mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Ben Hutchings <ben@decadent.org.uk>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: akpm@linux-foundation.org, "Tejun Heo" <tj@kernel.org>,
	"Fengguang Wu" <fengguang.wu@intel.com>,
	"Li Zefan" <lizefan@huawei.com>
Subject: [45/85] cgroup: fix RCU accesses to task->cgroups
Date: Wed, 24 Jul 2013 15:02:45 +0100	[thread overview]
Message-ID: <lsq.1374674565.581226504@decadent.org.uk> (raw)
In-Reply-To: <lsq.1374674564.753305150@decadent.org.uk>

3.2.49-rc1 review patch.  If anyone has any objections, please let me know.

------------------

From: Tejun Heo <tj@kernel.org>

commit 14611e51a57df10240817d8ada510842faf0ec51 upstream.

task->cgroups is a RCU pointer pointing to struct css_set.  A task
switches to a different css_set on cgroup migration but a css_set
doesn't change once created and its pointers to cgroup_subsys_states
aren't RCU protected.

task_subsys_state[_check]() is the macro to acquire css given a task
and subsys_id pair.  It RCU-dereferences task->cgroups->subsys[] not
task->cgroups, so the RCU pointer task->cgroups ends up being
dereferenced without read_barrier_depends() after it.  It's broken.

Fix it by introducing task_css_set[_check]() which does
RCU-dereference on task->cgroups.  task_subsys_state[_check]() is
reimplemented to directly dereference ->subsys[] of the css_set
returned from task_css_set[_check]().

This removes some of sparse RCU warnings in cgroup.

v2: Fixed unbalanced parenthsis and there's no need to use
    rcu_dereference_raw() when !CONFIG_PROVE_RCU.  Both spotted by Li.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Acked-by: Li Zefan <lizefan@huawei.com>
[bwh: Backported to 3.2:
 - Adjust context
 - Remove CONFIG_PROVE_RCU condition
 - s/lockdep_is_held(&cgroup_mutex)/cgroup_lock_is_held()/]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
---
 include/linux/cgroup.h | 58 +++++++++++++++++++++++++++++++++++++++++---------
 1 file changed, 48 insertions(+), 10 deletions(-)

--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -531,16 +531,54 @@ static inline struct cgroup_subsys_state
 	return cgrp->subsys[subsys_id];
 }
 
-/*
- * function to get the cgroup_subsys_state which allows for extra
- * rcu_dereference_check() conditions, such as locks used during the
- * cgroup_subsys::attach() methods.
+/**
+ * task_css_set_check - obtain a task's css_set with extra access conditions
+ * @task: the task to obtain css_set for
+ * @__c: extra condition expression to be passed to rcu_dereference_check()
+ *
+ * A task's css_set is RCU protected, initialized and exited while holding
+ * task_lock(), and can only be modified while holding both cgroup_mutex
+ * and task_lock() while the task is alive.  This macro verifies that the
+ * caller is inside proper critical section and returns @task's css_set.
+ *
+ * The caller can also specify additional allowed conditions via @__c, such
+ * as locks used during the cgroup_subsys::attach() methods.
+ */
+#define task_css_set_check(task, __c)					\
+	rcu_dereference_check((task)->cgroups,				\
+		lockdep_is_held(&(task)->alloc_lock) ||			\
+		cgroup_lock_is_held() || (__c))
+
+/**
+ * task_subsys_state_check - obtain css for (task, subsys) w/ extra access conds
+ * @task: the target task
+ * @subsys_id: the target subsystem ID
+ * @__c: extra condition expression to be passed to rcu_dereference_check()
+ *
+ * Return the cgroup_subsys_state for the (@task, @subsys_id) pair.  The
+ * synchronization rules are the same as task_css_set_check().
  */
 #define task_subsys_state_check(task, subsys_id, __c)			\
-	rcu_dereference_check(task->cgroups->subsys[subsys_id],		\
-			      lockdep_is_held(&task->alloc_lock) ||	\
-			      cgroup_lock_is_held() || (__c))
+	task_css_set_check((task), (__c))->subsys[(subsys_id)]
 
