information can d
                 maximum wsize currently allowed by CIFS is 57344 (fourteen
                 4096 byte pages)</para></listitem>
 	</varlistentry>
+	<varlistentry>
+		<term>fsc</term>
+
+		<listitem><para>Enable local disk caching using FS-Cache
+		for cifs. This option could be useful to improve performance
+		on a slow link, heavily loaded server and/or network
+		where reading from the disk is faster than reading from the
+		server (over the network). This could also impact the
+		scalability positively as the number of calls to the server
+		are reduced. But, be warned that local caching is not suitable
+		for all workloads, for e.g., read-once type workloads. So
+		you need to consider carefully the situation/workload before
+		using this option.
+		NOTE: You also need to have cachefilesd daemon installed and
+		running to make the cache operational.
+                </para></listitem>
+	</varlistentry>
 	 <varlistentry>
                 <term>--verbose</term>
                 <listitem><para>Print additional debugging information for the mount. Note that this parameter must be specified before the -o. For example:</para><para>mount -t cifs //server/share /mnt --verbose -o user=username</para></listitem>

Re: [PATCH] cifs: document 'fsc' mount option

[Jeff Layton]

On Tue, 27 Jul 2010 19:20:00 +0530
Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org> wrote:

> Document the newly added local caching feature using FS-Cache. This patch
> could be queued and considered once the local caching patches gets merged
> upstream.
> 
> Signed-off-by: Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org>
> ---
>  doc/mount.cifs.8.xml |   17 +++++++++++++++++
>  1 files changed, 17 insertions(+), 0 deletions(-)
> 
> diff --git a/doc/mount.cifs.8.xml b/doc/mount.cifs.8.xml
> index 4937550..2200746 100644
> --- a/doc/mount.cifs.8.xml
> +++ b/doc/mount.cifs.8.xml
> @@ -534,6 +534,23 @@ permissions in memory that can't be stored on the server. This information can d
>                  maximum wsize currently allowed by CIFS is 57344 (fourteen
>                  4096 byte pages)</para></listitem>
>  	</varlistentry>
> +	<varlistentry>
> +		<term>fsc</term>
> +
> +		<listitem><para>Enable local disk caching using FS-Cache
> +		for cifs. This option could be useful to improve performance
> +		on a slow link, heavily loaded server and/or network
> +		where reading from the disk is faster than reading from the
> +		server (over the network). This could also impact the
> +		scalability positively as the number of calls to the server
> +		are reduced. But, be warned that local caching is not suitable
> +		for all workloads, for e.g., read-once type workloads. So
> +		you need to consider carefully the situation/workload before
> +		using this option.
> +		NOTE: You also need to have cachefilesd daemon installed and
> +		running to make the cache operational.
> +                </para></listitem>
> +	</varlistentry>

You may want to mention somewhere that this is only available in recent
kernels that have been built with fsc turned on.

Other than that, this and the mount.cifs patch look fine and I'll plan
to commit them before the upcoming 4.6 release.

>  	 <varlistentry>
>                  <term>--verbose</term>
>                  <listitem><para>Print additional debugging information for the mount. Note that this parameter must be specified before the -o. For example:</para><para>mount -t cifs //server/share /mnt --verbose -o user=username</para></listitem>

-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH] cifs: document 'fsc' mount option

[Steve French]

Yes - I agree with Jeff, need to add the mention of the config option
and perhaps how to tell if fsc feature is available in kernel.

