- 09 Aug, 2013 15 commits
-
-
Tejun Heo authored
cgroup is currently in the process of transitioning to using css (cgroup_subsys_state) as the primary handle instead of cgroup in subsystem API. For hierarchy iterators, this is beneficial because * In most cases, css is the only thing subsystems care about anyway. * On the planned unified hierarchy, iterations for different subsystems will need to skip over different subtrees of the hierarchy depending on which subsystems are enabled on each cgroup. Passing around css makes it unnecessary to explicitly specify the subsystem in question as css is intersection between cgroup and subsystem * For the planned unified hierarchy, css's would need to be created and destroyed dynamically independent from cgroup hierarchy. Having cgroup core manage css iteration makes enforcing deref rules a lot easier. Most subsystem conversions are straight-forward. Noteworthy changes are * blkio: cgroup_to_blkcg() is no longer used. Removed. * freezer: cgroup_freezer() is no longer used. Removed. * devices: cgroup_to_devcgroup() is no longer used. Removed. Signed-off-by:
Tejun Heo <tj@kernel.org> Acked-by:
Li Zefan <lizefan@huawei.com> Acked-by:
Michal Hocko <mhocko@suse.cz> Acked-by:
Vivek Goyal <vgoyal@redhat.com> Acked-by:
Aristeu Rozanski <aris@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk>
-
Tejun Heo authored
There are several places where the children list is accessed directly. This patch converts those places to use cgroup_next_child(). This will help updating the hierarchy iterators to use @css instead of @cgrp. While cgroup_next_child() can be heavy in pathological cases - e.g. a lot of dead children, this shouldn't cause any noticeable behavior differences. Signed-off-by:
-
Tejun Heo authored
cgroup is transitioning to using css (cgroup_subsys_state) as the main subsys interface handle instead of cgroup and the iterators will be updated to use css too. The iterators need to walk the cgroup hierarchy and return the css's matching the origin css, which is a bit cumbersome to open code. This patch converts cgroup_next_sibling() to cgroup_next_child() so that it can handle all steps of direct child iteration. This will be used to update iterators to take @css instead of @cgrp. In addition to the new iteration init handling, cgroup_next_child() is restructured so that the different branches share the end of iteration condition check. This patch doesn't change any behavior. Signed-off-by:
-
Tejun Heo authored
cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by:
Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
-
Tejun Heo authored
cgroup subsystem API is being converted to use css (cgroup_subsys_state) as the main handle, which makes things a bit awkward for subsystem agnostic core features - the "cgroup.*" interface files and various iterations - a bit awkward as they don't have a css to use. This patch adds cgroup->dummy_css which has NULL ->ss and whose only role is pointing back to the cgroup. This will be used to support subsystem agnostic features on the coming css based API. css_parent() is updated to handle dummy_css's. Note that css will soon grow its own ->parent field and css_parent() will be made trivial. Signed-off-by:
-
Tejun Heo authored
Previously, each file read/write operation relied on the inode reference count pinning the cgroup and simply checked whether the cgroup was marked dead before proceeding to invoke the per-subsystem callback. This was rather silly as it didn't have any synchronization or css pinning around the check and the cgroup may be removed and all css refs drained between the DEAD check and actual method invocation. This patch pins the css between open() and release() so that it is guaranteed to be alive for all file operations and remove the silly DEAD checks from cgroup_file_read/write(). Signed-off-by:
-
Tejun Heo authored
cgroup is transitioning to using css (cgroup_subsys_state) instead of cgroup as the primary subsystem handle. The cgroupfs file interface will be converted to use css's which requires finding out the subsystem from cftype so that the matching css can be determined from the cgroup. This patch adds cftype->ss which points to the subsystem the file belongs to. The field is initialized while a cftype is being registered. This makes it unnecessary to explicitly specify the subsystem for other cftype handling functions. @ss argument dropped from various cftype handling functions. This patch shouldn't introduce any behavior differences. Signed-off-by:
-
Tejun Heo authored
cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup * in subsystem implementations for the following reasons. * With unified hierarchy, subsystems will be dynamically bound and unbound from cgroups and thus css's (cgroup_subsys_state) may be created and destroyed dynamically over the lifetime of a cgroup, which is different from the current state where all css's are allocated and destroyed together with the associated cgroup. This in turn means that cgroup_css() should be synchronized and may return NULL, making it more cumbersome to use. * Differing levels of per-subsystem granularity in the unified hierarchy means that the task and descendant iterators should behave differently depending on the specific subsystem the iteration is being performed for. * In majority of the cases, subsystems only care about its part in the cgroup hierarchy - ie. the hierarchy of css's. Subsystem methods often obtain the matching css pointer from the cgroup and don't bother with the cgroup pointer itself. Passing around css fits much better. This patch converts all cgroup_subsys methods to take @css instead of @cgroup. The conversions are mostly straight-forward. A few noteworthy changes are * ->css_alloc() now takes css of the parent cgroup rather than the pointer to the new cgroup as the css for the new cgroup doesn't exist yet. Knowing the parent css is enough for all the existing subsystems. * In kernel/cgroup.c::offline_css(), unnecessary open coded css dereference is replaced with local variable access. This patch shouldn't cause any behavior differences. v2: Unnecessary explicit cgrp->subsys[] deref in css_online() replaced with local variable @css as suggested by Li Zefan. Rebased on top of new for-3.12 which includes for-3.11-fixes so that ->css_free() invocation added by da0a12ca ("cgroup: fix a leak when percpu_ref_init() fails") is converted too. Suggested by Li Zefan. Signed-off-by: -
Tejun Heo authored
Currently, controllers have to explicitly follow the cgroup hierarchy to find the parent of a given css. cgroup is moving towards using cgroup_subsys_state as the main controller interface construct, so let's provide a way to climb the hierarchy using just csses. This patch implements css_parent() which, given a css, returns its parent. The function is guarnateed to valid non-NULL parent css as long as the target css is not at the top of the hierarchy. freezer, cpuset, cpu, cpuacct, hugetlb, memory, net_cls and devices are converted to use css_parent() instead of accessing cgroup->parent directly. * __parent_ca() is dropped from cpuacct and its usage is replaced with parent_ca(). The only difference between the two was NULL test on cgroup->parent which is now embedded in css_parent() making the distinction moot. Note that eventually a css->parent field will be added to css and the NULL check in css_parent() will go away. This patch shouldn't cause any behavior differences. Signed-off-by:
-
Tejun Heo authored
css (cgroup_subsys_state) is usually embedded in a subsys specific data structure. Subsystems either use container_of() directly to cast from css to such data structure or has an accessor function wrapping such cast. As cgroup as whole is moving towards using css as the main interface handle, add and update such accessors to ease dealing with css's. All accessors explicitly handle NULL input and return NULL in those cases. While this looks like an extra branch in the code, as all controllers specific data structures have css as the first field, the casting doesn't involve any offsetting and the compiler can trivially optimize out the branch. * blkio, freezer, cpuset, cpu, cpuacct and net_cls didn't have such accessor. Added. * memory, hugetlb and devices already had one but didn't explicitly handle NULL input. Updated. Signed-off-by:
-
Tejun Heo authored
Currently, given a cgroup_subsys_state, there's no way to find out which subsystem the css is for, which we'll need to convert the cgroup controller API to primarily use @css instead of @cgroup. This patch adds cgroup_subsys_state->ss which points to the subsystem the @css belongs to. While at it, remove the comment about accessing @css->cgroup to determine the hierarchy. cgroup core will provide API to traverse hierarchy of css'es and we don't want subsystems to directly walk cgroup hierarchies anymore. Signed-off-by:
-
Tejun Heo authored
cgroup controller API will be converted to primarily use struct cgroup_subsys_state instead of struct cgroup. In preparation, make hugetlb_cgroup functions pass around struct hugetlb_cgroup instead of struct cgroup. This patch shouldn't cause any behavior differences. Signed-off-by:
Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Reviewed-by:
-
Tejun Heo authored
cgroup controller API will be converted to primarily use struct cgroup_subsys_state instead of struct cgroup. In preparation, make the internal functions of netprio_cgroup pass around @css instead of @cgrp. While at it, kill struct cgroup_netprio_state which only contained struct cgroup_subsys_state without serving any purpose. All functions are converted to deal with @css directly. This patch shouldn't cause any behavior differences. Signed-off-by:
Neil Horman <nhorman@tuxdriver.com> Acked-by:
David S. Miller <davem@davemloft.net>
-
Tejun Heo authored
