linux/fs/notify/fanotify/fanotify.c

882 lines
25 KiB
C
Raw Normal View History

License cleanup: add SPDX GPL-2.0 license identifier to files with no license Many source files in the tree are missing licensing information, which makes it harder for compliance tools to determine the correct license. By default all files without license information are under the default license of the kernel, which is GPL version 2. Update the files which contain no license information with the 'GPL-2.0' SPDX license identifier. The SPDX identifier is a legally binding shorthand, which can be used instead of the full boiler plate text. This patch is based on work done by Thomas Gleixner and Kate Stewart and Philippe Ombredanne. How this work was done: Patches were generated and checked against linux-4.14-rc6 for a subset of the use cases: - file had no licensing information it it. - file was a */uapi/* one with no licensing information in it, - file was a */uapi/* one with existing licensing information, Further patches will be generated in subsequent months to fix up cases where non-standard license headers were used, and references to license had to be inferred by heuristics based on keywords. The analysis to determine which SPDX License Identifier to be applied to a file was done in a spreadsheet of side by side results from of the output of two independent scanners (ScanCode & Windriver) producing SPDX tag:value files created by Philippe Ombredanne. Philippe prepared the base worksheet, and did an initial spot review of a few 1000 files. The 4.13 kernel was the starting point of the analysis with 60,537 files assessed. Kate Stewart did a file by file comparison of the scanner results in the spreadsheet to determine which SPDX license identifier(s) to be applied to the file. She confirmed any determination that was not immediately clear with lawyers working with the Linux Foundation. Criteria used to select files for SPDX license identifier tagging was: - Files considered eligible had to be source code files. - Make and config files were included as candidates if they contained >5 lines of source - File already had some variant of a license header in it (even if <5 lines). All documentation files were explicitly excluded. The following heuristics were used to determine which SPDX license identifiers to apply. - when both scanners couldn't find any license traces, file was considered to have no license information in it, and the top level COPYING file license applied. For non */uapi/* files that summary was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 11139 and resulted in the first patch in this series. If that file was a */uapi/* path one, it was "GPL-2.0 WITH Linux-syscall-note" otherwise it was "GPL-2.0". Results of that was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 WITH Linux-syscall-note 930 and resulted in the second patch in this series. - if a file had some form of licensing information in it, and was one of the */uapi/* ones, it was denoted with the Linux-syscall-note if any GPL family license was found in the file or had no licensing in it (per prior point). Results summary: SPDX license identifier # files ---------------------------------------------------|------ GPL-2.0 WITH Linux-syscall-note 270 GPL-2.0+ WITH Linux-syscall-note 169 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-2-Clause) 21 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) 17 LGPL-2.1+ WITH Linux-syscall-note 15 GPL-1.0+ WITH Linux-syscall-note 14 ((GPL-2.0+ WITH Linux-syscall-note) OR BSD-3-Clause) 5 LGPL-2.0+ WITH Linux-syscall-note 4 LGPL-2.1 WITH Linux-syscall-note 3 ((GPL-2.0 WITH Linux-syscall-note) OR MIT) 3 ((GPL-2.0 WITH Linux-syscall-note) AND MIT) 1 and that resulted in the third patch in this series. - when the two scanners agreed on the detected license(s), that became the concluded license(s). - when there was disagreement between the two scanners (one detected a license but the other didn't, or they both detected different licenses) a manual inspection of the file occurred. - In most cases a manual inspection of the information in the file resulted in a clear resolution of the license that should apply (and which scanner probably needed to revisit its heuristics). - When it was not immediately clear, the license identifier was confirmed with lawyers working with the Linux Foundation. - If there was any question as to the appropriate license identifier, the file was flagged for further research and to be revisited later in time. In total, over 70 hours of logged manual review was done on the spreadsheet to determine the SPDX license identifiers to apply to the source files by Kate, Philippe, Thomas and, in some cases, confirmation by lawyers working with the Linux Foundation. Kate also obtained a third independent scan of the 4.13 code base from FOSSology, and compared selected files where the other two scanners disagreed against that SPDX file, to see if there was new insights. The Windriver scanner is based on an older version of FOSSology in part, so they are related. Thomas did random spot checks in about 500 files from the spreadsheets for the uapi headers and agreed with SPDX license identifier in the files he inspected. For the non-uapi files Thomas did random spot checks in about 15000 files. In initial set of patches against 4.14-rc6, 3 files were found to have copy/paste license identifier errors, and have been fixed to reflect the correct identifier. Additionally Philippe spent 10 hours this week doing a detailed manual inspection and review of the 12,461 patched files from the initial patch version early this week with: - a full scancode scan run, collecting the matched texts, detected license ids and scores - reviewing anything where there was a license detected (about 500+ files) to ensure that the applied SPDX license was correct - reviewing anything where there was no detection but the patch license was not GPL-2.0 WITH Linux-syscall-note to ensure that the applied SPDX license was correct This produced a worksheet with 20 files needing minor correction. This worksheet was then exported into 3 different .csv files for the different types of files to be modified. These .csv files were then reviewed by Greg. Thomas wrote a script to parse the csv files and add the proper SPDX tag to the file, in the format that the file expected. This script was further refined by Greg based on the output to detect more types of files automatically and to distinguish between header and source .c files (which need different comment types.) Finally Greg ran the script using the .csv files to generate the patches. Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org> Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2017-11-01 22:07:57 +08:00
// SPDX-License-Identifier: GPL-2.0
#include <linux/fanotify.h>
#include <linux/fdtable.h>
#include <linux/fsnotify_backend.h>
#include <linux/init.h>
#include <linux/jiffies.h>
#include <linux/kernel.h> /* UINT_MAX */
#include <linux/mount.h>
#include <linux/sched.h>
#include <linux/sched/user.h>
#include <linux/sched/signal.h>
#include <linux/types.h>
#include <linux/wait.h>
audit: Record fanotify access control decisions The fanotify interface allows user space daemons to make access control decisions. Under common criteria requirements, we need to optionally record decisions based on policy. This patch adds a bit mask, FAN_AUDIT, that a user space daemon can 'or' into the response decision which will tell the kernel that it made a decision and record it. It would be used something like this in user space code: response.response = FAN_DENY | FAN_AUDIT; write(fd, &response, sizeof(struct fanotify_response)); When the syscall ends, the audit system will record the decision as a AUDIT_FANOTIFY auxiliary record to denote that the reason this event occurred is the result of an access control decision from fanotify rather than DAC or MAC policy. A sample event looks like this: type=PATH msg=audit(1504310584.332:290): item=0 name="./evil-ls" inode=1319561 dev=fc:03 mode=0100755 ouid=1000 ogid=1000 rdev=00:00 obj=unconfined_u:object_r:user_home_t:s0 nametype=NORMAL type=CWD msg=audit(1504310584.332:290): cwd="/home/sgrubb" type=SYSCALL msg=audit(1504310584.332:290): arch=c000003e syscall=2 success=no exit=-1 a0=32cb3fca90 a1=0 a2=43 a3=8 items=1 ppid=901 pid=959 auid=1000 uid=1000 gid=1000 euid=1000 suid=1000 fsuid=1000 egid=1000 sgid=1000 fsgid=1000 tty=pts1 ses=3 comm="bash" exe="/usr/bin/bash" subj=unconfined_u:unconfined_r:unconfined_t: s0-s0:c0.c1023 key=(null) type=FANOTIFY msg=audit(1504310584.332:290): resp=2 Prior to using the audit flag, the developer needs to call fanotify_init or'ing in FAN_ENABLE_AUDIT to ensure that the kernel supports auditing. The calling process must also have the CAP_AUDIT_WRITE capability. Signed-off-by: sgrubb <sgrubb@redhat.com> Reviewed-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2017-10-03 08:21:39 +08:00
#include <linux/audit.h>
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-18 06:46:39 +08:00
#include <linux/sched/mm.h>
#include <linux/statfs.h>
#include <linux/stringhash.h>
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
#include "fanotify.h"
static bool fanotify_path_equal(struct path *p1, struct path *p2)
{
return p1->mnt == p2->mnt && p1->dentry == p2->dentry;
}
static unsigned int fanotify_hash_path(const struct path *path)
{
return hash_ptr(path->dentry, FANOTIFY_EVENT_HASH_BITS) ^
hash_ptr(path->mnt, FANOTIFY_EVENT_HASH_BITS);
}
static inline bool fanotify_fsid_equal(__kernel_fsid_t *fsid1,
__kernel_fsid_t *fsid2)
{
return fsid1->val[0] == fsid2->val[0] && fsid1->val[1] == fsid2->val[1];
}
static unsigned int fanotify_hash_fsid(__kernel_fsid_t *fsid)
{
return hash_32(fsid->val[0], FANOTIFY_EVENT_HASH_BITS) ^
hash_32(fsid->val[1], FANOTIFY_EVENT_HASH_BITS);
}
static bool fanotify_fh_equal(struct fanotify_fh *fh1,
struct fanotify_fh *fh2)
{
if (fh1->type != fh2->type || fh1->len != fh2->len)
return false;
return !fh1->len ||
!memcmp(fanotify_fh_buf(fh1), fanotify_fh_buf(fh2), fh1->len);
}
static unsigned int fanotify_hash_fh(struct fanotify_fh *fh)
{
long salt = (long)fh->type | (long)fh->len << 8;
/*
* full_name_hash() works long by long, so it handles fh buf optimally.