+/**
+ * task_css_set - obtain a task's css_set
+ * @task: the task to obtain css_set for
+ *
+ * See task_css_set_check().
+ */
+static inline struct css_set *task_css_set(struct task_struct *task)
+{
+	return task_css_set_check(task, false);
+}
+
+/**
+ * task_subsys_state - obtain css for (task, subsys)
+ * @task: the target task
+ * @subsys_id: the target subsystem ID
+ *
+ * See task_subsys_state_check().
+ */
 static inline struct cgroup_subsys_state *
 task_subsys_state(struct task_struct *task, int subsys_id)
 {


  parent reply	other threads:[~2013-07-24 14:14 UTC|newest]

Thread overview: 100+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-07-24 14:02 [00/85] 3.2.49-rc1 review Ben Hutchings
2013-07-24 14:02 ` [79/85] perf: Fix mmap() accounting hole Ben Hutchings
2013-07-24 14:02 ` [29/85] hw_breakpoint: Use cpu_possible_mask in {reserve,release}_bp_slot() Ben Hutchings
2013-07-24 14:02 ` [17/85] ath9k_hw: Assign default xlna config for AR9485 Ben Hutchings
2013-07-24 14:02 ` [56/85] powerpc/smp: Section mismatch from smp_release_cpus to __initdata spinning_secondaries Ben Hutchings
2013-07-24 14:02 ` [77/85] ceph: fix statvfs fr_size Ben Hutchings
2013-07-24 14:02 ` [11/85] zram: avoid access beyond the zram device Ben Hutchings
2013-07-24 14:02 ` [80/85] pch_uart: fix a deadlock when pch_uart as console Ben Hutchings
2013-07-24 14:02 ` [70/85] nbd: correct disconnect behavior Ben Hutchings
2013-07-24 14:02 ` [72/85] [SCSI] megaraid_sas: fix memory leak if SGL has zero length entries Ben Hutchings
2013-07-24 14:02 ` [12/85] zram: protect sysfs handler from invalid memory access Ben Hutchings
2013-07-24 14:02 ` [51/85] timer: Fix jiffies wrap behavior of round_jiffies_common() Ben Hutchings
2013-07-24 14:02 ` [71/85] hpfs: better test for errors Ben Hutchings
2013-07-24 14:02 ` [13/85] ahci: remove pmp link online check in FBS EH Ben Hutchings
2013-07-24 14:02 ` [16/85] b43: ensue that BCMA is "y" when B43 is "y" Ben Hutchings
2013-09-09 18:01   ` Steven Rostedt
2013-09-09 18:39     ` Ben Hutchings
2013-09-09 18:46       ` Steven Rostedt
2013-09-09 20:56         ` Konstantin Ryabitsev
2013-09-11  1:00           ` Ben Hutchings
2013-07-24 14:02 ` [85/85] MAINTAINERS: add stable_kernel_rules.txt to stable maintainer information Ben Hutchings
2013-07-24 14:02 ` [22/85] ASoC: wm8962: Remove remaining direct register cache accesses Ben Hutchings
2013-07-24 14:02 ` [34/85] ARM: 7765/1: perf: Record the user-mode PC in the call chain Ben Hutchings
2013-07-24 14:02 ` [63/85] Input: bcm5974 - add support for the 2013 MacBook Air Ben Hutchings
2013-07-25  6:48   ` rydberg
2013-07-24 14:02 ` [76/85] perf: Fix perf_lock_task_context() vs RCU Ben Hutchings
2013-07-24 14:02 ` [40/85] net/tg3: Avoid delay during MMIO access Ben Hutchings
2013-07-24 16:42   ` Luis Henriques
2013-07-24 19:45     ` David Miller
2013-07-24 21:09       ` David Miller
2013-07-25  4:16         ` Ben Hutchings
2013-07-25 21:03         ` Greg KH
2013-07-25  4:15     ` Ben Hutchings
2013-07-24 14:02 ` [57/85] ext3,ext4: don't mess with dir_file->f_pos in htree_dirblock_to_tree() Ben Hutchings
2013-07-24 14:02 ` [33/85] iommu/amd: Only unmap large pages from the first pte Ben Hutchings
