linux.git/drivers/block/rnbd, branch v5.11-rc1

block/rnbd-clt: Does not request pdu to rtrs-clt

2020-12-16T21:56:09+00:00

Previously the rnbd client requested the rtrs to allocate rnbd_iu
just after the rtrs_iu. So the rnbd client passes the size of
rnbd_iu for rtrs_clt_open() and rtrs creates an array of
rnbd_iu and rtrs_iu.

For IO handling, rnbd_iu exists after the request because we pass
the size of rnbd_iu when setting the tag-set. Therefore we do not
use the rnbd_iu allocated by rtrs for IO handling.
We only use the rnbd_iu allocated by rtrs when doing session
initialization. Almost all rnbd_iu allocated by rtrs are wasted.

By this patch the rnbd client does not request rnbd_iu allocation
to rtrs but allocate it for itself when doing session initialization.

Also remove unused rtrs_permit_to_pdu from rtrs.

Signed-off-by: Gioh Kim 
Signed-off-by: Jack Wang 
Signed-off-by: Jens Axboe

block/rnbd-clt: Dynamically allocate sglist for rnbd_iu

2020-12-16T21:56:03+00:00

The BMAX_SEGMENT static array for scatterlist is embedded in
rnbd_iu structure to avoid memory allocation in hot IO path.
In many cases, we do need only several sg entries because many IOs
have only several segments.

This patch change rnbd_iu to check the number of segments in the request
and allocate sglist dynamically.

For io path, use sg_alloc_table_chained to allocate sg list faster.

First it makes two sg entries after pdu of request.
The sg_alloc_table_chained uses the pre-allocated sg entries
if the number of segments of the request is less than two.
So it reduces the number of memory allocation.

Signed-off-by: Gioh Kim 
Signed-off-by: Jack Wang 
Signed-off-by: Jens Axboe

block/rnbd: Set write-back cache and fua same to the target device

2020-12-16T21:55:59+00:00

The rnbd-client always sets the write-back cache and fua attributes
of the rnbd device queue regardless of the target device on the server.
That generates IO hang issue when the target device does not
support both of write-back cacne and fua.

This patch adds more fields for the cache policy and fua into the
device opening message. The rnbd-server sends the information
if the target device supports the write-back cache and fua
and rnbd-client recevives it and set the device queue accordingly.

Signed-off-by: Gioh Kim 
[jwang: some minor change, rename a few varables, remove unrelated comments.]
Signed-off-by: Jack Wang 
Signed-off-by: Jens Axboe

block/rnbd: Fix typos

2020-12-16T21:55:55+00:00

Signed-off-by: Jack Wang 
Signed-off-by: Jens Axboe

block/rnbd-srv: Protect dev session sysfs removal

2020-12-16T21:55:50+00:00

Since the removal of the session sysfs can also be called from the
function destroy_sess, there is a need to protect the call from the
function rnbd_srv_sess_dev_force_close

Fixes: 786998050cbc ("block/rnbd-srv: close a mapped device from server side.")
Signed-off-by: Md Haris Iqbal 
Reviewed-by: Guoqing Jiang 
Signed-off-by: Jack Wang 
Signed-off-by: Jens Axboe

block/rnbd-clt: Fix possible memleak

2020-12-16T21:55:44+00:00

In error case, we do not free the memory for blk_symlink_name.

Do it by free the memory in error case, and set to NULL
afterwards.

Also fix the condition in rnbd_clt_remove_dev_symlink.

Fixes: 64e8a6ece1a5 ("block/rnbd-clt: Dynamically alloc buffer for pathname & blk_symlink_name")
Signed-off-by: Jack Wang 
Reviewed-by: Md Haris Iqbal 
Signed-off-by: Jens Axboe

block/rnbd-clt: Get rid of warning regarding size argument in strlcpy

2020-12-16T21:55:39+00:00

The kernel test robot triggerred the following warning,

>> drivers/block/rnbd/rnbd-clt.c:1397:42: warning: size argument in
'strlcpy' call appears to be size of the source; expected the size of the
destination [-Wstrlcpy-strlcat-size]
	strlcpy(dev->pathname, pathname, strlen(pathname) + 1);
					      ~~~~~~~^~~~~~~~~~~~~

To get rid of the above warning, use a kstrdup as Bart suggested.

