stable.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg KH <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: torvalds@linux-foundation.org, akpm@linux-foundation.org,
	alan@lxorguk.ukuu.org.uk, Tejun Heo <tj@kernel.org>,
	"Pavel V. Panteleev" <pp_84@mail.ru>
Subject: [ 18/47] percpu: pcpu_embed_first_chunk() should free unused parts after all allocs are complete
Date: Fri, 18 May 2012 14:27:07 -0700	[thread overview]
Message-ID: <20120518212650.828249272@linuxfoundation.org> (raw)
In-Reply-To: <20120518212701.GA5023@kroah.com>

3.3-stable review patch.  If anyone has any objections, please let me know.

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

From: Tejun Heo <tj@kernel.org>

commit 42b64281453249dac52861f9b97d18552a7ec62b upstream.

pcpu_embed_first_chunk() allocates memory for each node, copies percpu
data and frees unused portions of it before proceeding to the next
group.  This assumes that allocations for different nodes doesn't
overlap; however, depending on memory topology, the bootmem allocator
may end up allocating memory from a different node than the requested
one which may overlap with the portion freed from one of the previous
percpu areas.  This leads to percpu groups for different nodes
overlapping which is a serious bug.

This patch separates out copy & partial free from the allocation loop
such that all allocations are complete before partial frees happen.

This also fixes overlapping frees which could happen on allocation
failure path - out_free_areas path frees whole groups but the groups
could have portions freed at that point.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: "Pavel V. Panteleev" <pp_84@mail.ru>
Tested-by: "Pavel V. Panteleev" <pp_84@mail.ru>
LKML-Reference: <E1SNhwY-0007ui-V7.pp_84-mail-ru@f220.mail.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 mm/percpu.c |   10 ++++++++++
 1 file changed, 10 insertions(+)

--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -1650,6 +1650,16 @@ int __init pcpu_embed_first_chunk(size_t
 		areas[group] = ptr;
 
 		base = min(ptr, base);
+	}
+
+	/*
+	 * Copy data and free unused parts.  This should happen after all
+	 * allocations are complete; otherwise, we may end up with
+	 * overlapping groups.
+	 */
+	for (group = 0; group < ai->nr_groups; group++) {
+		struct pcpu_group_info *gi = &ai->groups[group];
+		void *ptr = areas[group];
 
 		for (i = 0; i < gi->nr_units; i++, ptr += ai->unit_size) {
 			if (gi->cpu_map[i] == NR_CPUS) {