2013-07-24 14:02 ` [84/85] MAINTAINERS: Greg's suse email address is dead Ben Hutchings
2013-07-24 14:02 ` [53/85] UBIFS: prepare to fix a horrid bug Ben Hutchings
2013-07-24 14:02 ` [62/85] HID: apple: Add support for the 2013 Macbook Air Ben Hutchings
2013-07-25  6:50   ` rydberg
2013-07-26  2:49     ` Ben Hutchings
2013-07-24 14:02 ` [21/85] Bluetooth: Fix crash in l2cap_build_cmd() with small MTU Ben Hutchings
2013-07-24 14:02 ` [06/85] ahci: Add AMD CZ SATA device ID Ben Hutchings
2013-07-24 14:02 ` [41/85] vgacon.c: add cond reschedule points in vgacon_do_font_op Ben Hutchings
2013-07-24 14:02 ` [49/85] genirq: Fix can_request_irq() for IRQs without an action Ben Hutchings
2013-07-24 14:02 ` [18/85] ath9k: fill channel mode in caldata Ben Hutchings
2013-07-24 14:02 ` [83/85] ext4: fix data offset overflow in ext4_xattr_fiemap() on 32-bit archs Ben Hutchings
2013-07-24 14:02 ` [54/85] UBIFS: fix a horrid bug Ben Hutchings
2013-07-24 14:02 ` [08/85] zram: avoid invalid memory access in zram_exit() Ben Hutchings
2013-07-24 14:02 ` [65/85] drivers/dma/pl330.c: fix locking in pl330_free_chan_resources() Ben Hutchings
2013-07-24 14:02 ` [36/85] [SCSI] mpt2sas: Fix for device scan following host reset could get stuck in a infinite loop Ben Hutchings
2013-07-24 14:02 ` [05/85] [SCSI] zfcp: status read buffers on first adapter open with link down Ben Hutchings
2013-07-24 14:02 ` [69/85] drivers/rtc/rtc-rv3029c2.c: fix disabling AIE irq Ben Hutchings
2013-07-24 14:02 ` [03/85] [SCSI] zfcp: fix adapter (re)open recovery while link to SAN is down Ben Hutchings
2013-07-24 14:02 ` [68/85] crypto: sanitize argument for format string Ben Hutchings
2013-07-24 14:02 ` [60/85] tracing: Fix irqs-off tag display in syscall tracing Ben Hutchings
2013-07-24 14:02 ` [14/85] usb: gadget: f_mass_storage: add missing memory barrier for thread_wakeup_needed Ben Hutchings
2013-07-24 14:02 ` [82/85] ext4: fix overflow when counting used blocks on 32-bit architectures Ben Hutchings
2013-07-24 14:02 ` [74/85] perf: Clone child context from parent context pmu Ben Hutchings
2013-07-24 14:02 ` [58/85] jbd2: fix theoretical race in jbd2__journal_restart Ben Hutchings
2013-07-24 14:02 ` [46/85] Handle big endianness in NTLM (ntlmv2) authentication Ben Hutchings
2013-07-24 14:02 ` [19/85] ath9k: Fix noisefloor calibration Ben Hutchings
2013-07-24 14:02 ` Ben Hutchings [this message]
2013-07-24 14:02 ` [73/85] iscsi-target: Fix tfc_tpg_nacl_auth_cit configfs length overflow Ben Hutchings
2013-07-24 14:02 ` [39/85] futex: Take hugepages into account when generating futex_key Ben Hutchings
2013-07-24 14:02 ` [23/85] rtlwifi: rtl8192cu: Add new USB ID for TP-Link TL-WN8200ND Ben Hutchings
2013-07-24 14:02 ` [38/85] USB: option,qcserial: move Novatel Gobi1K IDs to qcserial Ben Hutchings
2013-07-24 14:02 ` [15/85] libata: skip SRST for all SIMG [34]7x port-multipliers Ben Hutchings
2013-07-24 14:02 ` [61/85] HID: Add support for the MacBook Pro 10,2 keyboard / touchpad Ben Hutchings
2013-07-24 14:02 ` [50/85] writeback: Fix periodic writeback after fs mount Ben Hutchings
2013-07-24 14:02 ` [47/85] dlci: acquire rtnl_lock before calling __dev_get_by_name() Ben Hutchings
2013-07-24 14:02 ` [27/85] [media] media: dmxdev: remove dvb_ringbuffer_flush() on writer side Ben Hutchings
