commit d8204a37baf5474d3154eb536c936369be2bd5c0 Merge: 571020d 33619f0 Author: Linus Torvalds Date: Tue Feb 22 09:26:54 2011 -0800 Merge branch 'urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6 * 'urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6: pcmcia: re-enable Zoomed Video support cm4000_cs: Fix undefined ops warning pcmcia vs. MECR on pxa25x/sa1111 drivers/char/pcmcia/ipwireless/main.c: Convert release_resource to release_region/release_mem_region commit 571020df6fae18623498d70d3f2331d9eb7ad2be Merge: ee88347 6d83f94 Author: Linus Torvalds Date: Tue Feb 22 09:26:17 2011 -0800 Merge branch 'irq-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip * 'irq-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: genirq: Disable the SHIRQ_DEBUG call in request_threaded_irq for now genirq: Prevent access beyond allocated_irqs bitmap commit ee88347755460e8ceb37e8453adcf5fd7ca5ff00 Merge: 609b06f 4fe757d Author: Linus Torvalds Date: Tue Feb 22 09:25:55 2011 -0800 Merge branch 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip * 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: perf: Fix throttle logic perf, x86: P4 PMU: Fix spurious NMI messages commit 609b06f33506d92693cb89c303026f830f769c0d Merge: 790e10b 3017358 Author: Linus Torvalds Date: Tue Feb 22 08:20:02 2011 -0800 Merge branch 'fix/asoc' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6 * 'fix/asoc' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6: ASoC: Ensure supplies are maintained for force enabled widgets ASoC: WM8994: Improve playback robustness ASoC: WM8994: Improve robustness in some use cases ASoC: WM8903: Fix mic detection enable logic ASoC: WM8903: Fix mic detection register definitions ASoC: CX20442: fix wrong reg_cache_default content ASoC: Sync initial widget state with hardware commit 790e10ba6c93c123665a3cfa53d6dcf73923e458 Author: Romain Francoise Date: Tue Feb 22 10:48:06 2011 +0100 .gitignore: ignore *.xz files Building with CONFIG_KERNEL_XZ results in the following: # Untracked files: # (use "git add ..." to include in what will be committed) # # arch/x86/boot/compressed/vmlinux.bin.xz So ignore xz-compressed files at the top level like we already do for other compression types. Signed-off-by: Romain Francoise Signed-off-by: Linus Torvalds commit 33619f0d3ff715a2a5499520967d526ad931d70d Author: Dominik Brodowski Date: Sat Feb 19 12:35:15 2011 +0100 pcmcia: re-enable Zoomed Video support Allow drivers to enable Zoomed Video support. Currently, this is only used by out-of-tree drivers (L64020 DVB driver in particular). CC: [for 2.6.37] Signed-off-by: Dominik Brodowski commit 644e6e4a7fa6b11d59f24032997d90ea0d858d03 Author: Alan Cox Date: Tue Feb 1 15:46:05 2011 +0000 cm4000_cs: Fix undefined ops warning Signed-off-by: Alan Cox Signed-off-by: Dominik Brodowski commit 5b703683b6cc3cb97bbe6b1b14898b273eb59279 Author: Pavel Machek Date: Fri Feb 4 09:03:43 2011 +0100 pcmcia vs. MECR on pxa25x/sa1111 After 2.6.34 changes, __pxa2xx_drv_pcmcia_probe() was replaced by sa1111_pcmcia_add(). That unfortunately means that configure_sockets() is not called, leading to MECR not being set properly, leading to strange crashes. Tested on pxa255+sa1111, I do not have lubbock board nearby. Perhaps cleaner solution exists? Signed-off-by: Pavel Machek Signed-off-by: Dominik Brodowski commit d5bb2923cfa0a29c5854f9618703ff60849b949e Author: Julia Lawall Date: Sun Feb 13 13:12:10 2011 +0100 drivers/char/pcmcia/ipwireless/main.c: Convert release_resource to release_region/release_mem_region Request_region should be used with release_region, not release_resource. This patch contains a number of changes, related to calls to request_region, request_mem_region, and the associated error handling code. 1. For the call to request_region, the variable io_resource storing the result is dropped. The call to release_resource at the end of the function is changed to a call to release_region with the first two arguments of request_region as its arguments. The same call to release_region is also added to release_ipwireless. 2. The first call to request_mem_region is now tested and ret is set to -EBUSY if the the call has failed. This call was associated with the initialization of ipw->attr_memory. But the error handling code was testing ipw->common_memory. The definition of release_ipwireless also suggests that this call should be associated with ipw->common_memory, not ipw->attr_memory. 