  parent reply	other threads:[~2012-05-18 21:27 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-05-18 21:27 [ 00/47] 3.3.7-stable review Greg KH
2012-05-18 21:26 ` [ 01/47] ALSA: echoaudio: Remove incorrect part of assertion Greg KH
2012-05-18 21:26 ` [ 02/47] ALSA: HDA: Lessen CPU usage when waiting for chip to respond Greg KH
2012-05-18 21:26 ` [ 03/47] ALSA: hda/realtek - Add missing CD-input pin for MSI-7350 mobo Greg KH
2012-05-18 21:26 ` [ 04/47] ALSA: hda/idt - Fix power-map for speaker-pins with some HP laptops Greg KH
2012-05-18 21:26 ` [ 05/47] usbnet: fix skb traversing races during unlink(v2) Greg KH
2012-05-18 21:26 ` [ 06/47] namespaces, pid_ns: fix leakage on fork() failure Greg KH
2012-05-18 21:26 ` [ 07/47] sparc64: Do not clobber %g2 in xcall_fetch_glob_regs() Greg KH
2012-05-18 21:26 ` [ 08/47] media: marvell-cam: fix an ARM build error Greg KH
2012-05-18 21:26 ` [ 09/47] ARM: 7417/1: vfp: ensure preemption is disabled when enabling VFP access Greg KH
2012-05-18 21:26 ` [ 10/47] ARM: prevent VM_GROWSDOWN mmaps extending below FIRST_USER_ADDRESS Greg KH
2012-05-18 21:27 ` [ 11/47] media: s5p-fimc: Fix locking in subdev set_crop op Greg KH
2012-05-18 21:27 ` [ 12/47] media: rc: Postpone ISR registration Greg KH
2012-05-18 21:27 ` [ 13/47] media: dvb_frontend: fix a regression with DVB-S zig-zag Greg KH
2012-05-18 21:27 ` [ 14/47] ASoC: cs42l73: Sync digital mixer kcontrols to allow for 0dB Greg KH
2012-05-18 21:27 ` [ 15/47] ASoC: wm8994: Fix AIF2ADC power down Greg KH
2012-05-18 21:27 ` [ 16/47] cdc_ether: Ignore bogus union descriptor for RNDIS devices Greg KH
2012-05-18 21:27 ` [ 17/47] cdc_ether: add Novatel USB551L device IDs for FLAG_WWAN Greg KH
2012-05-18 21:27 ` Greg KH [this message]
2012-05-18 21:27 ` [ 19/47] kmemleak: Fix the kmemleak tracking of the percpu areas with !SMP Greg KH
2012-05-18 21:27 ` [ 20/47] mtd: fix oops in dataflash driver Greg KH
2012-05-18 21:27 ` [ 21/47] hugetlb: prevent BUG_ON in hugetlb_fault() -> hugetlb_cow() Greg KH
2012-05-18 21:27 ` [ 22/47] mm: nobootmem: fix sign extend problem in __free_pages_memory() Greg KH
2012-05-18 21:27 ` [ 23/47] jffs2: Fix lock acquisition order bug in gc path Greg KH
2012-05-18 21:27 ` [ 24/47] arch/tile: apply commit 74fca9da0 to the compat signal handling as well Greg KH
2012-05-18 21:27 ` [ 25/47] crypto: mv_cesa requires on CRYPTO_HASH to build Greg KH
2012-05-18 21:27 ` [ 26/47] target: Drop incorrect se_lun_acl release for dynamic -> explict ACL conversion Greg KH
2012-05-18 21:27 ` [ 27/47] target: Fix SPC-2 RELEASE bug for multi-session iSCSI client setups Greg KH
2012-05-18 21:27 ` [ 28/47] target: Fix bug in handling of FILEIO + block_device resize ops Greg KH
2012-05-18 21:27 ` [ 29/47] virtio: console: tell host of open ports after resume from s3/s4 Greg KH
2012-05-18 21:27 ` [ 30/47] dm mpath: check if scsi_dh module already loaded before trying to load Greg KH
2012-05-18 21:27 ` [ 31/47] e1000: Prevent reset task killing itself Greg KH
2012-05-18 21:27 ` [ 32/47] MD: Add del_timer_sync to mddev_suspend (fix nasty panic) Greg KH
2012-05-18 21:27 ` [ 33/47] tcp: do_tcp_sendpages() must try to push data out on oom conditions Greg KH
2012-05-18 21:27 ` [ 34/47] init: dont try mounting device as nfs root unless type fully matches Greg KH
2012-05-18 21:27 ` [ 35/47] ext4: avoid deadlock on sync-mounted FS w/o journal Greg KH
2012-05-18 21:27 ` [ 36/47] memcg: free spare array to avoid memory leak Greg KH
2012-05-18 21:27 ` [ 37/47] cifs: fix revalidation test in cifs_llseek() Greg KH
2012-05-18 21:27 ` [ 38/47] compat: Fix RT signal mask corruption via sigprocmask Greg KH
2012-05-18 21:27 ` [ 39/47] dl2k: Clean up rio_ioctl Greg KH
2012-05-18 21:27 ` [ 40/47] OMAPDSS: VENC: fix NULL pointer dereference in DSS2 VENC sysfs debug attr on OMAP4 Greg KH
2012-05-18 21:27 ` [ 41/47] i2c-eg20t: change timeout value 50msec to 1000msec Greg KH
2012-05-18 21:27 ` [ 42/47] spi-topcliff-pch: Modify pci-bus number dynamically to get DMA device info Greg KH
2012-05-18 21:27 ` [ 43/47] spi-topcliff-pch: Fix issue for transmitting over 4KByte Greg KH
2012-05-18 21:27 ` [ 44/47] spi-topcliff-pch: supports a spi mode setup and bit order setup by IO control Greg KH
2012-05-18 21:27 ` [ 45/47] spi-topcliff-pch: add recovery processing in case wait-event timeout Greg KH
2012-05-18 21:27 ` [ 46/47] Avoid beyond bounds copy while caching ACL Greg KH
2012-05-18 21:27 ` [ 47/47] Avoid reading past buffer when calling GETACL Greg KH

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=20120518212650.828249272@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=akpm@linux-foundation.org \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pp_84@mail.ru \
    --cc=stable@vger.kernel.org \
    --cc=tj@kernel.org \
    --cc=torvalds@linux-foundation.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).