cpuset uses "const" qualifiers on struct cpuset in some functions; however, it doesn't work well when a value derived from returned const pointer has to be passed to an accessor. It's C after all. Drop the "const" qualifiers except for the trivially leaf ones. This patch doesn't make any functional changes. Signed-off-by:
-
Tejun Heo authored
The names of the two struct cgroup_subsys_state accessors - cgroup_subsys_state() and task_subsys_state() - are somewhat awkward. The former clashes with the type name and the latter doesn't even indicate it's somehow related to cgroup. We're about to revamp large portion of cgroup API, so, let's rename them so that they're less awkward. Most per-controller usages of the accessors are localized in accessor wrappers and given the amount of scheduled changes, this isn't gonna add any noticeable headache. Rename cgroup_subsys_state() to cgroup_css() and task_subsys_state() to task_css(). This patch is pure rename. Signed-off-by:
-
- 02 Aug, 2013 1 commit
-
-
Tejun Heo authored
for-3.12 branch is about to receive invasive updates which are dependent on da0a12ca ("cgroup: fix a leak when percpu_ref_init() fails"). Given the amount of scheduled changes, I think it'd less painful to pull in for-3.11-fixes as preparation. Pull in for-3.11-fixes into for-3.12. Signed-off-by:
-
- 01 Aug, 2013 2 commits
-
-
Li Zefan authored
It's a rw_semaphore not a mutex. Signed-off-by:
-
Li Zefan authored
It uses a single label and checks the validity of each pointer. This is err-prone, and actually we had a bug because one of the check was insufficient. Use multi lables as we do in other places. v2: - drop initializations of local variables. Signed-off-by:
-
- 31 Jul, 2013 7 commits
-
-
Li Zefan authored
This will be used as a replacement for css_lookup(). There's a difference with cgroup id and css id. cgroup id starts with 0, while css id starts with 1. v4: - also check if cggroup_mutex is held. - make it an inline function. Signed-off-by:
-
Li Zefan authored
As cgroup id has been used in netprio cgroup and will be used in memcg, it's important to make it clear how a cgroup id is allocated. For example, in netprio cgroup, the id is used as index of anarray. Signed-off-by:
Li Zefan <lizefan@huwei.com> Reviewed-by:
-
Li Zefan authored
This enables us to lookup a cgroup by its id. v4: - add a comment for idr_remove() in cgroup_offline_fn(). v3: - on success, idr_alloc() returns the id but not 0, so fix the BUG_ON() in cgroup_init(). - pass the right value to idr_alloc() so that the id for dummy cgroup is 0. Signed-off-by:
-
Li Zefan authored
Constantly use @cset for css_set variables and use @cgrp as cgroup variables. Signed-off-by:
-
Li Zefan authored
We can use struct cfent instead. v2: - remove cgroup_seqfile_release(). Signed-off-by:
-
Li Zefan authored
This should have been removed in commit d7eeac19 ("cgroup: hold cgroup_mutex before calling css_offline"). While at it, update the comments. Signed-off-by:
-
Li Zefan authored
ss->css_free() is not called when perfcpu_ref_init() fails. Signed-off-by:
-
- 29 Jul, 2013 2 commits
-
-
Zhao Hongjiang authored
Comment for cpuset_css_offline() was on top of cpuset_css_free(). Move it. Signed-off-by:
Zhao Hongjiang <zhaohongjiang@huawei.com> Acked-by:
-
Zhao Hongjiang authored
get rid of the useless forward declaration of the struct cpuset cause the below define it. Signed-off-by:
-
- 16 Jul, 2013 2 commits
-
-
Tejun Heo authored
rebind_subsystems() performs santiy checks even on subsystems which aren't specified to be added or removed and the checks aren't all that useful given that these are in a very cold path while the violations they check would trip up in much hotter paths. Let's remove these from rebind_subsystems(). Signed-off-by:
-
Tejun Heo authored
Module ref handling in cgroup is rather weird. parse_cgroupfs_options() grabs all the modules for the specified subsystems. A module ref is kept if the specified subsystem is newly bound to the hierarchy. If not, or the operation fails, the refs are dropped. This scatters module ref handling across multiple functions making it difficult to track. It also make the function nasty to use for dynamic subsystem binding which is necessary for the planned unified hierarchy. There's nothing which requires the subsystem modules to be pinned between parse_cgroupfs_options() and rebind_subsystems() in both mount and remount paths. parse_cgroupfs_options() can just parse and rebind_subsystems() can handle pinning the subsystems that it wants to bind, which is a natural part of its task - binding - anyway. Move module ref handling into rebind_subsystems() which makes the code a lot simpler - modules are gotten iff it's gonna be bound and put iff unbound or binding fails. v2: Li pointed out that if a controller module is unloaded between parsing and binding, rebind_subsystems() won't notice the missing controller as it only iterates through existing controllers. Fix it by updating rebind_subsystems() to compare @added_mask to @pinned and fail with -ENOENT if they don't match. Signed-off-by:
-
- 12 Jul, 2013 9 commits
-
-
Tejun Heo authored
task_cgroup_path_from_hierarchy() was added for the planned new users and none of the currently planned users wants to know about multiple hierarchies. This patch drops the multiple hierarchy part and makes it always return the path in the first non-dummy hierarchy. As unified hierarchy will always have id 1, this is guaranteed to return the path for the unified hierarchy if mounted; otherwise, it will return the path from the hierarchy which happens to occupy the lowest hierarchy id, which will usually be the first hierarchy mounted after boot. Signed-off-by:
-
Tejun Heo authored
rebind_subsystems() currently fails if the hierarchy has any !root cgroups; however, on the planned unified hierarchy, rebind_subsystems() will be used while populated. Move the test to cgroup_remount(), which is the only place the test is necessary anyway. As it's impossible for the other two callers of rebind_subsystems() to have populated hierarchy, this doesn't make any behavior changes. Signed-off-by:
-
Tejun Heo authored
Currently, creating and removing cgroup files in the root directory are handled separately from the actual subsystem binding and unbinding which happens in rebind_subsystems(). Also, rebind_subsystems() users aren't handling file creation errors properly. Let's integrate top_cgroup file handling into rebind_subsystems() so that it's simpler to use and everyone handles file creation errors correctly. * On a successful return, rebind_subsystems() is guaranteed to have created all files of the new subsystems and deleted the ones belonging to the removed subsystems. After a failure, no file is created or removed. * cgroup_remount() no longer needs to make explicit populate/clear calls as it's all handled by rebind_subsystems(), and it gets proper error handling automatically. * cgroup_mount() has been updated such that the root dentry and cgroup are linked before rebind_subsystems(). Also, the init_cred dancing and base file handling are moved right above rebind_subsystems() call and proper error handling for the base files is added. While at it, add a comment explaining what's going on with the cred thing. * cgroup_kill_sb() calls rebind_subsystems() to unbind all subsystems which now implies removing all subsystem files which requires the directory's i_mutex. Grab it. This means that files on the root cgroup are removed earlier - they used to be deleted from generic super_block cleanup from vfs. This doesn't lead to any functional difference and it's cleaner to do the clean up explicitly for all files. Combined with the previous changes, this makes all cgroup file creation errors handled correctly. v2: Added comment on init_cred. v3: Li spotted that cgroup_mount() wasn't freeing tmp_links after base file addition failure. Fix it by adding free_tmp_links error handling label. v4: v3 introduced build bugs which got noticed by Fengguang's awesome kbuild test robot. Fixed, and shame on me. Signed-off-by: -
Tejun Heo authored
rebind_subsystems() will be updated to handle file creations and removals with proper error handling and to do that will need to perform file operations before actually adding the subsystem to the hierarchy. To enable such usage, update cgroup_populate/clear_dir() to use for_each_subsys() instead of for_each_root_subsys() so that they operate on all subsystems specified by @subsys_mask whether that subsystem is currently bound to the hierarchy or not. Signed-off-by:
-
Tejun Heo authored
cgroup_populate_dir() didn't use to check whether the actual file creations were successful and could return success with only subset of the requested files created, which is nasty. This patch udpates cgroup_populate_dir() so that it either succeeds with all files or fails with no file. v2: The original patch also converted for_each_root_subsys() usages to for_each_subsys() without explaining why. That part has been moved to a separate patch. Signed-off-by: -
Tejun Heo authored
cgroup_populate/clear_dir() currently take @base_files and adds and removes, respectively, cgroup_base_files[] to the directory. File additions and removals are being reorganized for proper error handling and more dynamic handling for the unified hierarchy, and mixing base and subsys file handling into the same functions gets a bit confusing. This patch moves base file handling out of cgroup_populate/clear_dir() into their users - cgroup_mount(), cgroup_create() and cgroup_destroy_locked(). Note that this changes the behavior of base file removal. If @base_files is %true, cgroup_clear_dir() used to delete files regardless of cftype until there's no files left. Now, only files with matching cfts are removed. As files can only be created by the base or registered cftypes, this shouldn't result in any behavior difference. Signed-off-by:
-
Tejun Heo authored
cgroup_add_cftypes() uses cgroup_cfts_commit() to actually create the files; however, both functions ignore actual file creation errors and just assume success. This can lead to, for example, blkio hierarchy with some of the cgroups with only subset of interface files populated after cfq-iosched is loaded under heavy memory pressure, which is nasty. This patch updates cgroup_cfts_commit() and cgroup_add_cftypes() to guarantee that all files are created on success and no file is created on failure. Signed-off-by:
-
Tejun Heo authored
cgroup_addrm_files() mishandled error return value from cgroup_add_file() and returns error iff the last file fails to create. As we're in the process of cleaning up file add/rm error handling and will reliably propagate file creation failures, there's no point in keeping adding files after a failure. Replace the broken error collection logic with immediate error return. While at it, add lockdep assertions and function comment. Signed-off-by:
-
Tejun Heo authored
* Rename it to cgroup_clear_dir() and make it take the pointer to the target cgroup instead of the the dentry. This makes the function consistent with its counterpart - cgroup_populate_dir(). * Move cgroup_clear_directory() invocation from cgroup_d_remove_dir() to cgroup_remount() so that the function doesn't have to determine the cgroup pointer back from the dentry. cgroup_d_remove_dir() now only deals with vfs, which is slightly cleaner. This patch doesn't introduce any functional differences. Signed-off-by:
-
- 11 Jul, 2013 2 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-dmLinus Torvalds authored
Pull device-mapper changes from Alasdair G Kergon: "Add a device-mapper target called dm-switch to provide a multipath framework for storage arrays that dynamically reconfigure their preferred paths for different device regions. Fix a bug in the verity target that prevented its use with some specific sizes of devices. Improve some locking mechanisms in the device-mapper core and bufio. Add Mike Snitzer as a device-mapper maintainer. A few more clean-ups and fixes" * tag 'dm-3.11-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-dm: dm: add switch target dm: update maintainers dm: optimize reorder structure dm: optimize use SRCU and RCU dm bufio: submit writes outside lock dm cache: fix arm link errors with inline dm verity: use __ffs and __fls dm flakey: correct ctr alloc failure mesg dm verity: remove pointless comparison dm: use __GFP_HIGHMEM in __vmalloc dm verity: fix inability to use a few specific devices sizes dm ioctl: set noio flag to avoid __vmalloc deadlock dm mpath: fix ioctl deadlock when no paths
-
git://git.kernel.dk/linux-blockLinus Torvalds authored
Pull core block IO updates from Jens Axboe: "Here are the core IO block bits for 3.11. It contains: - A tweak to the reserved tag logic from Jan, for weirdo devices with just 3 free tags. But for those it improves things substantially for random writes. - Periodic writeback fix from Jan. Marked for stable as well. - Fix for a race condition in IO scheduler switching from Jianpeng. - The hierarchical blk-cgroup support from Tejun. This is the grunt of the series. - blk-throttle fix from Vivek. Just a note that I'm in the middle of a relocation, whole family is flying out tomorrow. Hence I will be awal the remainder of this week, but back at work again on Monday the 15th. CC'ing Tejun, since any potential "surprises" will most likely be from the blk-cgroup work. But it's been brewing for a while and sitting in my tree and linux-next for a long time, so should be solid." * 'for-3.11/core' of git://git.kernel.dk/linux-block: (36 commits) elevator: Fix a race in elevator switching block: Reserve only one queue tag for sync IO if only 3 tags are available writeback: Fix periodic writeback after fs mount blk-throttle: implement proper hierarchy support blk-throttle: implement throtl_grp->has_rules[] blk-throttle: Account for child group's start time in parent while bio climbs up blk-throttle: add throtl_qnode for dispatch fairness blk-throttle: make throtl_pending_timer_fn() ready for hierarchy blk-throttle: make tg_dispatch_one_bio() ready for hierarchy blk-throttle: make blk_throtl_bio() ready for hierarchy blk-throttle: make blk_throtl_drain() ready for hierarchy blk-throttle: dispatch from throtl_pending_timer_fn() blk-throttle: implement dispatch looping blk-throttle: separate out throtl_service_queue->pending_timer from throtl_data->dispatch_work blk-throttle: set REQ_THROTTLED from throtl_charge_bio() and gate stats update with it blk-throttle: implement sq_to_tg(), sq_to_td() and throtl_log() blk-throttle: add throtl_service_queue->parent_sq blk-throttle: generalize update_disptime optimization in blk_throtl_bio() blk-throttle: dispatch to throtl_data->service_queue.bio_lists[] blk-throttle: move bio_lists[] and friends to throtl_service_queue ...
-