On Tue, Jul 27, 2010 at 9:33 AM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> On Tue, 27 Jul 2010 19:20:00 +0530
> Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org> wrote:
>
>> Document the newly added local caching feature using FS-Cache. This patch
>> could be queued and considered once the local caching patches gets merged
>> upstream.
>>
>> Signed-off-by: Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org>
>> ---
>>  doc/mount.cifs.8.xml |   17 +++++++++++++++++
>>  1 files changed, 17 insertions(+), 0 deletions(-)
>>
>> diff --git a/doc/mount.cifs.8.xml b/doc/mount.cifs.8.xml
>> index 4937550..2200746 100644
>> --- a/doc/mount.cifs.8.xml
>> +++ b/doc/mount.cifs.8.xml
>> @@ -534,6 +534,23 @@ permissions in memory that can't be stored on the server. This information can d
>>                  maximum wsize currently allowed by CIFS is 57344 (fourteen
>>                  4096 byte pages)</para></listitem>
>>       </varlistentry>
>> +     <varlistentry>
>> +             <term>fsc</term>
>> +
>> +             <listitem><para>Enable local disk caching using FS-Cache
>> +             for cifs. This option could be useful to improve performance
>> +             on a slow link, heavily loaded server and/or network
>> +             where reading from the disk is faster than reading from the
>> +             server (over the network). This could also impact the
>> +             scalability positively as the number of calls to the server
>> +             are reduced. But, be warned that local caching is not suitable
>> +             for all workloads, for e.g., read-once type workloads. So
>> +             you need to consider carefully the situation/workload before
>> +             using this option.
>> +             NOTE: You also need to have cachefilesd daemon installed and
>> +             running to make the cache operational.
>> +                </para></listitem>
>> +     </varlistentry>
>
> You may want to mention somewhere that this is only available in recent
> kernels that have been built with fsc turned on.
>
> Other than that, this and the mount.cifs patch look fine and I'll plan
> to commit them before the upcoming 4.6 release.
>
>>        <varlistentry>
>>                  <term>--verbose</term>
>>                  <listitem><para>Print additional debugging information for the mount. Note that this parameter must be specified before the -o. For example:</para><para>mount -t cifs //server/share /mnt --verbose -o user=username</para></listitem>
>
> --
> Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
>



-- 
Thanks,

Steve

Re: [PATCH] cifs: document 'fsc' mount option

[Jeff Layton]

On Tue, 27 Jul 2010 10:13:14 -0500
Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> Yes - I agree with Jeff, need to add the mention of the config option
> and perhaps how to tell if fsc feature is available in kernel.
> 

How would one tell? :)

> On Tue, Jul 27, 2010 at 9:33 AM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> > On Tue, 27 Jul 2010 19:20:00 +0530
> > Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org> wrote:
> >
> >> Document the newly added local caching feature using FS-Cache. This patch
> >> could be queued and considered once the local caching patches gets merged
> >> upstream.
> >>
> >> Signed-off-by: Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org>
> >> ---
> >>  doc/mount.cifs.8.xml |   17 +++++++++++++++++
> >>  1 files changed, 17 insertions(+), 0 deletions(-)
> >>
> >> diff --git a/doc/mount.cifs.8.xml b/doc/mount.cifs.8.xml
> >> index 4937550..2200746 100644
> >> --- a/doc/mount.cifs.8.xml
> >> +++ b/doc/mount.cifs.8.xml
> >> @@ -534,6 +534,23 @@ permissions in memory that can't be stored on the server. This information can d
> >>                  maximum wsize currently allowed by CIFS is 57344 (fourteen
> >>                  4096 byte pages)</para></listitem>
> >>       </varlistentry>
> >> +     <varlistentry>
> >> +             <term>fsc</term>
> >> +
> >> +             <listitem><para>Enable local disk caching using FS-Cache
> >> +             for cifs. This option could be useful to improve performance
> >> +             on a slow link, heavily loaded server and/or network
> >> +             where reading from the disk is faster than reading from the
> >> +             server (over the network). This could also impact the
> >> +             scalability positively as the number of calls to the server
> >> +             are reduced. But, be warned that local caching is not suitable
> >> +             for all workloads, for e.g., read-once type workloads. So
> >> +             you need to consider carefully the situation/workload before
> >> +             using this option.
> >> +             NOTE: You also need to have cachefilesd daemon installed and
> >> +             running to make the cache operational.
> >> +                </para></listitem>
> >> +     </varlistentry>
> >
> > You may want to mention somewhere that this is only available in recent
> > kernels that have been built with fsc turned on.
> >
> > Other than that, this and the mount.cifs patch look fine and I'll plan
> > to commit them before the upcoming 4.6 release.
> >
> >>        <varlistentry>
> >>                  <term>--verbose</term>
> >>                  <listitem><para>Print additional debugging information for the mount. Note that this parameter must be specified before the -o. For example:</para><para>mount -t cifs //server/share /mnt --verbose -o user=username</para></listitem>
> >
> > --
> > Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> >
> 
> 
> 


