glusterfs.git/xlators/performance/io-cache, branch v4.1.4

xlators/performance: Add pass-through option

2018-04-11T10:53:48+00:00

Add pass-through option in performance traslators. Set the option in
GF_OPTION_INIT() and GF_OPTION_RECONF()

Updates: #304

Change-Id: If1537450147d154905831e36f7162a32866d7ad6
Signed-off-by: Varsha Rao

io-cache: Fix coverity issue NEGATIVE_RETURNS

2018-03-05T01:37:27+00:00

Change-Id: I811225ad20e3bd9f05820212e6a843f05d96b246
Signed-off-by: Poornima G

io-cache: Fix coverity issue

2018-02-27T11:38:30+00:00

Coverity issue : FORWARD_NULL
fd is assigned within a condition, but the fd is used even outside
the condition.

Change-Id: I6548d605d8a8acc6a25f1657f9fb75586d513042
Signed-off-by: Poornima G

perf/io-cache: Update options for gd2

2018-01-12T15:34:32+00:00

Update io-cache options for GD2

Updates gluster/glusterfs#302

Change-Id: I7fabf912d1e3cd024989404526db0a29644768c7
Signed-off-by: N Balachandran

all: Simplify component message id's definition

2017-12-14T02:33:51+00:00

This patch creates a new way of defining message id's that is easier
and less error prone because it doesn't require so many manual changes
each time a new component is defined or a new message created.

Change-Id: I71ba8af9ac068f5add7e74f316a2478bc991c67b
Signed-off-by: Xavier Hernandez

xlators/performance/io-cache/: DEADCODE in ioc_open_cbk

2017-11-03T17:17:17+00:00

Coverity ID: 173

Issue: Event dead_error_line: Execution cannot reach this statement:
"fd_ctx_set(fd, this, 1UL);".

Removed if block as "weight == 0U" cannot be true, so if block will
never execute. Also removed weight variable because it is unused.

Change-Id: I4f028df29bfde91167fb15befa99b1fe1892adc6
BUG: 789278
Signed-off-by: Girjesh Rajoria

mem-pool: track glusterfs_ctx_t in struct mem_pool

2017-08-29T12:37:40+00:00

In order to generate statedumps per glusterfs_ctx_t, it is needed to
place all the memory pools in a structure that the context can reach.
The 'struct mem_pool' has been extended with a 'list_head owner' that is
linked with the glusterfs_ctx_t->mempool_list.

All callers of mem_pool_new() have been updated to pass the current
glusterfs_ctx_t along. This context is needed to add the new memory pool
to the list and for grabbing the ctx->lock while updating the
glusterfs_ctx_t->mempool_list.

Updates: #307
Change-Id: Ia9384424d8d1630ef3efc9d5d523bf739c356c6e
Signed-off-by: Niels de Vos 
Reviewed-on: https://review.gluster.org/18075
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Jeff Darcy

performance/io-cache: update inode contexts of each entry in readdirplus

2017-07-27T03:58:10+00:00

io-cache stores read-cache in inode which is currently created only in
lookup. But, with readdirplus and md-cache absorbing lookups, io-cache
need not receive a lookup before a fop like readv.

Change-Id: I6eba995b0a90d4d5055a4aef0489707b852da1b8
BUG: 1474180
Signed-off-by: Raghavendra G 
Signed-off-by: Raghavendra G 
Reviewed-on: https://review.gluster.org/5029
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System

perf/ioc: Fix race causing crash when accessing freed page

2017-05-30T15:51:46+00:00

ioc_inode_wakeup does not lock the ioc_inode for the duration
of the operation, leaving a window where ioc_prune could find
a NULL waitq and hence free the page which ioc_inode_wakeup later
tries to access.

Thanks to Mohit for the analysis.

credit: moagrawa@redhat.com

Change-Id: I54b064857e2694826d0c03b23f8014e3984a3330
BUG: 1456385
Signed-off-by: N Balachandran 
Reviewed-on: https://review.gluster.org/17410
Reviewed-by: Raghavendra G 
Tested-by: Raghavendra G 
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Jeff Darcy

libglusterfs: stop special casing "cache-size" in size_t validation

2017-05-08T13:37:54+00:00

The original situation was as follows:

The function that validates xlator options indicating a size,
xlator_option_validate_sizet(), handles the case when the name
of the option is "cache-size" in a special way.

- Xlator options (things of type volume_option_t) has a
  min and max attribute of type double.
- An xlator option is endowed with a gluster specific type (not
  C type). An instance of an xlator option goes through a validation
  process by a type specific validator function (which are collected
  in option.c).
- Validators of numeric types - size being one of them - make use the
  min and max attributes to perform a range check, except in one case:
  if an option is defined with min = max = 0, then this option will be
  exempt of range checking. (Note: the volume_option_t definition
  features the following comments along the min, max fields:

   double                  min;  /* 0 means no range */
   double                  max;  /* 0 means no range */

  which is slightly misleading as it lets one to conclude that
  zeroing min or max buys exemption from low or high boundary check,
  which is not true -- only *both* being zero buys exemption.)
- Besides this, the validator for options of size type,
  xlator_option_validate_sizet() special cases options
  named "cache-size" so that only min is enforced. (The only consequence
  of a value exceeding max is that glusterd logs a warning about it, but
  the cli user who makes such a setting gets no feedback on it.)
- This was introduced because a hard coded limit is not useful for
  io-cache and quick-read. They rather use a runtime calculated
  upper limit. (See changes
  I7dd4d8c53051b89a293696abf1ee8dc237e39a20
  I9c744b5ace10604d5a814e6218ca0d83c796db80
  about the last two points.)
- As an unintended consequence, the upper limit check of
  cache-size of write-behind, for which a conventional hard coded limit
  is specified, is defeated.

What we do about it:

- Remove the special casing clause for cache-size in
  xlator_option_validate_sizet. Thus the general range
  check policy (as described above) will apply to
  cache-size too.
- To implement a lower bound only check by the validator
  for cache-size of io-cache and quick-read, change the
  max attribute of these options to INFINITY.

The only behavioral difference is the omission of the warnings
about cache-size of io-cache and quick-read exceeding the former max
values. (They were rather heuristic anyway.)

BUG: 1445609
Change-Id: I0bd8bd391fa7d926f76e214a2178833fe4673b4a
Signed-off-by: Csaba Henk 
Reviewed-on: https://review.gluster.org/17125
Smoke: Gluster Build System 
Reviewed-by: Amar Tumballi 
Tested-by: Raghavendra G 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System