- 14 May, 2017 27 commits
-
-
Brian Norris authored
commit 6183468a upstream. Similar to commit fcd2042e ("mwifiex: printk() overflow with 32-byte SSIDs"), we failed to account for the existence of 32-char SSIDs in our debugfs code. Unlike in that case though, we zeroed out the containing struct first, and I'm pretty sure we're guaranteed to have some padding after the 'ssid.ssid' and 'ssid.ssid_len' fields (the struct is 33 bytes long). So, this is the difference between: # cat /sys/kernel/debug/mwifiex/mlan0/info ... essid="0123456789abcdef0123456789abcdef " ... and the correct output: # cat /sys/kernel/debug/mwifiex/mlan0/info ... essid="0123456789abcdef0123456789abcdef" ... Fixes: 5e6e3a92 ("wireless: mwifiex: initial commit for Marvell mwifiex driver") Signed-off-by:
Brian Norris <briannorris@chromium.org> Signed-off-by:
Kalle Valo <kvalo@codeaurora.org> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Florian Fainelli authored
commit 0cdefd5b upstream. The CPU port of the BCM53125 is configured with RGMII (no delays) but this should actually be RGMII with transmit delay (rgmii-txid) because STMMAC takes care of inserting the transmitter delay. This fixes occasional packet loss encountered. Fixes: d7b9eaff ("ARM: dts: sun7i: Add BCM53125 switch nodes to the lamobo-r1 board") Reported-by:
Hartmut Knaack <knaack.h@gmx.de> Signed-off-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
Maxime Ripard <maxime.ripard@free-electrons.com> Signed-off-by:
-
Matthijs van Duin authored
commit 448c077e upstream. 'adr' yields a data-pointer, not a function-pointer. Fixes: 999f934d ("ARM: omap5/dra7xx: Enable booting secondary CPU in HYP mode") Signed-off-by:
Matthijs van Duin <matthijsvanduin@gmail.com> Signed-off-by:
Tony Lindgren <tony@atomide.com> Signed-off-by:
-
Jon Mason authored
commit acfa28b3 upstream. The libgpio code pre-sets the GPIO values for the gpio-reset in the device tree. This results in the device being reset during bringup. To prevent this pre-setting, use the "open-source" flag in the device tree. Signed-off-by:
Jon Mason <jon.mason@broadcom.com> Fixes: b1aaf88b ("ARM: dts: NSP: Add GPIO reboot method to bcm958625hr DTS file") Fixes: 10baed1c ("ARM: dts: NSP: Add GPIO reboot method to bcm958625xmc DTS file") Fixes: 088e3148 ("ARM: dts: NSP: Add new DT file for bcm958522er") Fixes: e3227c12 ("ARM: dts: NSP: Add new DT file for bcm958525er") Fixes: 2f8bc002 ("ARM: dts: NSP: Add new DT file for bcm958622hr") Fixes: d454c376 ("ARM: dts: NSP: Add new DT file for bcm958623hr") Fixes: f27eacf2 ("ARM: dts: NSP: Add new DT file for bcm988312hr") Signed-off-by:
-
Arnd Bergmann authored
commit cbe99c53 upstream. gcc gets confused about the control flow in ktd2692_parse_dt(), causing it to warn about what seems like a potential bug: drivers/leds/leds-ktd2692.c: In function 'ktd2692_probe': drivers/leds/leds-ktd2692.c:244:15: error: '*((void *)&led_cfg+8)' may be used uninitialized in this function [-Werror=maybe-uninitialized] drivers/leds/leds-ktd2692.c:225:7: error: 'led_cfg.flash_max_microamp' may be used uninitialized in this function [-Werror=maybe-uninitialized] drivers/leds/leds-ktd2692.c:232:3: error: 'led_cfg.movie_max_microamp' may be used uninitialized in this function [-Werror=maybe-uninitialized] The code is fine, and slightly reworking it in an equivalent way lets gcc figure that out too, which gets rid of the warning. Fixes: 77e7915b ("leds: ktd2692: Add missing of_node_put") Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Acked-by:
Pavel Machek <pavel@ucw.cz> Signed-off-by:
Jacek Anaszewski <jacek.anaszewski@gmail.com> Signed-off-by:
-
Catalin Marinas authored
commit ec663d96 upstream. Commit cab15ce6 ("arm64: Introduce execute-only page access permissions") allowed a valid user PTE to have the PTE_USER bit clear. As a consequence, the pte_valid_not_user() macro in set_pte() was replaced with pte_valid_global() under the assumption that only user pages have the nG bit set. EFI mappings, however, also have the nG bit set and set_pte() wrongly ignores issuing the DSB+ISB. This patch reinstates the pte_valid_not_user() macro and adds the PTE_UXN bit check since all kernel mappings have this bit set. For clarity, pte_exec() is renamed to pte_user_exec() as it only checks for the absence of PTE_UXN. Consequently, the user executable check in set_pte_at() drops the pte_ng() test since pte_user_exec() is sufficient. Fixes: cab15ce6 ("arm64: Introduce execute-only page access permissions") Signed-off-by:
Catalin Marinas <catalin.marinas@arm.com> Signed-off-by:
Will Deacon <will.deacon@arm.com> Signed-off-by:
-
Stephen Boyd authored
commit 06dbf468 upstream. The ipq board has these rates as 25MHz, and not 19.2 and 27. I copy/pasted from other boards that have those rates but forgot to fix the rates here. Fixes: 30fc4212 ("arm: dts: qcom: Add more board clocks") Signed-off-by:
Stephen Boyd <sboyd@codeaurora.org> Signed-off-by:
Andy Gross <andy.gross@linaro.org> Signed-off-by:
-
Geert Uytterhoeven authored
commit 0d1390ff upstream. Device nodes representing I/O devices should be marked disabled in the SoC-specific DTS, and overridden by board-specific DTSes where needed. Fixes: a92843c8 ("arm64: dts: r8a7795: add EthernetAVB device node") Signed-off-by:
Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by:
Simon Horman <horms+renesas@verge.net.au> Signed-off-by:
-
Liam Breck authored
commit ba52e757 upstream. Reading both fault and status registers and logging any fault should take priority over handling status register update. Fix by moving the status handling to later in interrupt routine. Fixes: d7bf353f ("bq24190_charger: Add support for TI BQ24190 Battery Charger") Signed-off-by:
Liam Breck <kernel@networkimprov.net> Acked-by:
Mark Greer <mgreer@animalcreek.com> Acked-by:
Sebastian Reichel <sre@kernel.org> Signed-off-by:
-
Liam Breck authored
commit 68abfb80 upstream. Caching the fault register after a single I2C read may not keep an accurate value. Fix by doing two reads in irq_handle_thread() and using the cached value elsewhere. If a safety timer fault later clears itself, we apparently don't get an interrupt (INT), however other interrupts would refresh the register cache. From the data sheet: "When a fault occurs, the charger device sends out INT and keeps the fault state in REG09 until the host reads the fault register. Before the host reads REG09 and all the faults are cleared, the charger device would not send any INT upon new faults. In order to read the current fault status, the host has to read REG09 two times consecutively. The 1st reads fault register status from the last read [1] and the 2nd reads the current fault register status." [1] presumably a typo; should be "last fault" Fixes: d7bf353f ("bq24190_charger: Add support for TI BQ24190 Battery Charger") Signed-off-by:
-
Liam Breck authored
commit 2d9fee6a upstream. We wrongly get uevents for bq24190-charger and bq24190-battery on every register change. Fix by checking the association with charger and battery before emitting uevent(s). Fixes: d7bf353f ("bq24190_charger: Add support for TI BQ24190 Battery Charger") Signed-off-by:
-
Liam Breck authored
commit d62acc5e upstream. The device specific data is not fully initialized on request_threaded_irq(). This may cause a crash when the IRQ handler tries to reference them. Fix the issue by installing IRQ handler at the end of the probe. Fixes: d7bf353f ("bq24190_charger: Add support for TI BQ24190 Battery Charger") Signed-off-by:
-
Liam Breck authored
commit e05ad7e0 upstream. pm_resume() does a register_reset() which clears charger host mode. Fix by calling set_mode_host() after the reset. Fixes: d7bf353f ("bq24190_charger: Add support for TI BQ24190 Battery Charger") Signed-off-by:
-
Liam Breck authored
commit 767eee36 upstream. The interrupt signal is TRIGGER_FALLING. This is is specified in the data sheet PIN FUNCTIONS: "The INT pin sends active low, 256us pulse to host to report charger device status and fault." Also the direction can be seen in the data sheet Figure 37 "BQ24190 with D+/D- Detection and USB On-The-Go (OTG)" which shows a 10k pull-up resistor installed for the sample configurations. Fixes: d7bf353f ("bq24190_charger: Add support for TI BQ24190 Battery Charger") Signed-off-by:
-
Alexander Shishkin authored
commit 5443624b upstream. Commit: 8ee83b2a ("perf/x86/intel/pt: Add support for PTWRITE and power event tracing") forgot to add format strings to the PT driver. So one could enable these features by setting corresponding bits in the event config, but not by their mnemonic names. This patch adds the format strings. Signed-off-by:
Alexander Shishkin <alexander.shishkin@linux.intel.com> Signed-off-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@redhat.com> Cc: Borislav Petkov <bp@suse.de> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Vince Weaver <vincent.weaver@maine.edu> Cc: vince@deater.net Fixes: 8ee83b2a ("perf/x86/intel/pt: Add support for PTWRITE...") Link: http://lkml.kernel.org/r/20170127151644.8585-2-alexander.shishkin@linux.intel.comSigned-off-by:
Ingo Molnar <mingo@kernel.org> Signed-off-by:
-
Andrew Donnellan authored
commit eac6f8b0 upstream. Commit 38addce8 ("gcc-plugins: Add latent_entropy plugin") excludes certain powerpc early boot code from the latent entropy plugin by adding appropriate CFLAGS. It looks like this was supposed to cover prom_init.o, but ended up saying init.o (which doesn't exist) instead. Fix the typo. Fixes: 38addce8 ("gcc-plugins: Add latent_entropy plugin") Signed-off-by:
Andrew Donnellan <andrew.donnellan@au1.ibm.com> Signed-off-by:
Michael Ellerman <mpe@ellerman.id.au> Signed-off-by:
-
Anton Blanchard authored
commit 496e9cb5 upstream. The final paragraph of the help text is reversed. We want to enable this option by default, and disable it if the toolchain has a working -mprofile-kernel. Fixes: 8c50b72a ("powerpc/ftrace: Add Kconfig & Make glue for mprofile-kernel") Signed-off-by:
Anton Blanchard <anton@samba.org> Signed-off-by:
-
Michael Ellerman authored
commit a7e0fb6c upstream. Currently the opal_exit tracepoint usually shows the opcode as 0: <idle>-0 [047] d.h. 635.654292: opal_entry: opcode=63 <idle>-0 [047] d.h. 635.654296: opal_exit: opcode=0 retval=0 kopald-1209 [019] d... 636.420943: opal_entry: opcode=10 kopald-1209 [019] d... 636.420959: opal_exit: opcode=0 retval=0 This is because we incorrectly load the opcode into r0 before calling __trace_opal_exit(), whereas it expects the opcode in r3 (first function parameter). In fact we are leaving the retval in r3, so opcode and retval will always show the same value. Instead load the opcode into r3, resulting in: <idle>-0 [040] d.h. 636.618625: opal_entry: opcode=63 <idle>-0 [040] d.h. 636.618627: opal_exit: opcode=63 retval=0 Fixes: c49f6353 ("powernv: Add OPAL tracepoints") Signed-off-by:
-
Aneesh Kumar K.V authored
commit 4ab2537c upstream. In commit a4b34954 ("powerpc/mm: Cleanup LPCR defines") we updated LPCR_VRMASD wrongly as below. -#define LPCR_VRMASD (0x1ful << (63-16)) +#define LPCR_VRMASD_SH 47 +#define LPCR_VRMASD (ASM_CONST(1) << LPCR_VRMASD_SH) We initialize the VRMA bits in LPCR to 0x00 in kvm. Hence using a different mask value as above while updating lpcr should not have any impact. This patch updates it to the correct value. Fixes: a4b34954 ("powerpc/mm: Cleanup LPCR defines") Reported-by:
Ram Pai <linuxram@us.ibm.com> Signed-off-by:
Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by:
Jia He <hejianet@gmail.com> Acked-by:
Paul Mackerras <paulus@ozlabs.org> Signed-off-by:
-
Ben Hutchings authored
commit 4cca0457 upstream. The switch that conditionally sets CPUPOWER_CAP_HAS_TURBO_RATIO and CPUPOWER_CAP_IS_SNB flags is missing a break, so all cores get both flags set and an assumed base clock of 100 MHz for turbo values. Reported-by:
GSR <gsr.bugs@infernal-iceberg.com> Tested-by:
Ben Hutchings <ben@decadent.org.uk> Signed-off-by:
Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by:
-
Giedrius Statkevičius authored
commit bdd9968d upstream. val might become 7 in which case stime[7] (array of length 7) would be accessed during the scnprintf call later and that will cause issues. Obviously, string concatenation is not intended here so just a comma needs to be added to fix the issue. Fixes: 98a27664 ("power_supply: Add new lp8788 charger driver") Signed-off-by:
Giedrius Statkevičius <giedrius.statkevicius@gmail.com> Acked-by:
Milo Kim <milo.kim@ti.com> Signed-off-by:
-
Horia Geantă authored
commit 87ec02e7 upstream. In case ctx_dma dma mapping fails, ahash_unmap_ctx() tries to dma unmap an invalid address: map_seq_out_ptr_ctx() / ctx_map_to_sec4_sg() -> goto unmap_ctx -> -> ahash_unmap_ctx() -> dma unmap ctx_dma There is also possible to reach ahash_unmap_ctx() with ctx_dma uninitialzed or to try to unmap the same address twice. Fix these by setting ctx_dma = 0 where needed: -initialize ctx_dma in ahash_init() -clear ctx_dma in case of mapping error (instead of holding the error code returned by the dma map function) -clear ctx_dma after each unmapping Fixes: 32686d34 ("crypto: caam - ensure that we clean up after an error") Signed-off-by:
Horia Geantă <horia.geanta@nxp.com> Signed-off-by:
Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by:
-
Winkler, Tomas authored
commit 2998b02b upstream. The tpm stack uses pdev name convention for the parent device. Fix that also in tpm_chip_alloc(). Fixes: 3897cd9c ("tpm: Split out the devm stuff from tpmm_chip_alloc")' Signed-off-by:
Tomas Winkler <tomas.winkler@intel.com> Reviewed-by:
Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Tested-by:
-
Jarkko Sakkinen authored
commit 7d761119 upstream. The error code handling is broken as any error code that has the same bits set as TPM_RC_HASH passes. Implemented tpm2_rc_value() helper to parse the error value from FMT0 and FMT1 error codes so that these types of mistakes are prevented in the future. Fixes: 5ca4c20c ("keys, trusted: select hash algorithm for TPM2 chips") Signed-off-by:
Jason Gunthorpe <jgunthorpe@obsidianresearch.com> Signed-off-by:
-
Guenter Roeck authored
commit d66777ca upstream. pwm4 is enabled if bit 2 of GPIO control register 4 is disabled, not when it is enabled. Since the check is for the skip condition, it is reversed. This applies to both IT8620 and IT8628. Fixes: 36c4d98a ("hwmon: (it87) Add support for all pwm channels ...") Signed-off-by:
Guenter Roeck <linux@roeck-us.net> Signed-off-by:
-
Vincent Abriou authored
commit 2f410f88 upstream. On stih407-410 chip family the GDP layers are able to support up to UHD resolution (3840 x 2160). Signed-off-by:
Vincent Abriou <vincent.abriou@st.com> Acked-by:
Lee Jones <lee.jones@linaro.org> Tested-by:
-
Cong Wang authored
commit b5c66bab upstream. posix_acl_update_mode() could possibly clear 'acl', if so we leak the memory pointed by 'acl'. Save this pointer before calling posix_acl_update_mode() and release the memory if 'acl' really gets cleared. Link: http://lkml.kernel.org/r/1486678332-2430-1-git-send-email-xiyou.wangcong@gmail.comSigned-off-by:
Cong Wang <xiyou.wangcong@gmail.com> Reported-by:
Mark Salyzyn <salyzyn@android.com> Reviewed-by:
Jan Kara <jack@suse.cz> Reviewed-by:
Greg Kurz <groug@kaod.org> Cc: Eric Van Hensbergen <ericvh@gmail.com> Cc: Ron Minnich <rminnich@sandia.gov> Cc: Latchesar Ionkov <lucho@ionkov.net> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by:
-
- 08 May, 2017 10 commits
-
-
Greg Kroah-Hartman authored
-
Adrian Salido authored
commit 4617f564 upstream. When calling a dm ioctl that doesn't process any data (IOCTL_FLAGS_NO_PARAMS), the contents of the data field in struct dm_ioctl are left initialized. Current code is incorrectly extending the size of data copied back to user, causing the contents of kernel stack to be leaked to user. Fix by only copying contents before data and allow the functions processing the ioctl to override. Signed-off-by:
Adrian Salido <salidoa@google.com> Reviewed-by:
Alasdair G Kergon <agk@redhat.com> Signed-off-by:
Mike Snitzer <snitzer@redhat.com> Signed-off-by:
-
Sebastian Andrzej Siewior authored
