From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
stable@vger.kernel.org, linux-raid@vger.kernel.org,
Oleg Nesterov <oleg@redhat.com>,
"Srivatsa S. Bhat" <srivatsa.bhat@linux.vnet.ibm.com>,
NeilBrown <neilb@suse.de>
Subject: [PATCH 3.12 73/82] md/raid5: Fix CPU hotplug callback registration
Date: Thu, 20 Feb 2014 15:52:56 -0800 [thread overview]
Message-ID: <20140220235021.672984378@linuxfoundation.org> (raw)
In-Reply-To: <20140220235019.483734207@linuxfoundation.org>
3.12-stable review patch. If anyone has any objections, please let me know.
------------------
From: Oleg Nesterov <oleg@redhat.com>
commit 789b5e0315284463617e106baad360cb9e8db3ac upstream.
Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
register_cpu_notifier(&foobar_cpu_notifier);
put_online_cpus();
This is wrong, since it is prone to ABBA deadlocks involving the
cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently
with CPU hotplug operations).
Interestingly, the raid5 code can actually prevent double initialization and
hence can use the following simplified form of callback registration:
register_cpu_notifier(&foobar_cpu_notifier);
get_online_cpus();
for_each_online_cpu(cpu)
init_cpu(cpu);
put_online_cpus();
A hotplug operation that occurs between registering the notifier and calling
get_online_cpus(), won't disrupt anything, because the code takes care to
perform the memory allocations only once.
So reorganize the code in raid5 this way to fix the deadlock with callback
registration.
Cc: linux-raid@vger.kernel.org
Fixes: 36d1c6476be51101778882897b315bd928c8c7b5
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
[Srivatsa: Fixed the unregister_cpu_notifier() deadlock, added the
free_scratch_buffer() helper to condense code further and wrote the changelog.]
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
drivers/md/raid5.c | 90 +++++++++++++++++++++++++----------------------------
1 file changed, 44 insertions(+), 46 deletions(-)
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -5351,23 +5351,43 @@ raid5_size(struct mddev *mddev, sector_t
return sectors * (raid_disks - conf->max_degraded);
}
+static void free_scratch_buffer(struct r5conf *conf, struct raid5_percpu *percpu)
+{
+ safe_put_page(percpu->spare_page);
+ kfree(percpu->scribble);
+ percpu->spare_page = NULL;
+ percpu->scribble = NULL;
+}
+
+static int alloc_scratch_buffer(struct r5conf *conf, struct raid5_percpu *percpu)
+{
+ if (conf->level == 6 && !percpu->spare_page)
+ percpu->spare_page = alloc_page(GFP_KERNEL);
+ if (!percpu->scribble)
+ percpu->scribble = kmalloc(conf->scribble_len, GFP_KERNEL);
+
+ if (!percpu->scribble || (conf->level == 6 && !percpu->spare_page)) {
+ free_scratch_buffer(conf, percpu);
+ return -ENOMEM;
+ }
+
+ return 0;
+}
+
static void raid5_free_percpu(struct r5conf *conf)
{
- struct raid5_percpu *percpu;
unsigned long cpu;
if (!conf->percpu)
return;
- get_online_cpus();
- for_each_possible_cpu(cpu) {
- percpu = per_cpu_ptr(conf->percpu, cpu);
- safe_put_page(percpu->spare_page);
- kfree(percpu->scribble);
- }
#ifdef CONFIG_HOTPLUG_CPU
unregister_cpu_notifier(&conf->cpu_notify);
#endif
+
+ get_online_cpus();
+ for_each_possible_cpu(cpu)
+ free_scratch_buffer(conf, per_cpu_ptr(conf->percpu, cpu));
put_online_cpus();
free_percpu(conf->percpu);
@@ -5394,15 +5414,7 @@ static int raid456_cpu_notify(struct not
switch (action) {
case CPU_UP_PREPARE:
case CPU_UP_PREPARE_FROZEN:
- if (conf->level == 6 && !percpu->spare_page)
- percpu->spare_page = alloc_page(GFP_KERNEL);
- if (!percpu->scribble)
- percpu->scribble = kmalloc(conf->scribble_len, GFP_KERNEL);
-
- if (!percpu->scribble ||
- (conf->level == 6 && !percpu->spare_page)) {
- safe_put_page(percpu->spare_page);
- kfree(percpu->scribble);
+ if (alloc_scratch_buffer(conf, percpu)) {
pr_err("%s: failed memory allocation for cpu%ld\n",
__func__, cpu);
return notifier_from_errno(-ENOMEM);
@@ -5410,10 +5422,7 @@ static int raid456_cpu_notify(struct not
break;
case CPU_DEAD:
case CPU_DEAD_FROZEN:
- safe_put_page(percpu->spare_page);
- kfree(percpu->scribble);
- percpu->spare_page = NULL;
- percpu->scribble = NULL;
+ free_scratch_buffer(conf, per_cpu_ptr(conf->percpu, cpu));
break;
default:
break;
@@ -5425,40 +5434,29 @@ static int raid456_cpu_notify(struct not
static int raid5_alloc_percpu(struct r5conf *conf)
{
unsigned long cpu;
- struct page *spare_page;
- struct raid5_percpu __percpu *allcpus;
- void *scribble;
- int err;
+ int err = 0;
- allcpus = alloc_percpu(struct raid5_percpu);
- if (!allcpus)
+ conf->percpu = alloc_percpu(struct raid5_percpu);
+ if (!conf->percpu)
return -ENOMEM;
- conf->percpu = allcpus;
+
+#ifdef CONFIG_HOTPLUG_CPU
+ conf->cpu_notify.notifier_call = raid456_cpu_notify;
+ conf->cpu_notify.priority = 0;
+ err = register_cpu_notifier(&conf->cpu_notify);
+ if (err)
+ return err;
+#endif
get_online_cpus();
- err = 0;
for_each_present_cpu(cpu) {
- if (conf->level == 6) {
- spare_page = alloc_page(GFP_KERNEL);
- if (!spare_page) {