3. The second call to request_mem_region is now tested and ret is set to -EBUSY if the the call has failed. 4. The various gotos to the error handling code is adjusted so that there is no need for ifs. 5. Return the value stored in the ret variable rather than -1. The semantic match that finds this problem is as follows: (http://coccinelle.lip6.fr/) // @@ expression x,E; @@ ( *x = request_region(...) | *x = request_mem_region(...) ) ... when != release_region(x) when != x = E * release_resource(x); // Signed-off-by: Julia Lawall Signed-off-by: Jiri Kosina Signed-off-by: Dominik Brodowski commit 6d83f94db95cfe65d2a6359cccdf61cf087c2598 Author: Thomas Gleixner Date: Fri Feb 18 23:27:23 2011 +0100 genirq: Disable the SHIRQ_DEBUG call in request_threaded_irq for now With CONFIG_SHIRQ_DEBUG=y we call a newly installed interrupt handler in request_threaded_irq(). The original implementation (commit a304e1b8) called the handler _BEFORE_ it was installed, but that caused problems with handlers calling disable_irq_nosync(). See commit 377bf1e4. It's braindead in the first place to call disable_irq_nosync in shared handlers, but .... Moving this call after we installed the handler looks innocent, but it is very subtle broken on SMP. Interrupt handlers rely on the fact, that the irq core prevents reentrancy. Now this debug call violates that promise because we run the handler w/o the IRQ_INPROGRESS protection - which we cannot apply here because that would result in a possibly forever masked interrupt line. A concurrent real hardware interrupt on a different CPU results in handler reentrancy and can lead to complete wreckage, which was unfortunately observed in reality and took a fricking long time to debug. Leave the code here for now. We want this debug feature, but that's not easy to fix. We really should get rid of those disable_irq_nosync() abusers and remove that function completely. Signed-off-by: Thomas Gleixner Cc: Anton Vorontsov Cc: David Woodhouse Cc: Arjan van de Ven Cc: stable@kernel.org # .28 -> .37 commit c1ee6264280e740a9d3ff3feef38642cf0a57013 Author: Thomas Gleixner Date: Thu Feb 17 17:45:15 2011 +0100 genirq: Prevent access beyond allocated_irqs bitmap Lars-Peter Clausen pointed out: I stumbled upon this while looking through the existing archs using SPARSE_IRQ. Even with SPARSE_IRQ the NR_IRQS is still the upper limit for the number of IRQs. Both PXA and MMP set NR_IRQS to IRQ_BOARD_START, with IRQ_BOARD_START being the number of IRQs used by the core. In various machine files the nr_irqs field of the ARM machine defintion struct is then set to "IRQ_BOARD_START + NR_BOARD_IRQS". As a result "nr_irqs" will greater then NR_IRQS which then again causes the "allocated_irqs" bitmap in the core irq code to be accessed beyond its size overwriting unrelated data. The core code really misses a sanity check there. This went unnoticed so far as by chance the compiler/linker places data behind that bitmap which gets initialized later on those affected platforms. So the obvious fix would be to add a sanity check in early_irq_init() and break all affected platforms. Though that check wants to be backported to stable as well, which will require to fix all known problematic platforms and probably some more yet not known ones as well. Lots of churn. A way simpler solution is to allocate a slightly larger bitmap and avoid the whole churn w/o breaking anything. Add a few warnings when an arch returns utter crap. Reported-by: Lars-Peter Clausen Signed-off-by: Thomas Gleixner Cc: stable@kernel.org # .37 Cc: Haojian Zhuang Cc: Eric Miao Cc: Peter Zijlstra commit 4fe757dd48a9e95e1a071291f15dda5421dacb66 Author: Peter Zijlstra Date: Tue Feb 15 22:26:07 2011 +0100 perf: Fix throttle logic It was possible to call pmu::start() on an already running event. In particular this lead so some wreckage as the hrtimer events would re-initialize active timers. This was due to throttled events being activated again by scheduling. Scheduling in a context would add and force start events, resulting in running events with a possible throttle status. The next tick to hit that task will then try to unthrottle the event and call ->start() on an already running event. Reported-by: Jeff Moyer Cc: Signed-off-by: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar commit 7d44ec193d95416d1342cdd86392a1eeb7461186 Author: Cyrill Gorcunov