*/
return full_name_hash((void *)salt, fanotify_fh_buf(fh), fh->len);
}
static bool fanotify_fid_event_equal(struct fanotify_fid_event *ffe1,
struct fanotify_fid_event *ffe2)
{
/* Do not merge fid events without object fh */
if (!ffe1->object_fh.len)
return false;
return fanotify_fsid_equal(&ffe1->fsid, &ffe2->fsid) &&
fanotify_fh_equal(&ffe1->object_fh, &ffe2->object_fh);
}
static bool fanotify_info_equal(struct fanotify_info *info1,
struct fanotify_info *info2)
{
if (info1->dir_fh_totlen != info2->dir_fh_totlen ||
info1->file_fh_totlen != info2->file_fh_totlen ||
info1->name_len != info2->name_len)
return false;
if (info1->dir_fh_totlen &&
!fanotify_fh_equal(fanotify_info_dir_fh(info1),
fanotify_info_dir_fh(info2)))
return false;
if (info1->file_fh_totlen &&
!fanotify_fh_equal(fanotify_info_file_fh(info1),
fanotify_info_file_fh(info2)))
return false;
return !info1->name_len ||
!memcmp(fanotify_info_name(info1), fanotify_info_name(info2),
info1->name_len);
}
static bool fanotify_name_event_equal(struct fanotify_name_event *fne1,
struct fanotify_name_event *fne2)
{
struct fanotify_info *info1 = &fne1->info;
struct fanotify_info *info2 = &fne2->info;
/* Do not merge name events without dir fh */
if (!info1->dir_fh_totlen)
return false;
if (!fanotify_fsid_equal(&fne1->fsid, &fne2->fsid))
return false;
return fanotify_info_equal(info1, info2);
}
static bool fanotify_should_merge(struct fanotify_event *old,
struct fanotify_event *new)
{
pr_debug("%s: old=%p new=%p\n", __func__, old, new);
if (old->hash != new->hash ||
old->type != new->type || old->pid != new->pid)
return false;
/*
* We want to merge many dirent events in the same dir (i.e.
* creates/unlinks/renames), but we do not want to merge dirent
* events referring to subdirs with dirent events referring to
* non subdirs, otherwise, user won't be able to tell from a
* mask FAN_CREATE|FAN_DELETE|FAN_ONDIR if it describes mkdir+
* unlink pair or rmdir+create pair of events.
*/
if ((old->mask & FS_ISDIR) != (new->mask & FS_ISDIR))
return false;
switch (old->type) {
case FANOTIFY_EVENT_TYPE_PATH:
return fanotify_path_equal(fanotify_event_path(old),
fanotify_event_path(new));
case FANOTIFY_EVENT_TYPE_FID:
return fanotify_fid_event_equal(FANOTIFY_FE(old),
FANOTIFY_FE(new));
case FANOTIFY_EVENT_TYPE_FID_NAME:
return fanotify_name_event_equal(FANOTIFY_NE(old),
FANOTIFY_NE(new));
default:
WARN_ON_ONCE(1);
}
return false;
}
/* Limit event merges to limit CPU overhead per event */
#define FANOTIFY_MAX_MERGE_EVENTS 128
/* and the list better be locked by something too! */
static int fanotify_merge(struct fsnotify_group *group,
struct fsnotify_event *event)
{
struct fanotify_event *old, *new = FANOTIFY_E(event);
unsigned int bucket = fanotify_event_hash_bucket(group, new);
struct hlist_head *hlist = &group->fanotify_data.merge_hash[bucket];
int i = 0;
pr_debug("%s: group=%p event=%p bucket=%u\n", __func__,
group, event, bucket);
/*
* Don't merge a permission event with any other event so that we know
* the event structure we have created in fanotify_handle_event() is the
* one we should check for permission response.
*/
if (fanotify_is_perm_event(new->mask))
return 0;
hlist_for_each_entry(old, hlist, merge_list) {
if (++i > FANOTIFY_MAX_MERGE_EVENTS)
break;
if (fanotify_should_merge(old, new)) {
old->mask |= new->mask;
return 1;
}
}
return 0;
}
/*
* Wait for response to permission event. The function also takes care of
* freeing the permission event (or offloads that in case the wait is canceled
* by a signal). The function returns 0 in case access got allowed by userspace,
* -EPERM in case userspace disallowed the access, and -ERESTARTSYS in case
* the wait got interrupted by a signal.
*/
static int fanotify_get_response(struct fsnotify_group *group,
struct fanotify_perm_event *event,
struct fsnotify_iter_info *iter_info)
{
int ret;
pr_debug("%s: group=%p event=%p\n", __func__, group, event);
ret = wait_event_killable(group->fanotify_data.access_waitq,
event->state == FAN_EVENT_ANSWERED);
/* Signal pending? */
if (ret < 0) {
spin_lock(&group->notification_lock);
/* Event reported to userspace and no answer yet? */
if (event->state == FAN_EVENT_REPORTED) {
/* Event will get freed once userspace answers to it */
event->state = FAN_EVENT_CANCELED;
spin_unlock(&group->notification_lock);
return ret;
}
/* Event not yet reported? Just remove it. */
if (event->state == FAN_EVENT_INIT) {
fsnotify_remove_queued_event(group, &event->fae.fse);
/* Permission events are not supposed to be hashed */
WARN_ON_ONCE(!hlist_unhashed(&event->fae.merge_list));
}
/*
* Event may be also answered in case signal delivery raced
* with wakeup. In that case we have nothing to do besides
* freeing the event and reporting error.
