kernel/linux.git/drivers, branch v4.10.13

device-dax: switch to srcu, fix rcu_read_lock() vs pte allocation

2017-04-27T07:12:59+00:00

commit 956a4cd2c957acf638ff29951aabaa9d8e92bbc2 upstream. The following warning triggers with a new unit test that stresses the device-dax interface. =============================== [ ERR: suspicious RCU usage. ] 4.11.0-rc4+ #1049 Tainted: G O ------------------------------- ./include/linux/rcupdate.h:521 Illegal context switch in RCU read-side critical section! other info that might help us debug this: rcu_scheduler_active = 2, debug_locks = 0 2 locks held by fio/9070: #0: (&mm->mmap_sem){++++++}, at: [] __do_page_fault+0x167/0x4f0 #1: (rcu_read_lock){......}, at: [] dax_dev_huge_fault+0x32/0x620 [dax] Call Trace: dump_stack+0x86/0xc3 lockdep_rcu_suspicious+0xd7/0x110 ___might_sleep+0xac/0x250 __might_sleep+0x4a/0x80 __alloc_pages_nodemask+0x23a/0x360 alloc_pages_current+0xa1/0x1f0 pte_alloc_one+0x17/0x80 __pte_alloc+0x1e/0x120 __get_locked_pte+0x1bf/0x1d0 insert_pfn.isra.70+0x3a/0x100 ? lookup_memtype+0xa6/0xd0 vm_insert_mixed+0x64/0x90 dax_dev_huge_fault+0x520/0x620 [dax] ? dax_dev_huge_fault+0x32/0x620 [dax] dax_dev_fault+0x10/0x20 [dax] __do_fault+0x1e/0x140 __handle_mm_fault+0x9af/0x10d0 handle_mm_fault+0x16d/0x370 ? handle_mm_fault+0x47/0x370 __do_page_fault+0x28c/0x4f0 trace_do_page_fault+0x58/0x2a0 do_async_page_fault+0x1a/0xa0 async_page_fault+0x28/0x30 Inserting a page table entry may trigger an allocation while we are holding a read lock to keep the device instance alive for the duration of the fault. Use srcu for this keep-alive protection. Fixes: dee410792419 ("/dev/dax, core: file operations and dax-mmap") Signed-off-by: Dan Williams Signed-off-by: Greg Kroah-Hartman

ubi/upd: Always flush after prepared for an update

2017-04-27T07:12:59+00:00

commit 9cd9a21ce070be8a918ffd3381468315a7a76ba6 upstream. In commit 6afaf8a484cb ("UBI: flush wl before clearing update marker") I managed to trigger and fix a similar bug. Now here is another version of which I assumed it wouldn't matter back then but it turns out UBI has a check for it and will error out like this: |ubi0 warning: validate_vid_hdr: inconsistent used_ebs |ubi0 error: validate_vid_hdr: inconsistent VID header at PEB 592 All you need to trigger this is? "ubiupdatevol /dev/ubi0_0 file" + a powercut in the middle of the operation. ubi_start_update() sets the update-marker and puts all EBs on the erase list. After that userland can proceed to write new data while the old EB aren't erased completely. A powercut at this point is usually not that much of a tragedy. UBI won't give read access to the static volume because it has the update marker. It will most likely set the corrupted flag because it misses some EBs. So we are all good. Unless the size of the image that has been written differs from the old image in the magnitude of at least one EB. In that case UBI will find two different values for `used_ebs' and refuse to attach the image with the error message mentioned above. So in order not to get in the situation, the patch will ensure that we wait until everything is removed before it tries to write any data. The alternative would be to detect such a case and remove all EBs at the attached time after we processed the volume-table and see the update-marker set. The patch looks bigger and I doubt it is worth it since usually the write() will wait from time to time for a new EB since usually there not that many spare EB that can be used. Signed-off-by: Sebastian Andrzej Siewior Signed-off-by: Richard Weinberger Signed-off-by: Greg Kroah-Hartman