- err = -ENOMEM;
- break;
- }
- per_cpu_ptr(conf->percpu, cpu)->spare_page = spare_page;
- }
- scribble = kmalloc(conf->scribble_len, GFP_KERNEL);
- if (!scribble) {
- err = -ENOMEM;
+ err = alloc_scratch_buffer(conf, per_cpu_ptr(conf->percpu, cpu));
+ if (err) {
+ pr_err("%s: failed memory allocation for cpu%ld\n",
+ __func__, cpu);
break;
}
- per_cpu_ptr(conf->percpu, cpu)->scribble = scribble;
}
-#ifdef CONFIG_HOTPLUG_CPU
- conf->cpu_notify.notifier_call = raid456_cpu_notify;
- conf->cpu_notify.priority = 0;
- if (err == 0)
- err = register_cpu_notifier(&conf->cpu_notify);
-#endif
put_online_cpus();
return err;
next prev parent reply other threads:[~2014-02-21 0:24 UTC|newest]
Thread overview: 78+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-02-20 23:51 [PATCH 3.12 00/82] 3.12.13-stable review Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 01/82] xen: properly account for _PAGE_NUMA during xen pte translations Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 02/82] xen-blkfront: handle backend CLOSED without CLOSING Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 03/82] fs/file.c:fdtable: avoid triggering OOMs from alloc_fdmem Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 04/82] mm: fix page leak at nfs_symlink() Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 05/82] mm/memory-failure.c: move refcount only in !MF_COUNT_INCREASED Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 06/82] CIFS: Fix SMB2 mounts so they dont try to set or get xattrs via cifs Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 07/82] Add protocol specific operation for CIFS xattrs Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 08/82] retrieving CIFS ACLs when mounted with SMB2 fails dropping session Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 09/82] mac80211: move roc cookie assignment earlier Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 10/82] mac80211: release the channel in error path in start_ap Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 11/82] mac80211: fix fragmentation code, particularly for encryption Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 12/82] ath9k_htc: make ->sta_rc_update atomic for most calls Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 13/82] ath9k_htc: Do not support PowerSave by default Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 14/82] ath9k: " Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 15/82] ar5523: fix usb id for Gigaset Greg Kroah-Hartman
2014-02-20 23:51 ` [PATCH 3.12 16/82] s390/dump: Fix dump memory detection Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 17/82] s390: fix kernel crash due to linkage stack instructions Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 18/82] nl80211: Reset split_start when netlink skb is exhausted Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 19/82] spi: Fix crash with double message finalisation on error handling Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 20/82] iwlwifi: mvm: dont allow A band if SKU forbids it Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 21/82] iwlwifi: mvm: print the version of the firmware when it asserts Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 22/82] iwlwifi: mvm: BT Coex - disable BT when TXing probe request in scan Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 23/82] of: fix PCI bus match for PCIe slots Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 24/82] spi: nuc900: Set SPI_LSB_FIRST for master->mode_bits if hw->pdata->lsb is true Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 25/82] serial: sirf: fix kernel panic caused by unpaired spinlock Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 26/82] raw: test against runtime value of max_raw_minors Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 27/82] hwmon: (ntc_thermistor) Avoid math overflow Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 28/82] lockd: send correct lock when granting a delayed lock Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 29/82] tty: n_gsm: Fix for modems with brk in modem status control Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 30/82] tty: Set correct tty name in active sysfs attribute Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 31/82] drm/radeon: fix UVD IRQ support on 7xx Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 33/82] drm/i915: Pair va_copy with va_end in i915_error_vprintf Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 37/82] staging: r8188eu: Fix typo in USB_DEVICE list Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 38/82] staging: comedi: adv_pci1710: fix analog output readback value Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 39/82] staging:iio:ad799x fix error_free_irq which was freeing an irq that may not have been requested Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 40/82] iio: max1363: Use devm_regulator_get_optional for optional regulator Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 