2013-07-24 14:02 ` [81/85] drivers/cdrom/cdrom.c: use kzalloc() for failing hardware Ben Hutchings
2013-07-24 14:02 ` [26/85] drivers: hv: switch to use mb() instead of smp_mb() Ben Hutchings
2013-07-24 14:02 ` [37/85] [SCSI] mpt2sas: fix firmware failure with wrong task attribute Ben Hutchings
2013-07-24 14:02 ` [25/85] xhci: check for failed dma pool allocation Ben Hutchings
2013-07-24 14:02 ` [07/85] i2c-piix4: Add AMD CZ SMBus device ID Ben Hutchings
2013-07-24 14:02 ` [01/85] r8169: fix offloaded tx checksum for small packets Ben Hutchings
2013-07-24 14:02 ` [31/85] ahci: AHCI-mode SATA patch for Intel Coleto Creek DeviceIDs Ben Hutchings
2013-07-24 14:02 ` [09/85] zram: use zram->lock to protect zram_free_page() in swap free notify path Ben Hutchings
2013-07-24 14:02 ` [24/85] ALSA: hda - Cache the MUX selection for generic HDMI Ben Hutchings
2013-07-24 14:02 ` [20/85] ath9k: Do not assign noise for NULL caldata Ben Hutchings
2013-07-24 14:02 ` [42/85] perf: Disable monitoring on setuid processes for regular users Ben Hutchings
2013-07-24 14:02 ` [43/85] [SCSI] sd: fix array cache flushing bug causing performance problems Ben Hutchings
2013-07-24 14:02 ` [59/85] nfsd4: fix decoding of compounds across page boundaries Ben Hutchings
2013-07-24 14:02 ` [64/85] libceph: Fix NULL pointer dereference in auth client code Ben Hutchings
2013-07-24 14:02 ` [30/85] ata_piix: IDE-mode SATA patch for Intel Coleto Creek DeviceIDs Ben Hutchings
2013-07-24 14:02 ` [32/85] Bluetooth: Fix invalid length check in l2cap_information_rsp() Ben Hutchings
2013-07-24 14:02 ` [35/85] [SCSI] mpt2sas: Fix for issue Missing delay not getting set during system bootup Ben Hutchings
2013-07-24 14:02 ` [10/85] zram: destroy all devices on error recovery path in zram_init() Ben Hutchings
2013-07-24 14:02 ` [48/85] dlci: validate the net device in dlci_del() Ben Hutchings
2013-07-24 14:02 ` [04/85] [SCSI] zfcp: block queue limits with data router Ben Hutchings
2013-07-24 14:02 ` [78/85] perf: Fix perf mmap bugs Ben Hutchings
2013-07-24 14:02 ` [67/85] block: do not pass disk names as format strings Ben Hutchings
2013-07-24 14:02 ` [66/85] ocfs2: xattr: fix inlined xattr reflink Ben Hutchings
2013-07-24 14:02 ` [52/85] xen/time: remove blocked time accounting from xen "clockchip" Ben Hutchings
2013-07-24 14:02 ` [55/85] Revert "serial: 8250_pci: add support for another kind of NetMos Technology PCI 9835 Multi-I/O Controller" Ben Hutchings
2013-07-24 14:02 ` [02/85] printk: Fix rq->lock vs logbuf_lock unlock lock inversion Ben Hutchings
2013-07-24 14:02 ` [44/85] [SCSI] sd: Fix parsing of 'temporary ' cache mode prefix Ben Hutchings
2013-07-24 14:02 ` [75/85] perf: Remove WARN_ON_ONCE() check in __perf_event_enable() for valid scenario Ben Hutchings
2013-07-24 14:11 ` [00/85] 3.2.49-rc1 review Ben Hutchings

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=lsq.1374674565.581226504@decadent.org.uk \
    --to=ben@decadent.org.uk \
    --cc=akpm@linux-foundation.org \
    --cc=fengguang.wu@intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lizefan@huawei.com \
    --cc=stable@vger.kernel.org \
    --cc=tj@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®