*/
spin_unlock(&group->notification_lock);
goto out;
}
/* userspace responded, convert to something usable */
audit: Record fanotify access control decisions The fanotify interface allows user space daemons to make access control decisions. Under common criteria requirements, we need to optionally record decisions based on policy. This patch adds a bit mask, FAN_AUDIT, that a user space daemon can 'or' into the response decision which will tell the kernel that it made a decision and record it. It would be used something like this in user space code: response.response = FAN_DENY | FAN_AUDIT; write(fd, &response, sizeof(struct fanotify_response)); When the syscall ends, the audit system will record the decision as a AUDIT_FANOTIFY auxiliary record to denote that the reason this event occurred is the result of an access control decision from fanotify rather than DAC or MAC policy. A sample event looks like this: type=PATH msg=audit(1504310584.332:290): item=0 name="./evil-ls" inode=1319561 dev=fc:03 mode=0100755 ouid=1000 ogid=1000 rdev=00:00 obj=unconfined_u:object_r:user_home_t:s0 nametype=NORMAL type=CWD msg=audit(1504310584.332:290): cwd="/home/sgrubb" type=SYSCALL msg=audit(1504310584.332:290): arch=c000003e syscall=2 success=no exit=-1 a0=32cb3fca90 a1=0 a2=43 a3=8 items=1 ppid=901 pid=959 auid=1000 uid=1000 gid=1000 euid=1000 suid=1000 fsuid=1000 egid=1000 sgid=1000 fsgid=1000 tty=pts1 ses=3 comm="bash" exe="/usr/bin/bash" subj=unconfined_u:unconfined_r:unconfined_t: s0-s0:c0.c1023 key=(null) type=FANOTIFY msg=audit(1504310584.332:290): resp=2 Prior to using the audit flag, the developer needs to call fanotify_init or'ing in FAN_ENABLE_AUDIT to ensure that the kernel supports auditing. The calling process must also have the CAP_AUDIT_WRITE capability. Signed-off-by: sgrubb <sgrubb@redhat.com> Reviewed-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2017-10-03 08:21:39 +08:00
switch (event->response & ~FAN_AUDIT) {
case FAN_ALLOW:
ret = 0;
break;
case FAN_DENY:
default:
ret = -EPERM;
}
audit: Record fanotify access control decisions The fanotify interface allows user space daemons to make access control decisions. Under common criteria requirements, we need to optionally record decisions based on policy. This patch adds a bit mask, FAN_AUDIT, that a user space daemon can 'or' into the response decision which will tell the kernel that it made a decision and record it. It would be used something like this in user space code: response.response = FAN_DENY | FAN_AUDIT; write(fd, &response, sizeof(struct fanotify_response)); When the syscall ends, the audit system will record the decision as a AUDIT_FANOTIFY auxiliary record to denote that the reason this event occurred is the result of an access control decision from fanotify rather than DAC or MAC policy. A sample event looks like this: type=PATH msg=audit(1504310584.332:290): item=0 name="./evil-ls" inode=1319561 dev=fc:03 mode=0100755 ouid=1000 ogid=1000 rdev=00:00 obj=unconfined_u:object_r:user_home_t:s0 nametype=NORMAL type=CWD msg=audit(1504310584.332:290): cwd="/home/sgrubb" type=SYSCALL msg=audit(1504310584.332:290): arch=c000003e syscall=2 success=no exit=-1 a0=32cb3fca90 a1=0 a2=43 a3=8 items=1 ppid=901 pid=959 auid=1000 uid=1000 gid=1000 euid=1000 suid=1000 fsuid=1000 egid=1000 sgid=1000 fsgid=1000 tty=pts1 ses=3 comm="bash" exe="/usr/bin/bash" subj=unconfined_u:unconfined_r:unconfined_t: s0-s0:c0.c1023 key=(null) type=FANOTIFY msg=audit(1504310584.332:290): resp=2 Prior to using the audit flag, the developer needs to call fanotify_init or'ing in FAN_ENABLE_AUDIT to ensure that the kernel supports auditing. The calling process must also have the CAP_AUDIT_WRITE capability. Signed-off-by: sgrubb <sgrubb@redhat.com> Reviewed-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2017-10-03 08:21:39 +08:00
/* Check if the response should be audited */
if (event->response & FAN_AUDIT)
audit_fanotify(event->response & ~FAN_AUDIT);
pr_debug("%s: group=%p event=%p about to return ret=%d\n", __func__,
group, event, ret);
out:
fsnotify_destroy_event(group, &event->fae.fse);
return ret;
}
fanotify: return only user requested event types in event mask Modify fanotify_should_send_event() so that it now returns a mask for an event that contains ONLY flags for the event types that have been specifically requested by the user. Flags that may have been included within the event mask, but have not been explicitly requested by the user will not be present in the returned value. As an example, given the situation where a user requests events of type FAN_OPEN. Traditionally, the event mask returned within an event that occurred on a filesystem object that has been marked for monitoring and is opened, will only ever have the FAN_OPEN bit set. With the introduction of the new flags like FAN_OPEN_EXEC, and perhaps any other future event flags, there is a possibility of the returned event mask containing more than a single bit set, despite having only requested the single event type. Prior to these modifications performed to fanotify_should_send_event(), a user would have received a bundled event mask containing flags FAN_OPEN and FAN_OPEN_EXEC in the instance that a file was opened for execution via execve(), for example. This means that a user would receive event types in the returned event mask that have not been requested. This runs the possibility of breaking existing systems and causing other unforeseen issues. To mitigate this possibility, fanotify_should_send_event() has been modified to return the event mask containing ONLY event types explicitly requested by the user. This means that we will NOT report events that the user did no set a mask for, and we will NOT report events that the user has set an ignore mask for. The function name fanotify_should_send_event() has also been updated so that it's more relevant to what it has been designed to do. Signed-off-by: Matthew Bobrowski <mbobrowski@mbobrowski.org> Reviewed-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2018-11-08 11:05:49 +08:00
/*
* This function returns a mask for an event that only contains the flags
* that have been specifically requested by the user. Flags that may have
* been included within the event mask, but have not been explicitly
* requested by the user, will not be present in the returned mask.
*/
static u32 fanotify_group_event_mask(struct fsnotify_group *group,
struct fsnotify_iter_info *iter_info,
u32 event_mask, const void *data,
int data_type, struct inode *dir)
{
__u32 marks_mask = 0, marks_ignored_mask = 0;
__u32 test_mask, user_mask = FANOTIFY_OUTGOING_EVENTS |
FANOTIFY_EVENT_FLAGS;
const struct path *path = fsnotify_data_path(data, data_type);
unsigned int fid_mode = FAN_GROUP_FLAG(group, FANOTIFY_FID_BITS);
struct fsnotify_mark *mark;
int type;
pr_debug("%s: report_mask=%x mask=%x data=%p data_type=%d\n",
__func__, iter_info->report_mask, event_mask, data, data_type);
if (!fid_mode) {
/* Do we have path to open a file descriptor? */
if (!path)
return 0;
/* Path type events are only relevant for files and dirs */
if (!d_is_reg(path->dentry) && !d_can_lookup(path->dentry))
return 0;
} else if (!(fid_mode & FAN_REPORT_FID)) {
/* Do we have a directory inode to report? */
if (!dir && !(event_mask & FS_ISDIR))
return 0;
}
fsnotify_foreach_obj_type(type) {
if (!fsnotify_iter_should_report_type(iter_info, type))
continue;
mark = iter_info->marks[type];
/* Apply ignore mask regardless of ISDIR and ON_CHILD flags */
marks_ignored_mask |= mark->ignored_mask;
/*
* If the event is on dir and this mark doesn't care about
* events on dir, don't send it!