Fixes: 64e8a6ece1a5 ("block/rnbd-clt: Dynamically alloc buffer for pathname & blk_symlink_name")
Reported-by: kernel test robot 
Signed-off-by: Md Haris Iqbal 
Signed-off-by: Jack Wang 
Signed-off-by: Jens Axboe

Merge tag 'for-5.11/drivers-2020-12-14' of git://git.kernel.dk/linux-block

2020-12-16T21:09:32+00:00

Pull block driver updates from Jens Axboe:
 "Nothing major in here:

   - NVMe pull request from Christoph:
        - nvmet passthrough improvements (Chaitanya Kulkarni)
        - fcloop error injection support (James Smart)
        - read-only support for zoned namespaces without Zone Append
          (Javier González)
        - improve some error message (Minwoo Im)
        - reject I/O to offline fabrics namespaces (Victor Gladkov)
        - PCI queue allocation cleanups (Niklas Schnelle)
        - remove an unused allocation in nvmet (Amit Engel)
        - a Kconfig spelling fix (Colin Ian King)
        - nvme_req_qid simplication (Baolin Wang)

   - MD pull request from Song:
        - Fix race condition in md_ioctl() (Dae R. Jeong)
        - Initialize read_slot properly for raid10 (Kevin Vigor)
        - Code cleanup (Pankaj Gupta)
        - md-cluster resync/reshape fix (Zhao Heming)

   - Move null_blk into its own directory (Damien Le Moal)

   - null_blk zone and discard improvements (Damien Le Moal)

   - bcache race fix (Dongsheng Yang)

   - Set of rnbd fixes/improvements (Gioh Kim, Guoqing Jiang, Jack Wang,
     Lutz Pogrell, Md Haris Iqbal)

   - lightnvm NULL pointer deref fix (tangzhenhao)

   - sr in_interrupt() removal (Sebastian Andrzej Siewior)

   - FC endpoint security support for s390/dasd (Jan Höppner, Sebastian
     Ott, Vineeth Vijayan). From the s390 arch guys, arch bits included
     as it made it easier for them to funnel the feature through the
     block driver tree.

   - Follow up fixes (Colin Ian King)"

* tag 'for-5.11/drivers-2020-12-14' of git://git.kernel.dk/linux-block: (64 commits)
  block: drop dead assignments in loop_init()
  sr: Remove in_interrupt() usage in sr_init_command().
  sr: Switch the sector size back to 2048 if sr_read_sector() changed it.
  cdrom: Reset sector_size back it is not 2048.
  drivers/lightnvm: fix a null-ptr-deref bug in pblk-core.c
  null_blk: Move driver into its own directory
  null_blk: Allow controlling max_hw_sectors limit
  null_blk: discard zones on reset
  null_blk: cleanup discard handling
  null_blk: Improve implicit zone close
  null_blk: improve zone locking
  block: Align max_hw_sectors to logical blocksize
  null_blk: Fail zone append to conventional zones
  null_blk: Fix zone size initialization
  bcache: fix race between setting bdev state to none and new write request direct to backing
  block/rnbd: fix a null pointer dereference on dev->blk_symlink_name
  block/rnbd-clt: Dynamically alloc buffer for pathname & blk_symlink_name
  block/rnbd: call kobject_put in the failure path
  Documentation/ABI/rnbd-srv: add document for force_close
  block/rnbd-srv: close a mapped device from server side.
  ...

block/rnbd: fix a null pointer dereference on dev->blk_symlink_name

2020-12-07T15:00:32+00:00

Currently in the case where dev->blk_symlink_name fails to be allocates
the error return path attempts to set an end-of-string character to
the unallocated dev->blk_symlink_name causing a null pointer dereference
error. Fix this by returning with an explicity ENOMEM error (which also
is missing in the original code as was not initialized).

Fixes: 1eb54f8f5dd8 ("block/rnbd: client: sysfs interface functions")
Signed-off-by: Colin Ian King 
Addresses-Coverity: ("Dereference after null check")
Signed-off-by: Jens Axboe

block/rnbd-clt: Dynamically alloc buffer for pathname & blk_symlink_name

2020-12-04T16:41:10+00:00

For every rnbd_clt_dev, we alloc the pathname and blk_symlink_name
statically to NAME_MAX which is 255 bytes. In most of the cases we only
need less than 10 bytes, so 500 bytes per block device are wasted.

This commit dynamically allocates memory buffer for pathname and
blk_symlink_name.

Signed-off-by: Md Haris Iqbal 
Signed-off-by: Jack Wang 
Reviewed-by: Lutz Pogrell 
Signed-off-by: Jens Axboe