-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH] cifs: document 'fsc' mount option

[Steve French]

scripts/extract-ikconfig could be run against the image (or look at
/proc/config.gz) - but unless there is an easier way to view kernel
config options on particular distros via distro specific tools -
probably easier just to note that CONFIG_FSCACHE must be selected in
the kernel build (since it forces dependencies on the other kernel
options being set)

On Tue, Jul 27, 2010 at 10:15 AM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
> On Tue, 27 Jul 2010 10:13:14 -0500
> Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>
>> Yes - I agree with Jeff, need to add the mention of the config option
>> and perhaps how to tell if fsc feature is available in kernel.
>>
>
> How would one tell? :)
>
>> On Tue, Jul 27, 2010 at 9:33 AM, Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
>> > On Tue, 27 Jul 2010 19:20:00 +0530
>> > Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org> wrote:
>> >
>> >> Document the newly added local caching feature using FS-Cache. This patch
>> >> could be queued and considered once the local caching patches gets merged
>> >> upstream.
>> >>
>> >> Signed-off-by: Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org>
>> >> ---
>> >>  doc/mount.cifs.8.xml |   17 +++++++++++++++++
>> >>  1 files changed, 17 insertions(+), 0 deletions(-)
>> >>
>> >> diff --git a/doc/mount.cifs.8.xml b/doc/mount.cifs.8.xml
>> >> index 4937550..2200746 100644
>> >> --- a/doc/mount.cifs.8.xml
>> >> +++ b/doc/mount.cifs.8.xml
>> >> @@ -534,6 +534,23 @@ permissions in memory that can't be stored on the server. This information can d
>> >>                  maximum wsize currently allowed by CIFS is 57344 (fourteen
>> >>                  4096 byte pages)</para></listitem>
>> >>       </varlistentry>
>> >> +     <varlistentry>
>> >> +             <term>fsc</term>
>> >> +
>> >> +             <listitem><para>Enable local disk caching using FS-Cache
>> >> +             for cifs. This option could be useful to improve performance
>> >> +             on a slow link, heavily loaded server and/or network
>> >> +             where reading from the disk is faster than reading from the
>> >> +             server (over the network). This could also impact the
>> >> +             scalability positively as the number of calls to the server
>> >> +             are reduced. But, be warned that local caching is not suitable
>> >> +             for all workloads, for e.g., read-once type workloads. So
>> >> +             you need to consider carefully the situation/workload before
>> >> +             using this option.
>> >> +             NOTE: You also need to have cachefilesd daemon installed and
>> >> +             running to make the cache operational.
>> >> +                </para></listitem>
>> >> +     </varlistentry>
>> >
>> > You may want to mention somewhere that this is only available in recent
>> > kernels that have been built with fsc turned on.
>> >
>> > Other than that, this and the mount.cifs patch look fine and I'll plan
>> > to commit them before the upcoming 4.6 release.
>> >
>> >>        <varlistentry>
>> >>                  <term>--verbose</term>
>> >>                  <listitem><para>Print additional debugging information for the mount. Note that this parameter must be specified before the -o. For example:</para><para>mount -t cifs //server/share /mnt --verbose -o user=username</para></listitem>
>> >
>> > --
>> > Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
>> >
>>
>>
>>
>
>
> --
> Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
>



-- 
Thanks,

Steve

Re: [PATCH] cifs: document 'fsc' mount option

[Jeff Layton]