41/82] iio: adis16400: Set timestamp as the last element in chan_spec Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 42/82] iio: ak8975: Fix calculation formula for convert micro tesla to gauss unit Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 43/82] x86, smap: Dont enable SMAP if CONFIG_X86_SMAP is disabled Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 44/82] x86, smap: smap_violation() is bogus if CONFIG_X86_SMAP is off Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 45/82] ftrace/x86: Use breakpoints for converting function graph caller Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 46/82] ALSA: hda - Fix mic capture on Sony VAIO Pro 11 Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 47/82] mei: clear write cb from waiting list on reset Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 48/82] mei: dont unset read cb ptr " Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 49/82] VME: Correct read/write alignment algorithm Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 50/82] Drivers: hv: vmbus: Dont timeout during the initial connection with host Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 52/82] USB: ftdi_sio: add Tagsys RFID Reader IDs Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 53/82] usb-storage: add unusual-devs entry for BlackBerry 9000 Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 54/82] usb-storage: restrict bcdDevice range for Super Top in Cypress ATACB Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 55/82] usb-storage: enable multi-LUN scanning when needed Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 56/82] usb: option: blacklist ZTE MF667 net interface Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 59/82] Revert "xhci: Set scatter-gather limit to avoid failed block writes." Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 60/82] Revert "xhci: Avoid infinite loop when sg urb requires too many trbs" Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 61/82] Revert "usb: xhci: Link TRB must not occur within a USB payload burst" Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 62/82] Revert "usbcore: set lpm_capable field for LPM capable root hubs" Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 63/82] Modpost: fixed USB alias generation for ranges including 0x9 and 0xA Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 64/82] block: __elv_next_request() shouldnt call into the elevator if bypassing Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 65/82] block: Fix nr_vecs for inline integrity vectors Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 66/82] block: add cond_resched() to potentially long running ioctl discard loop Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 67/82] ACPI / hotplug / PCI: Relax the checking of _STA return values Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 68/82] compiler/gcc4: Make quirk for asm_volatile_goto() unconditional Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 69/82] IB/qib: Add missing serdes init sequence Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 70/82] KVM: return an error code in kvm_vm_ioctl_register_coalesced_mmio() Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 71/82] tick: Clear broadcast pending bit when switching to oneshot Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 72/82] md/raid1: restore ability for check and repair to fix read errors Greg Kroah-Hartman
2014-02-20 23:52 ` Greg Kroah-Hartman [this message]
2014-02-20 23:52 ` [PATCH 3.12 74/82] i2c: mv64xxx: refactor message start to ensure proper initialization Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 75/82] time: Fix overflow when HZ is smaller than 60 Greg Kroah-Hartman
2014-02-20 23:52 ` [PATCH 3.12 76/82] power: max17040: Fix NULL pointer dereference when there is no platform_data Greg Kroah-Hartman
2014-02-20 23:53 ` [PATCH 3.12 77/82] ring-buffer: Fix first commit on sub-buffer having non-zero delta Greg Kroah-Hartman
2014-02-20 23:53 ` [PATCH 3.12 78/82] target: Fix free-after-use regression in PR unregister Greg Kroah-Hartman
2014-02-20 23:53 ` [PATCH 3.12 79/82] genirq: Add missing irq_to_desc export for CONFIG_SPARSE_IRQ=n Greg Kroah-Hartman
2014-02-20 23:53 ` [PATCH 3.12 80/82] drivers/edac/edac_mc_sysfs.c: poll timeout cannot be zero Greg Kroah-Hartman
2014-02-20 23:53 ` [PATCH 3.12 81/82] EDAC: Poll timeout cannot be zero, p2 Greg Kroah-Hartman
2014-02-20 23:53 ` [PATCH 3.12 82/82] EDAC: Correct workqueue setup path Greg Kroah-Hartman
2014-02-21 5:04 ` [PATCH 3.12 00/82] 3.12.13-stable review Guenter Roeck
2014-02-21 23:41 ` Shuah Khan
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=20140220235021.672984378@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-raid@vger.kernel.org \
--cc=neilb@suse.de \
--cc=oleg@redhat.com \
--cc=srivatsa.bhat@linux.vnet.ibm.com \
--cc=stable@vger.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;
as well as URLs for NNTP newsgroup(s).