*/
if (event_mask & FS_ISDIR && !(mark->mask & FS_ISDIR))
continue;
/*
fsnotify: fix events reported to watching parent and child fsnotify_parent() used to send two separate events to backends when a parent inode is watching children and the child inode is also watching. In an attempt to avoid duplicate events in fanotify, we unified the two backend callbacks to a single callback and handled the reporting of the two separate events for the relevant backends (inotify and dnotify). However the handling is buggy and can result in inotify and dnotify listeners receiving events of the type they never asked for or spurious events. The problem is the unified event callback with two inode marks (parent and child) is called when any of the parent and child inodes are watched and interested in the event, but the parent inode's mark that is interested in the event on the child is not necessarily the one we are currently reporting to (it could belong to a different group). So before reporting the parent or child event flavor to backend we need to check that the mark is really interested in that event flavor. The semantics of INODE and CHILD marks were hard to follow and made the logic more complicated than it should have been. Replace it with INODE and PARENT marks semantics to hopefully make the logic more clear. Thanks to Hugh Dickins for spotting a bug in the earlier version of this patch. Fixes: 497b0c5a7c06 ("fsnotify: send event to parent and child with single callback") CC: stable@vger.kernel.org Link: https://lore.kernel.org/r/20201202120713.702387-4-amir73il@gmail.com Reported-by: Hugh Dickins <hughd@google.com> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2020-12-02 20:07:09 +08:00
* If the event is on a child and this mark is on a parent not
* watching children, don't send it!
*/
fsnotify: fix events reported to watching parent and child fsnotify_parent() used to send two separate events to backends when a parent inode is watching children and the child inode is also watching. In an attempt to avoid duplicate events in fanotify, we unified the two backend callbacks to a single callback and handled the reporting of the two separate events for the relevant backends (inotify and dnotify). However the handling is buggy and can result in inotify and dnotify listeners receiving events of the type they never asked for or spurious events. The problem is the unified event callback with two inode marks (parent and child) is called when any of the parent and child inodes are watched and interested in the event, but the parent inode's mark that is interested in the event on the child is not necessarily the one we are currently reporting to (it could belong to a different group). So before reporting the parent or child event flavor to backend we need to check that the mark is really interested in that event flavor. The semantics of INODE and CHILD marks were hard to follow and made the logic more complicated than it should have been. Replace it with INODE and PARENT marks semantics to hopefully make the logic more clear. Thanks to Hugh Dickins for spotting a bug in the earlier version of this patch. Fixes: 497b0c5a7c06 ("fsnotify: send event to parent and child with single callback") CC: stable@vger.kernel.org Link: https://lore.kernel.org/r/20201202120713.702387-4-amir73il@gmail.com Reported-by: Hugh Dickins <hughd@google.com> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2020-12-02 20:07:09 +08:00
if (type == FSNOTIFY_OBJ_TYPE_PARENT &&
!(mark->mask & FS_EVENT_ON_CHILD))
continue;
marks_mask |= mark->mask;
}
test_mask = event_mask & marks_mask & ~marks_ignored_mask;
/*
fanotify: send FAN_DIR_MODIFY event flavor with dir inode and name Dirent events are going to be supported in two flavors: 1. Directory fid info + mask that includes the specific event types (e.g. FAN_CREATE) and an optional FAN_ONDIR flag. 2. Directory fid info + name + mask that includes only FAN_DIR_MODIFY. To request the second event flavor, user needs to set the event type FAN_DIR_MODIFY in the mark mask. The first flavor is supported since kernel v5.1 for groups initialized with flag FAN_REPORT_FID. It is intended to be used for watching directories in "batch mode" - the watcher is notified when directory is changed and re-scans the directory content in response. This event flavor is stored more compactly in the event queue, so it is optimal for workloads with frequent directory changes. The second event flavor is intended to be used for watching large directories, where the cost of re-scan of the directory on every change is considered too high. The watcher getting the event with the directory fid and entry name is expected to call fstatat(2) to query the content of the entry after the change. Legacy inotify events are reported with name and event mask (e.g. "foo", FAN_CREATE | FAN_ONDIR). That can lead users to the conclusion that there is *currently* an entry "foo" that is a sub-directory, when in fact "foo" may be negative or non-dir by the time user gets the event. To make it clear that the current state of the named entry is unknown, when reporting an event with name info, fanotify obfuscates the specific event types (e.g. create,delete,rename) and uses a common event type - FAN_DIR_MODIFY to describe the change. This should make it harder for users to make wrong assumptions and write buggy filesystem monitors. At this point, name info reporting is not yet implemented, so trying to set FAN_DIR_MODIFY in mark mask will return -EINVAL. Link: https://lore.kernel.org/r/20200319151022.31456-12-amir73il@gmail.com Signed-off-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2020-03-19 23:10:19 +08:00
* For dirent modification events (create/delete/move) that do not carry
* the child entry name information, we report FAN_ONDIR for mkdir/rmdir
* so user can differentiate them from creat/unlink.
*
* For backward compatibility and consistency, do not report FAN_ONDIR
* to user in legacy fanotify mode (reporting fd) and report FAN_ONDIR
* to user in fid mode for all event types.
*
* We never report FAN_EVENT_ON_CHILD to user, but we do pass it in to
* fanotify_alloc_event() when group is reporting fid as indication
* that event happened on child.
*/
if (fid_mode) {
/* Do not report event flags without any event */
if (!(test_mask & ~FANOTIFY_EVENT_FLAGS))
return 0;
} else {
user_mask &= ~FANOTIFY_EVENT_FLAGS;
}
return test_mask & user_mask;
}
/*
* Check size needed to encode fanotify_fh.
*
* Return size of encoded fh without fanotify_fh header.
* Return 0 on failure to encode.
*/
static int fanotify_encode_fh_len(struct inode *inode)
{
int dwords = 0;
if (!inode)
return 0;
exportfs_encode_inode_fh(inode, NULL, &dwords, NULL);
return dwords << 2;
}
/*
* Encode fanotify_fh.
*
* Return total size of encoded fh including fanotify_fh header.
* Return 0 on failure to encode.
*/
static int fanotify_encode_fh(struct fanotify_fh *fh, struct inode *inode,
unsigned int fh_len, unsigned int *hash,
gfp_t gfp)
{
int dwords, type = 0;
char *ext_buf = NULL;
void *buf = fh->buf;
int err;
fh->type = FILEID_ROOT;
fh->len = 0;
fh->flags = 0;
if (!inode)
return 0;
/*
* !gpf means preallocated variable size fh, but fh_len could
* be zero in that case if encoding fh len failed.