commit dc434e05 upstream. The setup/remove_state/instance() functions in the hotplug core code are serialized against concurrent CPU hotplug, but unfortunately not serialized against themself. As a consequence a concurrent invocation of these function results in corruption of the callback machinery because two instances try to invoke callbacks on remote cpus at the same time. This results in missing callback invocations and initiator threads waiting forever on the completion. The obvious solution to replace get_cpu_online() with cpu_hotplug_begin() is not possible because at least one callsite calls into these functions from a get_online_cpu() locked region. Extend the protection scope of the cpuhp_state_mutex from solely protecting the state arrays to cover the callback invocation machinery as well. Fixes: 5b7aa87e ("cpu/hotplug: Implement setup/removal interface") Reported-and-tested-by:
Bart Van Assche <Bart.VanAssche@sandisk.com> Signed-off-by:
Sebastian Andrzej Siewior <bigeasy@linutronix.de> Cc: hpa@zytor.com Cc: mingo@kernel.org Cc: akpm@linux-foundation.org Cc: torvalds@linux-foundation.org Link: http://lkml.kernel.org/r/20170314150645.g4tdyoszlcbajmna@linutronix.deSigned-off-by:
Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
-
Yan, Zheng authored
commit 2b1ac852 upstream. For readahead/fadvise cases, caller of ceph_readpages does not hold buffer capability. Pages can be added to page cache while there is no buffer capability. This can cause data integrity issue. Signed-off-by:
Yan, Zheng <zyan@redhat.com> Signed-off-by:
-
Gabriel Krisman Bertazi authored
commit c130b666 upstream. Commit f209fa03 ("serial: 8250_pci: Detach low-level driver during PCI error recovery") introduces a potential use-after-free in case the pciserial_init_ports call in serial8250_io_resume fails, which may happen if a memory allocation fails or if the .init quirk failed for whatever reason). If this happen, further pci_get_drvdata will return a pointer to freed memory. This patch reworks the PCI recovery resume hook to restore the old priv structure in this case, which should be ok, since the ports were already detached. Such error during recovery causes us to give up on the recovery. Fixes: f209fa03 ("serial: 8250_pci: Detach low-level driver during PCI error recovery") Reported-by:
Michal Suchanek <msuchanek@suse.com> Signed-off-by:
Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com> Signed-off-by:
Guilherme G. Piccoli <gpiccoli@linux.vnet.ibm.com> Signed-off-by:
-
Guenter Roeck authored
commit 8358378b upstream. IT8705F is known to respond on both SIO addresses. Registering it twice may result in system lockups. Reported-by:
Russell King <linux@armlinux.org.uk> Fixes: e84bd953 ("hwmon: (it87) Add support for second Super-IO chip") Signed-off-by:
-
Stephen Hemminger authored
commit f1c635b4 upstream. Hyper-V host emulation of SCSI for virtual DVD device reports SCSI version 0 (UNKNOWN) but is still capable of supporting REPORTLUN. Without this patch, a GEN2 Linux guest on Hyper-V will not boot 4.11 successfully with virtual DVD ROM device. What happens is that the SCSI scan process falls back to doing sequential probing by INQUIRY. But the storvsc driver has a previous workaround that masks/blocks all errors reports from INQUIRY (or MODE_SENSE) commands. This workaround causes the scan to then populate a full set of bogus LUN's on the target and then sends kernel spinning off into a death spiral doing block reads on the non-existent LUNs. By setting the correct blacklist flags, the target with the DVD device is scanned with REPORTLUN and that works correctly. Patch needs to go in current 4.11, it is safe but not necessary in older kernels. Signed-off-by:
Stephen Hemminger <sthemmin@microsoft.com> Reviewed-by:
K. Y. Srinivasan <kys@microsoft.com> Reviewed-by:
Christoph Hellwig <hch@lst.de> Signed-off-by:
Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by:
-
Maciej S. Szmigiero authored
