Skip to content

L
linux
Switch branch/tag
  1. 24 Sep, 2014 1 commit
    • Wanpeng Li's avatar
      sched: Fix unreleased llc_shared_mask bit during CPU hotplug · 03bd4e1f
      Wanpeng Li authored 
      The following bug can be triggered by hot adding and removing a large number of
xen domain0's vcpus repeatedly:

	BUG: unable to handle kernel NULL pointer dereference at 0000000000000004 IP: [..] find_busiest_group
	PGD 5a9d5067 PUD 13067 PMD 0
	Oops: 0000 [#3] SMP
	[...]
	Call Trace:
	load_balance
	? _raw_spin_unlock_irqrestore
	idle_balance
	__schedule
	schedule
	schedule_timeout
	? lock_timer_base
	schedule_timeout_uninterruptible
	msleep
	lock_device_hotplug_sysfs
	online_store
	dev_attr_store
	sysfs_write_file
	vfs_write
	SyS_write
	system_call_fastpath

Last level cache shared mask is built during CPU up and the
build_sched_domain() routine takes advantage of it to setup
the sched domain CPU topology.

However, llc_shared_mask is not released during CPU disable,
which leads to an invalid sched domainCPU topology.

This patch fix it by releasing the llc_shared_mask correctly
during CPU disable.

Yasuaki also reported that this can happen on real hardware:

  https://lkml.org/lkml/2014/7/22/1018

His case is here:

	==
	Here is an example on my system.
	My system has 4 sockets and each socket has 15 cores and HT is
	enabled. In this case, each core of sockes is numbered as
	follows:

		 | CPU#
	Socket#0 | 0-14 , 60-74
	Socket#1 | 15-29, 75-89
	Socket#2 | 30-44, 90-104
	Socket#3 | 45-59, 105-119

	Then llc_shared_mask of CPU#30 has 0x3fff80000001fffc0000000.

	It means that last level cache of Socket#2 is shared with
	CPU#30-44 and 90-104.

	When hot-removing socket#2 and #3, each core of sockets is
	numbered as follows:

		 | CPU#
	Socket#0 | 0-14 , 60-74
	Socket#1 | 15-29, 75-89

	But llc_shared_mask is not cleared. So llc_shared_mask of CPU#30
	remains having 0x3fff80000001fffc0000000.

	After that, when hot-adding socket#2 and #3, each core of
	sockets is numbered as follows:

		 | CPU#
	Socket#0 | 0-14 , 60-74
	Socket#1 | 15-29, 75-89
	Socket#2 | 30-59
	Socket#3 | 90-119

	Then llc_shared_mask of CPU#30 becomes
	0x3fff8000fffffffc0000000. It means that last level cache of
	Socket#2 is shared with CPU#30-59 and 90-104. So the mask has
	the wrong value.
Signed-off-by: default avatarWanpeng Li <wanpeng.li@linux.intel.com>
Tested-by: default avatarLinn Crosetto <linn@hp.com>
Reviewed-by: default avatarBorislav Petkov <bp@suse.de>
Reviewed-by: default avatarToshi Kani <toshi.kani@hp.com>
Reviewed-by: default avatarYasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: <stable@vger.kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Steven Rostedt <srostedt@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1411547885-48165-1-git-send-email-wanpeng.li@linux.intel.comSigned-off-by: default avatarIngo Molnar <mingo@kernel.org>
      03bd4e1f
  3. 20 Sep, 2014 1 commit
  5. 19 Sep, 2014 11 commits
  7. 18 Sep, 2014 8 commits
  9. 17 Sep, 2014 2 commits
    • Paul Gortmaker's avatar
      Revert "init: make rootdelay=N consistent with rootwait behaviour" · 8ba4caf1
      Paul Gortmaker authored 
      This reverts commit 4dfe694f.

In that, we did:

  Here we move the rootdelay code to be right beside the rootwait code, so
  that their behaviour is consistent.

...which is fine, but in hindsight, perhaps moving the rootwait to be
beside the rootdelay would have been better.  We also indicated:

  It should be noted that in doing so, the actions based on the
  saved_root_name[0] and initrd_load() were previously put on hold by
  rootdelay=N and now currently will not be delayed.  However, I think
  consistent behaviour is more important than matching historical behaviour
  of delaying the above two operations.

But Pavel reported an instance where an ARM target with root on MMC
was failing to mount root, and Russell diagnosed it to the fact that
the call to set ROOT_DEV within the saved_root_name[0] processing
block mentioned above was no longer being delayed.

Rather than moving both wait clauses to the original position of
rootdelay and risking unearthing other possible corner case breakage
at this point in time, we simply revert now and we can revisit
trying the alternate/earlier location in another development cycle.

Cc: Pavel Machek <pavel@denx.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarPaul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      8ba4caf1
    • Markos Chandras's avatar
      MIPS: SmartMIPS: Disable assembler warnings · dab1b445
      Markos Chandras authored 
      The kernel code overrides the default ISA as passed by the compiler
in quite a few places. This has unfortunate side effects when smartmips
is enabled leading to hundreds of warnings during build such as:

{standard input}: Assembler messages:
{standard input}:411: Warning: the `smartmips' extension requires MIPS32
revision 1 or greater
{standard input}: Assembler messages:
{standard input}:43: Warning: the 64-bit MIPS architecture does not support the
`smartmips' extension
[...]

Until the kernel code is fixed properly (if possible), disable all the
assembler warning messages to make the build logs readable again.
This has no runtime side effects but it makes it easier to spot
more critical warnings and problems during build.
Signed-off-by: default avatarMarkos Chandras <markos.chandras@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/7356/Signed-off-by: default avatarRalf Baechle <ralf@linux-mips.org>
      dab1b445
  11. 16 Sep, 2014 13 commits
  13. 15 Sep, 2014 4 commits
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7