*/
err = -ENOENT;
if (fh_len < 4 || WARN_ON_ONCE(fh_len % 4))
goto out_err;
/* No external buffer in a variable size allocated fh */
if (gfp && fh_len > FANOTIFY_INLINE_FH_LEN) {
/* Treat failure to allocate fh as failure to encode fh */
err = -ENOMEM;
ext_buf = kmalloc(fh_len, gfp);
if (!ext_buf)
goto out_err;
*fanotify_fh_ext_buf_ptr(fh) = ext_buf;
buf = ext_buf;
fh->flags |= FANOTIFY_FH_FLAG_EXT_BUF;
}
dwords = fh_len >> 2;
type = exportfs_encode_inode_fh(inode, buf, &dwords, NULL);
err = -EINVAL;
if (!type || type == FILEID_INVALID || fh_len != dwords << 2)
goto out_err;
fh->type = type;
fh->len = fh_len;
/* Mix fh into event merge key */
*hash ^= fanotify_hash_fh(fh);
return FANOTIFY_FH_HDR_LEN + fh_len;
out_err:
pr_warn_ratelimited("fanotify: failed to encode fid (type=%d, len=%d, err=%i)\n",
type, fh_len, err);
kfree(ext_buf);
*fanotify_fh_ext_buf_ptr(fh) = NULL;
/* Report the event without a file identifier on encode error */
fh->type = FILEID_INVALID;
fh->len = 0;
return 0;
}
/*
* The inode to use as identifier when reporting fid depends on the event.
* Report the modified directory inode on dirent modification events.
* Report the "victim" inode otherwise.
* For example:
* FS_ATTRIB reports the child inode even if reported on a watched parent.
* FS_CREATE reports the modified dir inode and not the created inode.
*/
static struct inode *fanotify_fid_inode(u32 event_mask, const void *data,
int data_type, struct inode *dir)
{
if (event_mask & ALL_FSNOTIFY_DIRENT_EVENTS)
return dir;
return fsnotify_data_inode(data, data_type);
}
/*
* The inode to use as identifier when reporting dir fid depends on the event.
* Report the modified directory inode on dirent modification events.
* Report the "victim" inode if "victim" is a directory.
* Report the parent inode if "victim" is not a directory and event is
* reported to parent.
* Otherwise, do not report dir fid.
*/
static struct inode *fanotify_dfid_inode(u32 event_mask, const void *data,
int data_type, struct inode *dir)
{
struct inode *inode = fsnotify_data_inode(data, data_type);
if (event_mask & ALL_FSNOTIFY_DIRENT_EVENTS)
return dir;
if (S_ISDIR(inode->i_mode))
return inode;
return dir;
}
static struct fanotify_event *fanotify_alloc_path_event(const struct path *path,
unsigned int *hash,
gfp_t gfp)
{
struct fanotify_path_event *pevent;
pevent = kmem_cache_alloc(fanotify_path_event_cachep, gfp);
if (!pevent)
return NULL;
pevent->fae.type = FANOTIFY_EVENT_TYPE_PATH;
pevent->path = *path;
*hash ^= fanotify_hash_path(path);
path_get(path);
return &pevent->fae;
}
static struct fanotify_event *fanotify_alloc_perm_event(const struct path *path,
gfp_t gfp)
{
struct fanotify_perm_event *pevent;
pevent = kmem_cache_alloc(fanotify_perm_event_cachep, gfp);
if (!pevent)
return NULL;
pevent->fae.type = FANOTIFY_EVENT_TYPE_PATH_PERM;
pevent->response = 0;
pevent->state = FAN_EVENT_INIT;
pevent->path = *path;
path_get(path);
return &pevent->fae;
}
static struct fanotify_event *fanotify_alloc_fid_event(struct inode *id,
__kernel_fsid_t *fsid,
unsigned int *hash,
gfp_t gfp)
{
struct fanotify_fid_event *ffe;
ffe = kmem_cache_alloc(fanotify_fid_event_cachep, gfp);
if (!ffe)
return NULL;
ffe->fae.type = FANOTIFY_EVENT_TYPE_FID;
ffe->fsid = *fsid;
*hash ^= fanotify_hash_fsid(fsid);
fanotify_encode_fh(&ffe->object_fh, id, fanotify_encode_fh_len(id),
hash, gfp);
return &ffe->fae;
}
static struct fanotify_event *fanotify_alloc_name_event(struct inode *id,
__kernel_fsid_t *fsid,
const struct qstr *name,
struct inode *child,
unsigned int *hash,
gfp_t gfp)
{
struct fanotify_name_event *fne;
struct fanotify_info *info;
struct fanotify_fh *dfh, *ffh;
unsigned int dir_fh_len = fanotify_encode_fh_len(id);
unsigned int child_fh_len = fanotify_encode_fh_len(child);
unsigned int size;
size = sizeof(*fne) + FANOTIFY_FH_HDR_LEN + dir_fh_len;
if (child_fh_len)
size += FANOTIFY_FH_HDR_LEN + child_fh_len;
if (name)
size += name->len + 1;
fne = kmalloc(size, gfp);
if (!fne)
return NULL;
fne->fae.type = FANOTIFY_EVENT_TYPE_FID_NAME;
fne->fsid = *fsid;
*hash ^= fanotify_hash_fsid(fsid);
info = &fne->info;
fanotify_info_init(info);
dfh = fanotify_info_dir_fh(info);
info->dir_fh_totlen = fanotify_encode_fh(dfh, id, dir_fh_len, hash, 0);
if (child_fh_len) {
ffh = fanotify_info_file_fh(info);
info->file_fh_totlen = fanotify_encode_fh(ffh, child,
child_fh_len, hash, 0);
}
if (name) {
long salt = name->len;
fanotify_info_copy_name(info, name);
*hash ^= full_name_hash((void *)salt, name->name, name->len);
}
pr_debug("%s: ino=%lu size=%u dir_fh_len=%u child_fh_len=%u name_len=%u name='%.*s'\n",
__func__, id->i_ino, size, dir_fh_len, child_fh_len,
info->name_len, info->name_len, fanotify_info_name(info));
return &fne->fae;
}
static struct fanotify_event *fanotify_alloc_event(struct fsnotify_group *group,
u32 mask, const void *data,
int data_type, struct inode *dir,
const struct qstr *file_name,
__kernel_fsid_t *fsid)
{
struct fanotify_event *event = NULL;
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-18 06:46:39 +08:00
gfp_t gfp = GFP_KERNEL_ACCOUNT;
struct inode *id = fanotify_fid_inode(mask, data, data_type, dir);
struct inode *dirid = fanotify_dfid_inode(mask, data, data_type, dir);
const struct path *path = fsnotify_data_path(data, data_type);
unsigned int fid_mode = FAN_GROUP_FLAG(group, FANOTIFY_FID_BITS);
mm, memcg: rework remote charging API to support nesting Currently the remote memcg charging API consists of two functions: memalloc_use_memcg() and memalloc_unuse_memcg(), which set and clear the memcg value, which overwrites the memcg of the current task. memalloc_use_memcg(target_memcg); <...> memalloc_unuse_memcg(); It works perfectly for allocations performed from a normal context, however an attempt to call it from an interrupt context or just nest two remote charging blocks will lead to an incorrect accounting. On exit from the inner block the active memcg will be cleared instead of being restored. memalloc_use_memcg(target_memcg); memalloc_use_memcg(target_memcg_2); <...> memalloc_unuse_memcg(); Error: allocation here are charged to the memcg of the current process instead of target_memcg. memalloc_unuse_memcg(); This patch extends the remote charging API by switching to a single function: struct mem_cgroup *set_active_memcg(struct mem_cgroup *memcg), which sets the new value and returns the old one. So a remote charging block will look like: old_memcg = set_active_memcg(target_memcg); <...> set_active_memcg(old_memcg); This patch is heavily based on the patch by Johannes Weiner, which can be found here: https://lkml.org/lkml/2020/5/28/806 . Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Dan Schatzberg <dschatzberg@fb.com> Link: https://lkml.kernel.org/r/20200821212056.3769116-1-guro@fb.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-10-18 07:13:40 +08:00
struct mem_cgroup *old_memcg;
struct inode *child = NULL;
bool name_event = false;
unsigned int hash = 0;
bool ondir = mask & FAN_ONDIR;
struct pid *pid;
if ((fid_mode & FAN_REPORT_DIR_FID) && dirid) {
/*
* With both flags FAN_REPORT_DIR_FID and FAN_REPORT_FID, we
* report the child fid for events reported on a non-dir child
* in addition to reporting the parent fid and maybe child name.