Date: Wed Feb 16 14:08:02 2011 +0300 perf, x86: P4 PMU: Fix spurious NMI messages Several people have reported spurious unknown NMI messages on some P4 CPUs. This patch fixes it by checking for an overflow (negative counter values) directly, instead of relying on the P4_CCCR_OVF bit. Reported-by: George Spelvin Reported-by: Meelis Roos Reported-by: Don Zickus Reported-by: Dave Airlie Signed-off-by: Cyrill Gorcunov Cc: Lin Ming Cc: Don Zickus Cc: Peter Zijlstra LKML-Reference: Signed-off-by: Ingo Molnar commit 3017358a75917b5ed5ad361c02ba2a7e257d3b2a Author: Mark Brown Date: Fri Feb 11 11:42:19 2011 +0000 ASoC: Ensure supplies are maintained for force enabled widgets If a widget has been force enabled then not only do we need to keep the widget itself enabled, we also need to keep any supplies the widget requires enabled. The user could force all the individual widgets on but this requires too much knowledge of device internals. Signed-off-by: Mark Brown Tested-by: Stephen Warren Acked-by: Liam Girdwood commit c52fd021bc027a90a10782c0dcf667ac0135e478 Author: Dimitris Papastamos Date: Fri Feb 11 16:32:12 2011 +0000 ASoC: WM8994: Improve playback robustness On WM8994 revision D and earlier ensure proper playback robustness as some rare use cases can trigger issues. Signed-off-by: Dimitris Papastamos Acked-by: Liam Girdwood Signed-off-by: Mark Brown Cc: stable@kernel.org commit 173efa09e4c807a2a764509ddd593ad13a44d1df Author: Dimitris Papastamos Date: Fri Feb 11 16:32:11 2011 +0000 ASoC: WM8994: Improve robustness in some use cases Ensure that on disabling certain registers such as AIF1DAC1L, AIF1DAC1R etc. the AIF1CLK and AIF2CLK remain enabled. Similarly when enabling those registers, AIF1CLK and AIF2CLK will remain disabled. Signed-off-by: Dimitris Papastamos Acked-by: Liam Girdwood Signed-off-by: Mark Brown Cc: stable@kernel.org commit 3088e3b4963d26d6f6f54987f595b974ed6d48d8 Author: Stephen Warren Date: Thu Feb 10 15:37:14 2011 -0700 ASoC: WM8903: Fix mic detection enable logic The mic detection HW should be enabled when either mic or short detection is required, not when only both are required. Signed-off-by: Stephen Warren Acked-by: Liam Girdwood Signed-off-by: Mark Brown Cc: stable@kernel.org commit 28d639f7bd2559e3d6dc10c0316d19154b28a2a4 Author: Stephen Warren Date: Thu Feb 10 15:37:13 2011 -0700 ASoC: WM8903: Fix mic detection register definitions * There is no hysteresis enable field in the current datasheet. * Mic detection threshold field is only 2 bits wide. Signed-off-by: Stephen Warren Acked-by: Liam Girdwood Signed-off-by: Mark Brown commit 8e6bfb9b1f79e07c18b0ae406c7c678fc54e4d8e Author: Janusz Krzysztofik Date: Thu Feb 10 13:24:32 2011 +0100 ASoC: CX20442: fix wrong reg_cache_default content Content of the CX20442's snd_soc_codec_driver.reg_cache_default pointed area, introduced with my recent NULL pointer dereferece fix (commit f019ee5feb344ff0b22b58df4568676295aae14f), occured wrong after further testing, more thorough than just booting successfully. There are two problems with it: 1) It should read (1 << CX20442_TELOUT) | (1 << CX20442_MIC), not CX20442_TELOUT | CX20442_MIC. 2) While correctly matching actual codec hardware state on boot when fixed per 1), a few more code modifications would still be required to reflect that state not only into register cache, but also force them into DAPM pins state, otherwise an inconsitency occures which may prevent further codec state changes from being applied correctly. As a result, the phone stops ringing after reboot, until someone picks up the handset for the first time. Revert that reg_cache_default content to a working, previous de facto default value of 0, in hope this change can still be accepted as an rc cycle fix. Created and tested against linux-2.6.38-rc4 Signed-off-by: Janusz Krzysztofik Acked-by: Liam Girdwood Signed-off-by: Mark Brown commit b66a70d5e9929f3b0df5a7177bba75652d2f4c3e Author: Mark Brown Date: Wed Feb 9 18:04:11 2011 +0000 ASoC: Sync initial widget state with hardware ASoC generally uses the register defaults for everything, but in some cases the hardware will default to enabling some of the DAPM widgets (clocks for example). Ensure that DAPM knows about the actual widget state at initialisation by reading the enable bits after instantiating the widgets so they don't get left enabled needlessly. Signed-off-by: Mark Brown Acked-by: Liam Girdwood