[PATCH 00/11] rbd: another set of patches

[PATCH 00/11] rbd: another set of patches

[Alex Elder]

I keep moving simple and simplifying patches to the front of
my patch queue.  This is my latest set of things that are
ready to get reviewed and committed.

					-Alex

[PATCH 01/11] rbd: handle locking inside __rbd_client_find()
    This simplifies a little code.
[PATCH 02/11] rbd: don't over-allocate space for object prefix
    Don't allocate the max required space for the object prefix.
    This also ensures that the object prefix is either terminated
    with '\0' or else is maximally-sized.
[PATCH 03/11] rbd: kill incore snap_names_len
    Eliminate an unnecessary field in struct rbd_image_header.
[PATCH 04/11] rbd: more cleanup in rbd_header_from_disk()
    Rearrange some code, for better logical grouping.
[PATCH 05/11] rbd: move rbd_opts to struct rbd_device
    Tie rbd-specific options to the rbd_device, not the client.
[PATCH 06/11] rbd: add read_only rbd map option
[PATCH 07/11] rbd: kill notify_timeout option
    These two were done together, in this order, to preserve the
    argument parsing code.  The first adds the ability to map an
    rbd image read-only.  The second eliminates the only other
    rbd option, which is otherwise unused.
[PATCH 08/11] rbd: bio_chain_clone() cleanups
    Simple refactoring.
[PATCH 09/11] rbd: drop needless test in rbd_rq_fn()
    Trivial.
[PATCH 10/11] rbd: check for overflow in rbd_get_num_segments()
    Make sure we don't exceed 64-bit maximum when adding offset
    and length.
[PATCH 11/11] rbd: split up rbd_get_segment()
    Refactor rbd_get_segment() into three separate functions,
    which return a segment name, offset, and length.

Re: [PATCH 00/11] rbd: another set of patches

[Alex Elder]

On 08/24/2012 11:26 AM, Alex Elder wrote:
> I keep moving simple and simplifying patches to the front of
> my patch queue.  This is my latest set of things that are
> ready to get reviewed and committed.

Anyone have time to review these?  They're mostly very simple.
Thanks.

					-Alex

> [PATCH 01/11] rbd: handle locking inside __rbd_client_find()
>     This simplifies a little code.
> [PATCH 02/11] rbd: don't over-allocate space for object prefix
>     Don't allocate the max required space for the object prefix.
>     This also ensures that the object prefix is either terminated
>     with '\0' or else is maximally-sized.
> [PATCH 03/11] rbd: kill incore snap_names_len
>     Eliminate an unnecessary field in struct rbd_image_header.
> [PATCH 04/11] rbd: more cleanup in rbd_header_from_disk()
>     Rearrange some code, for better logical grouping.
> [PATCH 05/11] rbd: move rbd_opts to struct rbd_device
>     Tie rbd-specific options to the rbd_device, not the client.
> [PATCH 06/11] rbd: add read_only rbd map option
> [PATCH 07/11] rbd: kill notify_timeout option
>     These two were done together, in this order, to preserve the
>     argument parsing code.  The first adds the ability to map an
>     rbd image read-only.  The second eliminates the only other
>     rbd option, which is otherwise unused.
> [PATCH 08/11] rbd: bio_chain_clone() cleanups
>     Simple refactoring.
> [PATCH 09/11] rbd: drop needless test in rbd_rq_fn()
>     Trivial.
> [PATCH 10/11] rbd: check for overflow in rbd_get_num_segments()
>     Make sure we don't exceed 64-bit maximum when adding offset
>     and length.
> [PATCH 11/11] rbd: split up rbd_get_segment()
>     Refactor rbd_get_segment() into three separate functions,
>     which return a segment name, offset, and length.
> 

Re: [PATCH 00/11] rbd: another set of patches

[Alex Elder]

On 08/24/2012 11:26 AM, Alex Elder wrote:
> I keep moving simple and simplifying patches to the front of
> my patch queue.  This is my latest set of things that are
> ready to get reviewed and committed.
> 
> 					-Alex

I'm about to send updates to two patches in this series, based
on comments from Yehuda's review.  I'm not going to re-send the
entire series.

Updated are patches 03 and 06.  I also am still awaiting
confirmation of review for patch 05 (which has not changed
and I am not re-sending).

					-Alex

> 
> [PATCH 01/11] rbd: handle locking inside __rbd_client_find()
>     This simplifies a little code.
> [PATCH 02/11] rbd: don't over-allocate space for object prefix
>     Don't allocate the max required space for the object prefix.
>     This also ensures that the object prefix is either terminated
>     with '\0' or else is maximally-sized.
> [PATCH 03/11] rbd: kill incore snap_names_len
>     Eliminate an unnecessary field in struct rbd_image_header.
> [PATCH 04/11] rbd: more cleanup in rbd_header_from_disk()
>     Rearrange some code, for better logical grouping.
> [PATCH 05/11] rbd: move rbd_opts to struct rbd_device
>     Tie rbd-specific options to the rbd_device, not the client.
> [PATCH 06/11] rbd: add read_only rbd map option
> [PATCH 07/11] rbd: kill notify_timeout option
>     These two were done together, in this order, to preserve the
>     argument parsing code.  The first adds the ability to map an
>     rbd image read-only.  The second eliminates the only other
>     rbd option, which is otherwise unused.
> [PATCH 08/11] rbd: bio_chain_clone() cleanups
>     Simple refactoring.
> [PATCH 09/11] rbd: drop needless test in rbd_rq_fn()
>     Trivial.
> [PATCH 10/11] rbd: check for overflow in rbd_get_num_segments()
>     Make sure we don't exceed 64-bit maximum when adding offset
>     and length.
> [PATCH 11/11] rbd: split up rbd_get_segment()
>     Refactor rbd_get_segment() into three separate functions,
>     which return a segment name, offset, and length.
> 

[PATCH 00/11] rbd: another set of patches

[Alex Elder]

The first set of patches has already been posted for review.

It is based on the current state of the ceph-client/testing
branch, and is available as branch "wip-rbd-review-1" on
the ceph-client git repository.

    https://github.com/ceph/ceph-client/tree/wip-rbd-review-1

Most of these have been reviewed, and I will not be re-posting
them here today.

					-Alex

[PATCH 01/11]  rbd: handle locking inside __rbd_client_find()
[PATCH 02/11]  rbd: don't over-allocate space for object prefix
[PATCH 03/11]  rbd: kill incore snap_names_len
[PATCH 04/11]  rbd: more cleanup in rbd_header_from_disk()
[PATCH 05/11]  rbd: move rbd_opts to struct rbd_device
[PATCH 06/11]  rbd: add read_only rbd map option
[PATCH 07/11]  rbd: kill notify_timeout option
[PATCH 08/11]  rbd: bio_chain_clone() cleanups
[PATCH 09/11]  rbd: drop needless test in rbd_rq_fn()
[PATCH 10/11]  rbd: check for overflow in rbd_get_num_segments()
[PATCH 11/11]  rbd: split up rbd_get_segment()