*/
if ((fid_mode & FAN_REPORT_FID) && id != dirid && !ondir)
child = id;
id = dirid;
/*
* We record file name only in a group with FAN_REPORT_NAME
* and when we have a directory inode to report.
*
* For directory entry modification event, we record the fid of
* the directory and the name of the modified entry.
*
* For event on non-directory that is reported to parent, we
* record the fid of the parent and the name of the child.
*
* Even if not reporting name, we need a variable length
* fanotify_name_event if reporting both parent and child fids.
*/
if (!(fid_mode & FAN_REPORT_NAME)) {
name_event = !!child;
file_name = NULL;
} else if ((mask & ALL_FSNOTIFY_DIRENT_EVENTS) || !ondir) {
name_event = true;
}
}
/*
* For queues with unlimited length lost events are not expected and
* can possibly have security implications. Avoid losing events when
* memory is short. For the limited size queues, avoid OOM killer in the
* target monitoring memcg as it may have security repercussion.
*/
if (group->max_events == UINT_MAX)
gfp |= __GFP_NOFAIL;
else
gfp |= __GFP_RETRY_MAYFAIL;
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-18 06:46:39 +08:00
/* Whoever is interested in the event, pays for the allocation. */
mm, memcg: rework remote charging API to support nesting Currently the remote memcg charging API consists of two functions: memalloc_use_memcg() and memalloc_unuse_memcg(), which set and clear the memcg value, which overwrites the memcg of the current task. memalloc_use_memcg(target_memcg); <...> memalloc_unuse_memcg(); It works perfectly for allocations performed from a normal context, however an attempt to call it from an interrupt context or just nest two remote charging blocks will lead to an incorrect accounting. On exit from the inner block the active memcg will be cleared instead of being restored. memalloc_use_memcg(target_memcg); memalloc_use_memcg(target_memcg_2); <...> memalloc_unuse_memcg(); Error: allocation here are charged to the memcg of the current process instead of target_memcg. memalloc_unuse_memcg(); This patch extends the remote charging API by switching to a single function: struct mem_cgroup *set_active_memcg(struct mem_cgroup *memcg), which sets the new value and returns the old one. So a remote charging block will look like: old_memcg = set_active_memcg(target_memcg); <...> set_active_memcg(old_memcg); This patch is heavily based on the patch by Johannes Weiner, which can be found here: https://lkml.org/lkml/2020/5/28/806 . Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Dan Schatzberg <dschatzberg@fb.com> Link: https://lkml.kernel.org/r/20200821212056.3769116-1-guro@fb.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-10-18 07:13:40 +08:00
old_memcg = set_active_memcg(group->memcg);
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-18 06:46:39 +08:00
if (fanotify_is_perm_event(mask)) {
event = fanotify_alloc_perm_event(path, gfp);
} else if (name_event && (file_name || child)) {
event = fanotify_alloc_name_event(id, fsid, file_name, child,
&hash, gfp);
} else if (fid_mode) {
event = fanotify_alloc_fid_event(id, fsid, &hash, gfp);
} else {
event = fanotify_alloc_path_event(path, &hash, gfp);
}
if (!event)
goto out;
if (FAN_GROUP_FLAG(group, FAN_REPORT_TID))
pid = get_pid(task_pid(current));
else
pid = get_pid(task_tgid(current));
/* Mix event info, FAN_ONDIR flag and pid into event merge key */
hash ^= hash_long((unsigned long)pid | ondir, FANOTIFY_EVENT_HASH_BITS);
fanotify_init_event(event, hash, mask);
event->pid = pid;
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-18 06:46:39 +08:00
out:
mm, memcg: rework remote charging API to support nesting Currently the remote memcg charging API consists of two functions: memalloc_use_memcg() and memalloc_unuse_memcg(), which set and clear the memcg value, which overwrites the memcg of the current task. memalloc_use_memcg(target_memcg); <...> memalloc_unuse_memcg(); It works perfectly for allocations performed from a normal context, however an attempt to call it from an interrupt context or just nest two remote charging blocks will lead to an incorrect accounting. On exit from the inner block the active memcg will be cleared instead of being restored. memalloc_use_memcg(target_memcg); memalloc_use_memcg(target_memcg_2); <...> memalloc_unuse_memcg(); Error: allocation here are charged to the memcg of the current process instead of target_memcg. memalloc_unuse_memcg(); This patch extends the remote charging API by switching to a single function: struct mem_cgroup *set_active_memcg(struct mem_cgroup *memcg), which sets the new value and returns the old one. So a remote charging block will look like: old_memcg = set_active_memcg(target_memcg); <...> set_active_memcg(old_memcg); This patch is heavily based on the patch by Johannes Weiner, which can be found here: https://lkml.org/lkml/2020/5/28/806 . Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Dan Schatzberg <dschatzberg@fb.com> Link: https://lkml.kernel.org/r/20200821212056.3769116-1-guro@fb.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-10-18 07:13:40 +08:00
set_active_memcg(old_memcg);
return event;
}
/*
* Get cached fsid of the filesystem containing the object from any connector.
* All connectors are supposed to have the same fsid, but we do not verify that
* here.
*/
static __kernel_fsid_t fanotify_get_fsid(struct fsnotify_iter_info *iter_info)
{
int type;
__kernel_fsid_t fsid = {};
fsnotify_foreach_obj_type(type) {
struct fsnotify_mark_connector *conn;
if (!fsnotify_iter_should_report_type(iter_info, type))
continue;
conn = READ_ONCE(iter_info->marks[type]->connector);
/* Mark is just getting destroyed or created? */
if (!conn)
continue;
if (!(conn->flags & FSNOTIFY_CONN_FLAG_HAS_FSID))
continue;
/* Pairs with smp_wmb() in fsnotify_add_mark_list() */
smp_rmb();
fsid = conn->fsid;
if (WARN_ON_ONCE(!fsid.val[0] && !fsid.val[1]))
continue;
return fsid;
}
return fsid;
}
/*
* Add an event to hash table for faster merge.