On Tue, 27 Jul 2010 10:21:51 -0500
Steve French <smfrench-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> scripts/extract-ikconfig could be run against the image (or look at
> /proc/config.gz) - but unless there is an easier way to view kernel
> config options on particular distros via distro specific tools -
> probably easier just to note that CONFIG_FSCACHE must be selected in
> the kernel build (since it forces dependencies on the other kernel
> options being set)
> 

It would be nice if there were a definitive way to tell what's been
compiled in. Not just for this, but for DFS/krb5 upcalls, etc. Maybe we
should consider a /proc/fs/cifs/features file that would display this
info? For instance:

$ cat /proc/fs/cifs/features
dfs spnego fsc

...or something like that?

-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH] cifs: document 'fsc' mount option

[Suresh Jayaraman]

On 07/27/2010 08:51 PM, Steve French wrote:
> scripts/extract-ikconfig could be run against the image (or look at
> /proc/config.gz) - but unless there is an easier way to view kernel
> config options on particular distros via distro specific tools -
> probably easier just to note that CONFIG_FSCACHE must be selected in
> the kernel build (since it forces dependencies on the other kernel
> options being set)

Does something like
	$zcat /proc/config.gz | grep CONFIG_CIFS_FSCACHE

might work on all distros?

But, Jeff's idea of having a features file in /proc sounds good. Apart
from using them as a reference to check in documentation, it could be a
useful information while collecting debugging output as well..?


Thanks,

-- 
Suresh Jayaraman

Re: [PATCH] cifs: document 'fsc' mount option

[Steve French]

On Tue, Jul 27, 2010 at 11:33 AM, Suresh Jayaraman <sjayaraman@suse.de> wrote:
> On 07/27/2010 08:51 PM, Steve French wrote:
>> scripts/extract-ikconfig could be run against the image (or look at
>> /proc/config.gz) - but unless there is an easier way to view kernel
>> config options on particular distros via distro specific tools -
>> probably easier just to note that CONFIG_FSCACHE must be selected in
>> the kernel build (since it forces dependencies on the other kernel
>> options being set)
>
> Does something like
>        $zcat /proc/config.gz | grep CONFIG_CIFS_FSCACHE
>
> might work on all distros?
>
> But, Jeff's idea of having a features file in /proc sounds good. Apart
> from using them as a reference to check in documentation, it could be a
> useful information while collecting debugging output as well..?

Seems reasonable - but wondering if other file systems have run into the same
problem - how to view their build/config options - and how they have
solved it (if at all) so ccing fsdevel



-- 
Thanks,

Steve
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] cifs: document 'fsc' mount option

[Jeff Layton]

On Tue, 27 Jul 2010 19:20:00 +0530
Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org> wrote:

> Document the newly added local caching feature using FS-Cache. This patch
> could be queued and considered once the local caching patches gets merged
> upstream.
> 
> Signed-off-by: Suresh Jayaraman <sjayaraman-l3A5Bk7waGM@public.gmane.org>
> ---
>  doc/mount.cifs.8.xml |   17 +++++++++++++++++
>  1 files changed, 17 insertions(+), 0 deletions(-)
> 
> diff --git a/doc/mount.cifs.8.xml b/doc/mount.cifs.8.xml
> index 4937550..2200746 100644
> --- a/doc/mount.cifs.8.xml
> +++ b/doc/mount.cifs.8.xml
> @@ -534,6 +534,23 @@ permissions in memory that can't be stored on the server. This information can d
>                  maximum wsize currently allowed by CIFS is 57344 (fourteen
>                  4096 byte pages)</para></listitem>
>  	</varlistentry>
> +	<varlistentry>
> +		<term>fsc</term>
> +
> +		<listitem><para>Enable local disk caching using FS-Cache
> +		for cifs. This option could be useful to improve performance
> +		on a slow link, heavily loaded server and/or network
> +		where reading from the disk is faster than reading from the
> +		server (over the network). This could also impact the
> +		scalability positively as the number of calls to the server
> +		are reduced. But, be warned that local caching is not suitable
> +		for all workloads, for e.g., read-once type workloads. So
> +		you need to consider carefully the situation/workload before
> +		using this option.
> +		NOTE: You also need to have cachefilesd daemon installed and
> +		running to make the cache operational.
> +                </para></listitem>
> +	</varlistentry>
>  	 <varlistentry>
>                  <term>--verbose</term>
>                  <listitem><para>Print additional debugging information for the mount. Note that this parameter must be specified before the -o. For example:</para><para>mount -t cifs //server/share /mnt --verbose -o user=username</para></listitem>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-cifs" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