mmc: sdhci-esdhc-imx: increase the pad I/O drive strength for DDR50 card

2017-04-27T07:12:58+00:00

commit 9f327845358d3dd0d8a5a7a5436b0aa5c432e757 upstream. Currently for DDR50 card, it need tuning in default. We meet tuning fail issue for DDR50 card and some data CRC error when DDR50 sd card works. This is because the default pad I/O drive strength can't make sure DDR50 card work stable. So increase the pad I/O drive strength for DDR50 card, and use pins_100mhz. This fixes DDR50 card support for IMX since DDR50 tuning was enabled from commit 9faac7b95ea4 ("mmc: sdhci: enable tuning for DDR50") Tested-and-reported-by: Tim Harvey Signed-off-by: Haibo Chen Acked-by: Dong Aisheng Acked-by: Adrian Hunter Signed-off-by: Ulf Hansson Signed-off-by: Greg Kroah-Hartman

mmc: dw_mmc: Don't allow Runtime PM for SDIO cards

2017-04-27T07:12:58+00:00

commit a6db2c86033bc41329770e90c20d4f1fec3824e4 upstream. According to the SDIO standard interrupts are normally signalled in a very complicated way. They require the card clock to be running and require the controller to be paying close attention to the signals coming from the card. This simply can't happen with the clock stopped or with the controller in a low power mode. To that end, we'll disable runtime_pm when we detect that an SDIO card was inserted. This is much like with what we do with the special "SDMMC_CLKEN_LOW_PWR" bit that dw_mmc supports. NOTE: we specifically do this Runtime PM disabling at card init time rather than in the enable_sdio_irq() callback. This is _different_ than how SDHCI does it. Why do we do it differently? - Unlike SDHCI, dw_mmc uses the standard sdio_irq code in Linux (AKA dw_mmc doesn't set MMC_CAP2_SDIO_IRQ_NOTHREAD). - Because we use the standard sdio_irq code: - We see a constant stream of enable_sdio_irq(0) and enable_sdio_irq(1) calls. This is because the standard code disables interrupts while processing and re-enables them after. - While interrupts are disabled, there's technically a period where we could get runtime disabled while processing interrupts. - If we are runtime disabled while processing interrupts, we'll reset the controller at resume time (see dw_mci_runtime_resume), which seems like a terrible idea because we could possibly have another interrupt pending. To fix the above isues we'd want to put something in the standard sdio_irq code that makes sure to call pm_runtime get/put when interrupts are being actively being processed. That's possible to do, but it seems like a more complicated mechanism when we really just want the runtime pm disabled always for SDIO cards given that all the other bits needed to get Runtime PM vs. SDIO just aren't there. NOTE: at some point in time someone might come up with a fancy way to do SDIO interrupts and still allow (some) amount of runtime PM. Technically we could turn off the card clock if we used an alternate way of signaling SDIO interrupts (and out of band interrupt is one way to do this). We probably wouldn't actually want to fully runtime suspend in this case though--at least not with the current dw_mci_runtime_resume() which basically fully resets the controller at resume time. Fixes: e9ed8835e990 ("mmc: dw_mmc: add runtime PM callback") Reported-by: Brian Norris Signed-off-by: Douglas Anderson Acked-by: Jaehoon Chung Reviewed-by: Shawn Lin Signed-off-by: Ulf Hansson Signed-off-by: Greg Kroah-Hartman

ACPI / power: Avoid maybe-uninitialized warning

2017-04-27T07:12:58+00:00

commit fe8c470ab87d90e4b5115902dd94eced7e3305c3 upstream. gcc -O2 cannot always prove that the loop in acpi_power_get_inferred_state() is enterered at least once, so it assumes that cur_state might not get initialized: drivers/acpi/power.c: In function 'acpi_power_get_inferred_state': drivers/acpi/power.c:222:9: error: 'cur_state' may be used uninitialized in this function [-Werror=maybe-uninitialized] This sets the variable to zero at the start of the loop, to ensure that there is well-defined behavior even for an empty list. This gets rid of the warning. The warning first showed up when the -Os flag got removed in a bug fix patch in linux-4.11-rc5. I would suggest merging this addon patch on top of that bug fix to avoid introducing a new warning in the stable kernels. Fixes: 61b79e16c68d (ACPI: Fix incompatibility with mcount-based function graph tracing) Signed-off-by: Arnd Bergmann Signed-off-by: Rafael J. Wysocki Signed-off-by: Greg Kroah-Hartman