*/
static void fanotify_insert_event(struct fsnotify_group *group,
struct fsnotify_event *fsn_event)
{
struct fanotify_event *event = FANOTIFY_E(fsn_event);
unsigned int bucket = fanotify_event_hash_bucket(group, event);
struct hlist_head *hlist = &group->fanotify_data.merge_hash[bucket];
assert_spin_locked(&group->notification_lock);
pr_debug("%s: group=%p event=%p bucket=%u\n", __func__,
group, event, bucket);
hlist_add_head(&event->merge_list, hlist);
}
static int fanotify_handle_event(struct fsnotify_group *group, u32 mask,
const void *data, int data_type,
struct inode *dir,
const struct qstr *file_name, u32 cookie,
struct fsnotify_iter_info *iter_info)
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
{
int ret = 0;
struct fanotify_event *event;
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
struct fsnotify_event *fsn_event;
__kernel_fsid_t fsid = {};
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
BUILD_BUG_ON(FAN_ACCESS != FS_ACCESS);
BUILD_BUG_ON(FAN_MODIFY != FS_MODIFY);
BUILD_BUG_ON(FAN_ATTRIB != FS_ATTRIB);
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
BUILD_BUG_ON(FAN_CLOSE_NOWRITE != FS_CLOSE_NOWRITE);
BUILD_BUG_ON(FAN_CLOSE_WRITE != FS_CLOSE_WRITE);
BUILD_BUG_ON(FAN_OPEN != FS_OPEN);
BUILD_BUG_ON(FAN_MOVED_TO != FS_MOVED_TO);
BUILD_BUG_ON(FAN_MOVED_FROM != FS_MOVED_FROM);
BUILD_BUG_ON(FAN_CREATE != FS_CREATE);
BUILD_BUG_ON(FAN_DELETE != FS_DELETE);
BUILD_BUG_ON(FAN_DELETE_SELF != FS_DELETE_SELF);
BUILD_BUG_ON(FAN_MOVE_SELF != FS_MOVE_SELF);
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
BUILD_BUG_ON(FAN_EVENT_ON_CHILD != FS_EVENT_ON_CHILD);
BUILD_BUG_ON(FAN_Q_OVERFLOW != FS_Q_OVERFLOW);
BUILD_BUG_ON(FAN_OPEN_PERM != FS_OPEN_PERM);
BUILD_BUG_ON(FAN_ACCESS_PERM != FS_ACCESS_PERM);
BUILD_BUG_ON(FAN_ONDIR != FS_ISDIR);
BUILD_BUG_ON(FAN_OPEN_EXEC != FS_OPEN_EXEC);
BUILD_BUG_ON(FAN_OPEN_EXEC_PERM != FS_OPEN_EXEC_PERM);
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
BUILD_BUG_ON(HWEIGHT32(ALL_FANOTIFY_EVENT_BITS) != 19);
mask = fanotify_group_event_mask(group, iter_info, mask, data,
data_type, dir);
fanotify: return only user requested event types in event mask Modify fanotify_should_send_event() so that it now returns a mask for an event that contains ONLY flags for the event types that have been specifically requested by the user. Flags that may have been included within the event mask, but have not been explicitly requested by the user will not be present in the returned value. As an example, given the situation where a user requests events of type FAN_OPEN. Traditionally, the event mask returned within an event that occurred on a filesystem object that has been marked for monitoring and is opened, will only ever have the FAN_OPEN bit set. With the introduction of the new flags like FAN_OPEN_EXEC, and perhaps any other future event flags, there is a possibility of the returned event mask containing more than a single bit set, despite having only requested the single event type. Prior to these modifications performed to fanotify_should_send_event(), a user would have received a bundled event mask containing flags FAN_OPEN and FAN_OPEN_EXEC in the instance that a file was opened for execution via execve(), for example. This means that a user would receive event types in the returned event mask that have not been requested. This runs the possibility of breaking existing systems and causing other unforeseen issues. To mitigate this possibility, fanotify_should_send_event() has been modified to return the event mask containing ONLY event types explicitly requested by the user. This means that we will NOT report events that the user did no set a mask for, and we will NOT report events that the user has set an ignore mask for. The function name fanotify_should_send_event() has also been updated so that it's more relevant to what it has been designed to do. Signed-off-by: Matthew Bobrowski <mbobrowski@mbobrowski.org> Reviewed-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2018-11-08 11:05:49 +08:00
if (!mask)
return 0;
pr_debug("%s: group=%p mask=%x\n", __func__, group, mask);
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
if (fanotify_is_perm_event(mask)) {
/*
* fsnotify_prepare_user_wait() fails if we race with mark
* deletion. Just let the operation pass in that case.
*/
if (!fsnotify_prepare_user_wait(iter_info))
return 0;
}
if (FAN_GROUP_FLAG(group, FANOTIFY_FID_BITS)) {
fsid = fanotify_get_fsid(iter_info);
/* Racing with mark destruction or creation? */
if (!fsid.val[0] && !fsid.val[1])
return 0;
}
event = fanotify_alloc_event(group, mask, data, data_type, dir,
file_name, &fsid);
ret = -ENOMEM;
if (unlikely(!event)) {
/*
* We don't queue overflow events for permission events as
* there the access is denied and so no event is in fact lost.
*/
if (!fanotify_is_perm_event(mask))
fsnotify_queue_overflow(group);
goto finish;
}
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
fsn_event = &event->fse;
ret = fsnotify_add_event(group, fsn_event, fanotify_merge,
fanotify_is_hashed_event(mask) ?
fanotify_insert_event : NULL);
if (ret) {
/* Permission events shouldn't be merged */
BUG_ON(ret == 1 && mask & FANOTIFY_PERM_EVENTS);
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
/* Our event wasn't used in the end. Free it. */
fsnotify_destroy_event(group, fsn_event);
ret = 0;
} else if (fanotify_is_perm_event(mask)) {
ret = fanotify_get_response(group, FANOTIFY_PERM(event),
iter_info);
}
finish:
if (fanotify_is_perm_event(mask))
fsnotify_finish_user_wait(iter_info);
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
return ret;
}
static void fanotify_free_group_priv(struct fsnotify_group *group)
{
kfree(group->fanotify_data.merge_hash);
fanotify: configurable limits via sysfs fanotify has some hardcoded limits. The only APIs to escape those limits are FAN_UNLIMITED_QUEUE and FAN_UNLIMITED_MARKS. Allow finer grained tuning of the system limits via sysfs tunables under /proc/sys/fs/fanotify, similar to tunables under /proc/sys/fs/inotify, with some minor differences. - max_queued_events - global system tunable for group queue size limit. Like the inotify tunable with the same name, it defaults to 16384 and applies on initialization of a new group. - max_user_marks - user ns tunable for marks limit per user. Like the inotify tunable named max_user_watches, on a machine with sufficient RAM and it defaults to 1048576 in init userns and can be further limited per containing user ns. - max_user_groups - user ns tunable for number of groups per user. Like the inotify tunable named max_user_instances, it defaults to 128 in init userns and can be further limited per containing user ns. The slightly different tunable names used for fanotify are derived from the "group" and "mark" terminology used in the fanotify man pages and throughout the code. Considering the fact that the default value for max_user_instances was increased in kernel v5.10 from 8192 to 1048576, leaving the legacy fanotify limit of 8192 marks per group in addition to the max_user_marks limit makes little sense, so the per group marks limit has been removed. Note that when a group is initialized with FAN_UNLIMITED_MARKS, its own marks are not accounted in the per user marks account, so in effect the limit of max_user_marks is only for the collection of groups that are not initialized with FAN_UNLIMITED_MARKS. Link: https://lore.kernel.org/r/20210304112921.3996419-2-amir73il@gmail.com Suggested-by: Jan Kara <jack@suse.cz> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2021-03-04 19:29:20 +08:00
if (group->fanotify_data.ucounts)
dec_ucount(group->fanotify_data.ucounts,
UCOUNT_FANOTIFY_GROUPS);
}
static void fanotify_free_path_event(struct fanotify_event *event)
{
path_put(fanotify_event_path(event));
kmem_cache_free(fanotify_path_event_cachep, FANOTIFY_PE(event));