commit 1d70fe9d upstream. Since commit 1107d065 ("tpm_tis: Introduce intermediate layer for TPM access") Atmel 3203 TPM on ThinkPad X61S (TPM firmware version 13.9) no longer works. The initialization proceeds fine until we get and start using chip-reported timeouts - and the chip reports C and D timeouts of zero. It turns out that until commit 8e54caf4 ("tpm: Provide a generic means to override the chip returned timeouts") we had actually let default timeout values remain in this case, so let's bring back this behavior to make chips like Atmel 3203 work again. Use a common code that was introduced by that commit so a warning is printed in this case and /sys/class/tpm/tpm*/timeouts correctly says the timeouts aren't chip-original. This is a backport for 4.9 kernel version of the original commit, with renaming of "TPM_TIS_ITPM_POSSIBLE" flag removed since it was only a cosmetic change and not a part of the real bug fix. Fixes: 1107d065 ("tpm_tis: Introduce intermediate layer for TPM access") Cc: stable@vger.kernel.org Signed-off-by:
Maciej S. Szmigiero <mail@maciej.szmigiero.name> Reviewed-by:
-
Sachin Prabhu authored
commit 38bd4906 upstream. A signal can interrupt a SendReceive call which result in incoming responses to the call being ignored. This is a problem for calls such as open which results in the successful response being ignored. This results in an open file resource on the server. The patch looks into responses which were cancelled after being sent and in case of successful open closes the open fids. For this patch, the check is only done in SendReceive2() RH-bz: 1403319 Signed-off-by:
Sachin Prabhu <sprabhu@redhat.com> Reviewed-by:
Pavel Shilovsky <pshilov@microsoft.com> Acked-by:
-
Thomas Gleixner authored
commit 1e38da30 upstream. The handling of the might_cancel queueing is not properly protected, so parallel operations on the file descriptor can race with each other and lead to list corruptions or use after free. Protect the context for these operations with a seperate lock. The wait queue lock cannot be reused for this because that would create a lock inversion scenario vs. the cancel lock. Replacing might_cancel with an atomic (atomic_t or atomic bit) does not help either because it still can race vs. the actual list operation. Reported-by:
Dmitry Vyukov <dvyukov@google.com> Signed-off-by:
-
- 03 May, 2017 3 commits
-
-
Greg Kroah-Hartman authored
-
Josh Poimboeuf authored
commit 34a477e5 upstream. On x86-32, with CONFIG_FIRMWARE and multiple CPUs, if you enable function graph tracing and then suspend to RAM, it will triple fault and reboot when it resumes. The first fault happens when booting a secondary CPU: startup_32_smp() load_ucode_ap() prepare_ftrace_return() ftrace_graph_is_dead() (accesses 'kill_ftrace_graph') The early head_32.S code calls into load_ucode_ap(), which has an an ftrace hook, so it calls prepare_ftrace_return(), which calls ftrace_graph_is_dead(), which tries to access the global 'kill_ftrace_graph' variable with a virtual address, causing a fault because the CPU is still in real mode. The fix is to add a check in prepare_ftrace_return() to make sure it's running in protected mode before continuing. The check makes sure the stack pointer is a virtual kernel address. It's a bit of a hack, but it's not very intrusive and it works well enough. For reference, here are a few other (more difficult) ways this could have potentially been fixed: - Move startup_32_smp()'s call to load_ucode_ap() down to *after* paging is enabled. (No idea what that would break.) - Track down load_ucode_ap()'s entire callee tree and mark all the functions 'notrace'. (Probably not realistic.) - Pause graph tracing in ftrace_suspend_notifier_call() or bringup_cpu() or __cpu_up(), and ensure that the pause facility can be queried from real mode. Reported-by:
Paul Menzel <pmenzel@molgen.mpg.de> Signed-off-by:
Josh Poimboeuf <jpoimboe@redhat.com> Tested-by:
Steven Rostedt (VMware) <rostedt@goodmis.org> Cc: "Rafael J . Wysocki" <rjw@rjwysocki.net> Cc: linux-acpi@vger.kernel.org Cc: Borislav Petkov <bp@alien8.de> Cc: Len Brown <lenb@kernel.org> Link: http://lkml.kernel.org/r/5c1272269a580660703ed2eccf44308e790c7a98.1492123841.git.jpoimboe@redhat.comSigned-off-by:
-
Vineet Gupta authored
commit ecd43afd upstream. This is not exposed to userspace debugers yet, which can be done independently as a seperate patch ! Signed-off-by:
Vineet Gupta <vgupta@synopsys.com> Signed-off-by:
-