Input: elantech - add Fujitsu Lifebook E547 to force crc_enabled

2017-04-27T07:12:58+00:00

commit 704de489e0e3640a2ee2d0daf173e9f7375582ba upstream. Temporary got a Lifebook E547 into my hands and noticed the touchpad only works after running: echo "1" > /sys/devices/platform/i8042/serio2/crc_enabled Add it to the list of machines that need this workaround. Signed-off-by: Thorsten Leemhuis Reviewed-by: Ulrik De Bie Signed-off-by: Dmitry Torokhov Signed-off-by: Greg Kroah-Hartman

mmc: dw_mmc: silent verbose log when calling from PM context

2017-04-27T07:12:57+00:00

commit ce69e2fea093b7fa3991c87849c4955cd47796c9 upstream. When deploying runtime PM, it's quite verbose to print the log of ios setting. Also it's useless to print it from system PM as it should be the same with booting time. We also have sysfs to get all these information from ios attribute, so let's skip this print from PM context. Signed-off-by: Shawn Lin Signed-off-by: Jaehoon Chung Signed-off-by: Ulf Hansson Cc: Alexander Kochetkov Signed-off-by: Greg Kroah-Hartman

HID: wacom: Treat HID_DG_TOOLSERIALNUMBER as unsigned

2017-04-27T07:12:57+00:00

commit 286f3f478796fb4f9e003e9f7d649f3c33f08d2f upstream. Because HID_DG_TOOLSERIALNUMBER doesn't first cast the value recieved from HID to an unsigned type, sign-extension rules can cause the value of wacom_wac->serial[0] to inadvertently wind up with all 32 of its highest bits set if the highest bit of "value" was set. This can cause problems for Tablet PC devices which use AES sensors and the xf86-input-wacom userspace driver. It is not uncommon for AES sensors to send a serial number of '0' while the pen is entering or leaving proximity. The xf86-input-wacom driver ignores events with a serial number of '0' since it cannot match them up to an in-use tool. To ensure the xf86-input-wacom driver does not ignore the final out-of-proximity event, the kernel does not send MSC_SERIAL events when the value of wacom_wac->serial[0] is '0'. If the highest bit of HID_DG_TOOLSERIALNUMBER is set by an in-prox pen which later leaves proximity and sends a '0' for HID_DG_TOOLSERIALNUMBER, then only the lowest 32 bits of wacom_wac->serial[0] are actually cleared, causing the kernel to send an MSC_SERIAL event. Since the 'input_event' function takes an 'int' as argument, only those lowest (now-cleared) 32 bits of wacom_wac->serial[0] are sent to userspace, causing xf86-input-wacom to ignore the event. If the event was the final out-of-prox event, then xf86-input-wacom may remain in a state where it believes the pen is in proximity and refuses to allow other devices under its control (e.g. the touchscreen) to move the cursor. It should be noted that EMR devices and devices which use both the HID_DG_TOOLSERIALNUMBER and WACOM_HID_WD_SERIALHI usages (in that order) would be immune to this issue. It appears only AES devices are affected. Fixes: f85c9dc678a ("HID: wacom: generic: Support tool ID and additional tool types") Signed-off-by: Jason Gerecke Acked-by: Benjamin Tissoires Signed-off-by: Jiri Kosina Signed-off-by: Greg Kroah-Hartman

virtio-console: avoid DMA from stack

2017-04-21T07:32:42+00:00

commit c4baad50297d84bde1a7ad45e50c73adae4a2192 upstream. put_chars() stuffs the buffer it gets into an sg, but that buffer may be on the stack. This breaks with CONFIG_VMAP_STACK=y (for me, it manifested as printks getting turned into NUL bytes). Signed-off-by: Omar Sandoval Signed-off-by: Michael S. Tsirkin Reviewed-by: Amit Shah Cc: Ben Hutchings Cc: Brad Spengler Signed-off-by: Greg Kroah-Hartman

cxusb: Use a dma capable buffer also for reading

2017-04-21T07:32:42+00:00

commit 3f190e3aec212fc8c61e202c51400afa7384d4bc upstream. Commit 17ce039b4e54 ("[media] cxusb: don't do DMA on stack") added a kmalloc'ed bounce buffer for writes, but missed to do the same for reads. As the read only happens after the write is finished, we can reuse the same buffer. As dvb_usb_generic_rw handles a read length of 0 by itself, avoid calling it using the dvb_usb_generic_read wrapper function. Signed-off-by: Stefan Brüns Signed-off-by: Mauro Carvalho Chehab Cc: Ben Hutchings Cc: Brad Spengler Signed-off-by: Greg Kroah-Hartman