Skip to content

L
linux
Switch branch/tag
  1. 16 Feb, 2012 1 commit
    • Linus Torvalds's avatar
      i387: fix x86-64 preemption-unsafe user stack save/restore · 15d8791c
      Linus Torvalds authored 
      Commit 5b1cbac3 ("i387: make irq_fpu_usable() tests more robust")
added a sanity check to the #NM handler to verify that we never cause
the "Device Not Available" exception in kernel mode.

However, that check actually pinpointed a (fundamental) race where we do
cause that exception as part of the signal stack FPU state save/restore
code.

Because we use the floating point instructions themselves to save and
restore state directly from user mode, we cannot do that atomically with
testing the TS_USEDFPU bit: the user mode access itself may cause a page
fault, which causes a task switch, which saves and restores the FP/MMX
state from the kernel buffers.

This kind of "recursive" FP state save is fine per se, but it means that
when the signal stack save/restore gets restarted, it will now take the
'#NM' exception we originally tried to avoid.  With preemption this can
happen even without the page fault - but because of the user access, we
cannot just disable preemption around the save/restore instruction.

There are various ways to solve this, including using the
"enable/disable_page_fault()" helpers to not allow page faults at all
during the sequence, and fall back to copying things by hand without the
use of the native FP state save/restore instructions.

However, the simplest thing to do is to just allow the #NM from kernel
space, but fix the race in setting and clearing CR0.TS that this all
exposed: the TS bit changes and the TS_USEDFPU bit absolutely have to be
atomic wrt scheduling, so while the actual state save/restore can be
interrupted and restarted, the act of actually clearing/setting CR0.TS
and the TS_USEDFPU bit together must not.

Instead of just adding random "preempt_disable/enable()" calls to what
is already excessively ugly code, this introduces some helper functions
that mostly mirror the "kernel_fpu_begin/end()" functionality, just for
the user state instead.

Those helper functions should probably eventually replace the other
ad-hoc CR0.TS and TS_USEDFPU tests too, but I'll need to think about it
some more: the task switching functionality in particular needs to
expose the difference between the 'prev' and 'next' threads, while the
new helper functions intentionally were written to only work with
'current'.
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      15d8791c
  3. 15 Feb, 2012 1 commit
    • Linus Torvalds's avatar
      i387: fix sense of sanity check · c38e2345
      Linus Torvalds authored 
      The check for save_init_fpu() (introduced in commit 5b1cbac3: "i387:
make irq_fpu_usable() tests more robust") was the wrong way around, but
I hadn't noticed, because my "tests" were bogus: the FPU exceptions are
disabled by default, so even doing a divide by zero never actually
triggers this code at all unless you do extra work to enable them.

So if anybody did enable them, they'd get one spurious warning.
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      c38e2345
  5. 14 Feb, 2012 34 commits
  7. 13 Feb, 2012 4 commits
    • Linus Torvalds's avatar
      Merge tag 'battery-fixes-for-v3.3-rc2' of git://git.infradead.org/users/cbou/battery-urgent · 3b582f39
      Linus Torvalds authored 
      Just a few small fixes for a bunch of drivers. Nothing noteworthy.

* tag 'battery-fixes-for-v3.3-rc2' of git://git.infradead.org/users/cbou/battery-urgent:
  lp8727_charger: Add terminating entry for i2c_device_id table
  power_supply: Fix modalias for charger-manager
  lp8727_chager: Fix permissions on a header file
  bq27x00_battery: Fix flag register read
  Revert "bq27x00_battery: Fix reporting status value for bq27500 battery"
      3b582f39
    • Linus Torvalds's avatar
      Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs · 19be13cf
      Linus Torvalds authored 
      Two bugfixes in XFS for 3.3: one fix passes KMEM_SLEEP to kmem_realloc
instead of 0, and the other resolves a possible deadlock in xfs quotas.

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: use a normal shrinker for the dquot freelist
  xfs: pass KM_SLEEP flag to kmem_realloc() in xlog_recover_add_to_cnt_trans()
      19be13cf
    • Linus Torvalds's avatar
      Merge branch 'omap-fixes-warnings' of git://git.linaro.org/people/rmk/linux-arm · b14a2998
      Linus Torvalds authored 
      This set of changes are fixing various section mismatch warnings which
look to be completely valid.  Primerily, those which are fixed are those
which can cause oopses by manipulation of driver binding via sysfs.  For
example: calling code marked __init from driver probe __devinit
functions.

Some of these changes will be reworked at the next merge window when the
underlying reasons are sorted out.  In the mean time, I think it's
important to have this fixed for correctness.

Also included in this set are fixes to various error messages in OMAP -
including making them gramatically correct, fixing a few spelling
errors, and more importantly, making them greppable by unwrapping them.

Tony Lindgren has acked all these patches, put them out for testing a
week ago, and I've tested them on the platforms I have.

* 'omap-fixes-warnings' of git://git.linaro.org/people/rmk/linux-arm:
  ARM: omap: resolve nebulous 'Error setting wl12xx data'
  ARM: omap: fix wrapped error messages in omap_hwmod.c
  ARM: omap: fix section mismatch warnings in mux.c caused by hsmmc.c
  ARM: omap: fix section mismatch warning for sdp3430_twl_gpio_setup()
  ARM: omap: fix section mismatch error for omap_4430sdp_display_init()
  ARM: omap: fix section mismatch warning for omap_secondary_startup()
  ARM: omap: preemptively fix section mismatch in omap4_sdp4430_wifi_mux_init()
  ARM: omap: fix section mismatch warning in mux.c
  ARM: omap: fix section mismatch errors in TWL PMIC driver
  ARM: omap: fix uninformative vc/i2c configuration error message
  ARM: omap: fix vc.c PMIC error message
  ARM: omap: fix prm44xx.c OMAP44XX_IRQ_PRCM build error
      b14a2998
    • Linus Torvalds's avatar
      Merge branch 'omap-fixes-urgent' of git://git.linaro.org/people/rmk/linux-arm · a269c2f5
      Linus Torvalds authored 
      This pull request covers the major oopsing issues with OMAP, caused by
the lack of the TWL driver.  Even when the TWL driver is not built in,
we shouldn't oops.

* 'omap-fixes-urgent' of git://git.linaro.org/people/rmk/linux-arm:
  ARM: omap: fix broken twl-core dependencies and ifdefs
  ARM: omap: fix oops in drivers/video/omap2/dss/dpi.c
  ARM: omap: fix oops in arch/arm/mach-omap2/vp.c when pmic is not found
      a269c2f5
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7