merged
-- 
Jeff Layton <jlayton-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Re: [PATCH] cifs: document 'fsc' mount option

[Randy Dunlap]

On Tue, 27 Jul 2010 11:38:20 -0500 Steve French wrote:

> On Tue, Jul 27, 2010 at 11:33 AM, Suresh Jayaraman <sjayaraman@suse.de> wrote:
> > On 07/27/2010 08:51 PM, Steve French wrote:
> >> scripts/extract-ikconfig could be run against the image (or look at
> >> /proc/config.gz) - but unless there is an easier way to view kernel
> >> config options on particular distros via distro specific tools -
> >> probably easier just to note that CONFIG_FSCACHE must be selected in
> >> the kernel build (since it forces dependencies on the other kernel
> >> options being set)
> >
> > Does something like
> >        $zcat /proc/config.gz | grep CONFIG_CIFS_FSCACHE
> >
> > might work on all distros?
> >
> > But, Jeff's idea of having a features file in /proc sounds good. Apart
> > from using them as a reference to check in documentation, it could be a
> > useful information while collecting debugging output as well..?
> 
> Seems reasonable - but wondering if other file systems have run into the same
> problem - how to view their build/config options - and how they have
> solved it (if at all) so ccing fsdevel

There is no guarantee that scripts/extract-ikconfig can find anything,
or that proc/config.gz is present.

The safe, sure-to-work thing to do is create some strings or files or whatever
based on CONFIG_* kconfig symbols during kernel build, then those hints will
be available at kernel run-time.

---
~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***

Re: [PATCH] cifs: document 'fsc' mount option

[Suresh Jayaraman]

On 07/27/2010 10:08 PM, Steve French wrote:
> On Tue, Jul 27, 2010 at 11:33 AM, Suresh Jayaraman <sjayaraman@suse.de> wrote:
>> On 07/27/2010 08:51 PM, Steve French wrote:
>>> scripts/extract-ikconfig could be run against the image (or look at
>>> /proc/config.gz) - but unless there is an easier way to view kernel
>>> config options on particular distros via distro specific tools -
>>> probably easier just to note that CONFIG_FSCACHE must be selected in
>>> the kernel build (since it forces dependencies on the other kernel
>>> options being set)
>>
>> Does something like
>> � � � �$zcat /proc/config.gz | grep CONFIG_CIFS_FSCACHE
>>
>> might work on all distros?
>>
>> But, Jeff's idea of having a features file in /proc sounds good. Apart
>> from using them as a reference to check in documentation, it could be a
>> useful information while collecting debugging output as well..?
> 
> Seems reasonable - but wondering if other file systems have run into the same
> problem - how to view their build/config options - and how they have
> solved it (if at all) so ccing fsdevel
> 

I just checked some of the other filesystems. Most of them just mention
the CONFIG_XXX option has to be set/enabled.

For e.g.

(1) Mount options for jfs

iocharset=name
	...
	This requires CONFIG_NLS_UTF8 to be set in the kernel .config
file.

(2) Mount options for tmpfs

mpol=[default|prefer:Node|bind:NodeList|interleave|interleave:NodeList]
	Set the NUMA  memory  allocation  policy  for  all  files  in  that
instance  (if  the  kernel  CONFIG_NUMA  is enabled) ...


Thanks,

-- 
Suresh Jayaraman