linux.git/drivers/staging/dst, branch v2.6.31

Staging: DST: fix build dependancy

2009-04-03T21:53:33+00:00

On Mon, Jan 19, 2009 at 10:12:24AM -0800, Randy Dunlap (randy.dunlap@oracle.com) wrote:
> 
> DST build fails when CONFIG_BLOCK=n:

DST should depend on block and block device, in the original patch its
kconfig entry was in the BLK_DEV menu, so this dependency was satisfied
automatically. Should attached patch be pushed into drivers/staging?

Reported-by: Randy Dunlap 
Acked-by: Randy Dunlap 
Signed-off-by: Evgeniy Polyakov 
Signed-off-by: Greg Kroah-Hartman

Staging: DST: Kconfig text update.

2009-04-03T21:53:33+00:00

Kconfig help text update

Signed-off-by: Evgeniy Polyakov 
Signed-off-by: Greg Kroah-Hartman

Staging: DST: Do not allow empty barriers.

2009-04-03T21:53:33+00:00

Do not allow empty barriers or generic_make_request() ->  scsi_setup_fs_cmnd()
will explode

Signed-off-by: Evgeniy Polyakov 
Signed-off-by: Greg Kroah-Hartman

Staging: DST: extend thread pool exit conditions.

2009-04-03T21:53:33+00:00

Added thread pool exit condition into the thread_pool_del_worker(). If
called in parallel another thread can steal

Signed-off-by: Evgeniy Polyakov 
Signed-off-by: Greg Kroah-Hartman

Staging: DST: optimize bio allocation.

2009-04-03T21:53:33+00:00

Use bio prepend feature as suggested by Jens Axboe.

Signed-off-by: Evgeniy Polyakov 
Signed-off-by: Greg Kroah-Hartman

Staging: dst: kconfig update.

2009-04-03T21:53:33+00:00

On Wed, Jan 14, 2009 at 02:05:33AM +0300, Evgeniy Polyakov (zbr@ioremap.net) wrote:
> --- /dev/null
> +++ b/drivers/staging/dst/Kconfig
> @@ -0,0 +1,71 @@
> +config DST
> +	tristate "Distributed storage"
> +	depends on NET && CRYPTO && SYSFS
> +	select CONNECTOR
> +	select LIBCRC32C

Above is not needed.

Signed-off-by: Evgeniy Polyakov 
Cc: Randy Dunlap 
Signed-off-by: Greg Kroah-Hartman

Staging: dst: kconfig and makefile changes.

2009-04-03T21:53:32+00:00

Signed-off-by: Evgeniy Polaykov 
Signed-off-by: Greg Kroah-Hartman

staging: dst: replace bus_id with dev_set_name

2009-04-03T21:53:32+00:00

bus_id is going away, use the dev_set_name() function instead.

Cc: Kay Sievers 
Cc: Evgeniy Polyakov 
Signed-off-by: Greg Kroah-Hartman

Staging: dst: crypto processing.

2009-04-03T21:53:32+00:00

DST may fully encrypt the data channel in case of untrusted channel and implement
strong checksum of the transferred data. It is possible to configure algorithms
and crypto keys, they should match on both sides of the network channel.
Crypto processing does not introduce noticeble performance overhead, since DST
uses configurable pool of threads to perform crypto processing.

This patch introduces crypto processing helpers and crypto engine initialization:
glueing with the crypto layer, allocation and initialization of the crypto
processing thread pool, allocation of the cached pages, which are used to temporary
encrypt data into, since it is forbidden to encrypt data in-place, since pages
are used by the higher layers.

Signed-off-by: Evgeniy Polyakov 
Signed-off-by: Greg Kroah-Hartman

Staging: dst: transactions.

2009-04-03T21:53:32+00:00

DST uses transaction model, when each store has to be explicitly acked
from the remote node to be considered as successfully written. There
may be lots of in-flight transactions. When remote host does not ack
the transaction it will be resent predefined number of times with specified
timeouts between them. All those parameters are configurable. Transactions
are marked as failed after all resends completed unsuccessfully, having
long enough resend timeout and/or large number of resends allows not to
return error to the higher (FS usually) layer in case of short network
problems or remote node outages. In case of network RAID setup this means
that storage will not degrade until transactions are marked as failed, and
thus will not force checksum recalculation and data rebuild. In case of
connection failure DST will try to reconnect to the remote node automatically.
DST sends ping commands at idle time to detect if remote node is alive.

Because of transactional model it is possible to use zero-copy sending
without worry of data corruption (which in turn could be detected by the
strong checksums though).

Transactions are handled in this patch: allocation/freeing/completion,
scanning for stall and to-be-resent transactions and overall management
of the storing tree.

Signed-off-by: Evgeniy Polyakov 
Signed-off-by: Greg Kroah-Hartman