}
static void fanotify_free_perm_event(struct fanotify_event *event)
{
path_put(fanotify_event_path(event));
kmem_cache_free(fanotify_perm_event_cachep, FANOTIFY_PERM(event));
}
static void fanotify_free_fid_event(struct fanotify_event *event)
{
struct fanotify_fid_event *ffe = FANOTIFY_FE(event);
if (fanotify_fh_has_ext_buf(&ffe->object_fh))
kfree(fanotify_fh_ext_buf(&ffe->object_fh));
kmem_cache_free(fanotify_fid_event_cachep, ffe);
}
static void fanotify_free_name_event(struct fanotify_event *event)
{
kfree(FANOTIFY_NE(event));
}
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
static void fanotify_free_event(struct fsnotify_event *fsn_event)
{
struct fanotify_event *event;
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
event = FANOTIFY_E(fsn_event);
put_pid(event->pid);
switch (event->type) {
case FANOTIFY_EVENT_TYPE_PATH:
fanotify_free_path_event(event);
break;
case FANOTIFY_EVENT_TYPE_PATH_PERM:
fanotify_free_perm_event(event);
break;
case FANOTIFY_EVENT_TYPE_FID:
fanotify_free_fid_event(event);
break;
case FANOTIFY_EVENT_TYPE_FID_NAME:
fanotify_free_name_event(event);
break;
case FANOTIFY_EVENT_TYPE_OVERFLOW:
kfree(event);
break;
default:
WARN_ON_ONCE(1);
}
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
}
fanotify: configurable limits via sysfs fanotify has some hardcoded limits. The only APIs to escape those limits are FAN_UNLIMITED_QUEUE and FAN_UNLIMITED_MARKS. Allow finer grained tuning of the system limits via sysfs tunables under /proc/sys/fs/fanotify, similar to tunables under /proc/sys/fs/inotify, with some minor differences. - max_queued_events - global system tunable for group queue size limit. Like the inotify tunable with the same name, it defaults to 16384 and applies on initialization of a new group. - max_user_marks - user ns tunable for marks limit per user. Like the inotify tunable named max_user_watches, on a machine with sufficient RAM and it defaults to 1048576 in init userns and can be further limited per containing user ns. - max_user_groups - user ns tunable for number of groups per user. Like the inotify tunable named max_user_instances, it defaults to 128 in init userns and can be further limited per containing user ns. The slightly different tunable names used for fanotify are derived from the "group" and "mark" terminology used in the fanotify man pages and throughout the code. Considering the fact that the default value for max_user_instances was increased in kernel v5.10 from 8192 to 1048576, leaving the legacy fanotify limit of 8192 marks per group in addition to the max_user_marks limit makes little sense, so the per group marks limit has been removed. Note that when a group is initialized with FAN_UNLIMITED_MARKS, its own marks are not accounted in the per user marks account, so in effect the limit of max_user_marks is only for the collection of groups that are not initialized with FAN_UNLIMITED_MARKS. Link: https://lore.kernel.org/r/20210304112921.3996419-2-amir73il@gmail.com Suggested-by: Jan Kara <jack@suse.cz> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2021-03-04 19:29:20 +08:00
static void fanotify_freeing_mark(struct fsnotify_mark *mark,
struct fsnotify_group *group)
{
if (!FAN_GROUP_FLAG(group, FAN_UNLIMITED_MARKS))
dec_ucount(group->fanotify_data.ucounts, UCOUNT_FANOTIFY_MARKS);
}
static void fanotify_free_mark(struct fsnotify_mark *fsn_mark)
{
kmem_cache_free(fanotify_mark_cache, fsn_mark);
}
const struct fsnotify_ops fanotify_fsnotify_ops = {
.handle_event = fanotify_handle_event,
.free_group_priv = fanotify_free_group_priv,
fsnotify: do not share events between notification groups Currently fsnotify framework creates one event structure for each notification event and links this event into all interested notification groups. This is done so that we save memory when several notification groups are interested in the event. However the need for event structure shared between inotify & fanotify bloats the event structure so the result is often higher memory consumption. Another problem is that fsnotify framework keeps path references with outstanding events so that fanotify can return open file descriptors with its events. This has the undesirable effect that filesystem cannot be unmounted while there are outstanding events - a regression for inotify compared to a situation before it was converted to fsnotify framework. For fanotify this problem is hard to avoid and users of fanotify should kind of expect this behavior when they ask for file descriptors from notified files. This patch changes fsnotify and its users to create separate event structure for each group. This allows for much simpler code (~400 lines removed by this patch) and also smaller event structures. For example on 64-bit system original struct fsnotify_event consumes 120 bytes, plus additional space for file name, additional 24 bytes for second and each subsequent group linking the event, and additional 32 bytes for each inotify group for private data. After the conversion inotify event consumes 48 bytes plus space for file name which is considerably less memory unless file names are long and there are several groups interested in the events (both of which are uncommon). Fanotify event fits in 56 bytes after the conversion (fanotify doesn't care about file names so its events don't have to have it allocated). A win unless there are four or more fanotify groups interested in the event. The conversion also solves the problem with unmount when only inotify is used as we don't have to grab path references for inotify events. [hughd@google.com: fanotify: fix corruption preventing startup] Signed-off-by: Jan Kara <jack@suse.cz> Reviewed-by: Christoph Hellwig <hch@lst.de> Cc: Eric Paris <eparis@parisplace.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-22 07:48:14 +08:00
.free_event = fanotify_free_event,
fanotify: configurable limits via sysfs fanotify has some hardcoded limits. The only APIs to escape those limits are FAN_UNLIMITED_QUEUE and FAN_UNLIMITED_MARKS. Allow finer grained tuning of the system limits via sysfs tunables under /proc/sys/fs/fanotify, similar to tunables under /proc/sys/fs/inotify, with some minor differences. - max_queued_events - global system tunable for group queue size limit. Like the inotify tunable with the same name, it defaults to 16384 and applies on initialization of a new group. - max_user_marks - user ns tunable for marks limit per user. Like the inotify tunable named max_user_watches, on a machine with sufficient RAM and it defaults to 1048576 in init userns and can be further limited per containing user ns. - max_user_groups - user ns tunable for number of groups per user. Like the inotify tunable named max_user_instances, it defaults to 128 in init userns and can be further limited per containing user ns. The slightly different tunable names used for fanotify are derived from the "group" and "mark" terminology used in the fanotify man pages and throughout the code. Considering the fact that the default value for max_user_instances was increased in kernel v5.10 from 8192 to 1048576, leaving the legacy fanotify limit of 8192 marks per group in addition to the max_user_marks limit makes little sense, so the per group marks limit has been removed. Note that when a group is initialized with FAN_UNLIMITED_MARKS, its own marks are not accounted in the per user marks account, so in effect the limit of max_user_marks is only for the collection of groups that are not initialized with FAN_UNLIMITED_MARKS. Link: https://lore.kernel.org/r/20210304112921.3996419-2-amir73il@gmail.com Suggested-by: Jan Kara <jack@suse.cz> Signed-off-by: Amir Goldstein <amir73il@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz>
2021-03-04 19:29:20 +08:00
.freeing_mark = fanotify_freeing_mark,
.free_mark = fanotify_free_mark,
};