- 06 May, 2015 40 commits
-
-
Felipe Balbi authored
commit 7e136bb7 upstream. Make sure we're using the new macro, so our resume signaling will always pass certification. Signed-off-by:
Felipe Balbi <balbi@ti.com> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Felipe Balbi authored
commit b8fb6f79 upstream. Make sure we're using the new macro, so our resume signaling will always pass certification. Signed-off-by:
-
Felipe Balbi authored
commit 595227db upstream. Make sure we're using the new macro, so our resume signaling will always pass certification. Signed-off-by:
-
Felipe Balbi authored
commit 84c0d178 upstream. Make sure we're using the new macro, so our resume signaling will always pass certification. Signed-off-by:
-
Felipe Balbi authored
commit 309be239 upstream. Make sure we're using the new macro, so our resume signaling will always pass certification. Based on original work by Bin Liu <Bin Liu <b-liu@ti.com>> Cc: Bin Liu <b-liu@ti.com> Signed-off-by:
-
Felipe Balbi authored
commit 62f0342d upstream. Every USB Host controller should use this new macro to define for how long resume signalling should be driven on the bus. Currently, almost every single USB controller is using a 20ms timeout for resume signalling. That's problematic for two reasons: a) sometimes that 20ms timer expires a little before 20ms, which makes us fail certification b) some (many) devices actually need more than 20ms resume signalling. Sure, in case of (b) we can state that the device is against the USB spec, but the fact is that we have no control over which device the certification lab will use. We also have no control over which host they will use. Most likely they'll be using a Windows PC which, again, we have no control over how that USB stack is written and how long resume signalling they are using. At the end of the day, we must make sure Linux passes electrical compliance when working as Host or as Device and currently we don't pass compliance as host because we're driving resume signallig for exactly 20ms and that confuses certification test setup resulting in Certification failure. Acked-by:
Peter Chen <peter.chen@freescale.com> Signed-off-by:
-
Axel Lin authored
commit 869aee0f upstream. The res parameter passed to devm_usb_phy_match() is the location where the pointer to the usb_phy is stored, hence it needs to be dereferenced before comparing to the match data in order to find the correct match. Fixes: 410219dc ("usb: otg: utils: devres: Add API's to associate a device with the phy") Signed-off-by:
Axel Lin <axel.lin@ingics.com> Signed-off-by:
-
Felipe Balbi authored
commit e3c93e1a upstream. As per Mentor Graphics' documentation, we should always handle TX endpoints before RX endpoints. This patch fixes that error while also updating some hard-to-read comments which were scattered around musb_interrupt(). This patch should be backported as far back as possible since this error has been in the driver since it's conception. Signed-off-by:
-
Andreas Faerber authored
commit 7e9e20b1 upstream. Resolve a merge conflict with mmc refactoring aaa25a5a ("ARM: dts: unuse the slot-node and deprecate the supports-highspeed for dw-mmc in exynos") by dropping the slot@0 nodes, moving its bus-width property to the mmc node and replacing supports-highspeed with cap-{mmc,sd}-highspeed, matching exynos5250-snow. Cc: Jaehoon Chung <jh80.chung@samsung.com> Fixes: 53dd4138 ("ARM: dts: Add exynos5250-spring device tree") Signed-off-by:
Andreas Faerber <afaerber@suse.de> Reviewed-by:
Javier Martinez Canillas <javier.martinez@collabora.co.uk> Signed-off-by:
Kukjin Kim <kgene@kernel.org> Signed-off-by:
-
Sebastian Hesselbarth authored
commit a74cd13b upstream. Fix Dove's register addresses of uart2 and uart3 nodes that seem to be broken since ages due to a copy-and-paste error. Signed-off-by:
Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com> Acked-by:
Gregory CLEMENT <gregory.clement@free-electrons.com> Signed-off-by:
-
Nicolas Ferre authored
commit 98b80987 upstream. After 57a38eff (net: phy: micrel: disable broadcast for KSZ8081/KSZ8091) the macb1 interface refuses to work properly because it tries to cling to address 0 which isn't able to communicate in broadcast with the mac anymore. The micrel phy on the board is actually configured to show up at address 1. Adding the phy node and its real address fixes the issue. Signed-off-by:
Nicolas Ferre <nicolas.ferre@atmel.com> Cc: Johan Hovold <johan@kernel.org> Signed-off-by:
Olof Johansson <olof@lixom.net> Signed-off-by:
-
Charles Keepax authored
commit 4e330ae4 upstream. There are two PMICs on Cragganmore, currently one dynamically assign its IRQ base and the other uses a fixed base. It is possible for the statically assigned PMIC to fail if its IRQ is taken by the dynamically assigned one. Fix this by statically assigning both the IRQ bases. Signed-off-by:
Charles Keepax <ckeepax@opensource.wolfsonmicro.com> Signed-off-by:
-
Gregory CLEMENT authored
commit 548ae94c upstream. On Armada 38x SoCs, under heavy I/O load, the system hangs when CPU Idle is enabled. Waiting for a solution to this issue, this patch disables the CPU Idle support for this SoC. As CPU Hot plug support also uses some of the CPU Idle functions it is also affected by the same issue. This patch disables it also for the Armada 38x SoCs. Signed-off-by:
Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Signed-off-by:
-
Andrey Ryabinin authored
commit 8defb336 upstream. Usually ELF_ET_DYN_BASE is 2/3 of TASK_SIZE. With 3G/1G user/kernel split this is not so, because 2*TASK_SIZE overflows 32 bits, so the actual value of ELF_ET_DYN_BASE is: (2 * TASK_SIZE / 3) = 0x2a000000 When ASLR is disabled PIE binaries will load at ELF_ET_DYN_BASE address. On 32bit platforms AddressSanitzer uses addresses [0x20000000 - 0x40000000] for shadow memory [1]. So ASan doesn't work for PIE binaries when ASLR disabled as it fails to map shadow memory. Also after Kees's 'split ET_DYN ASLR from mmap ASLR' patchset PIE binaries has a high chance of loading somewhere in between [0x2a000000 - 0x40000000] even if ASLR enabled. This makes ASan with PIE absolutely incompatible. Fix overflow by dividing TASK_SIZE prior to multiplying. After this patch ELF_ET_DYN_BASE equals to (for CONFIG_VMSPLIT_3G=y): (TASK_SIZE / 3 * 2) = 0x7f555554 [1] https://code.google.com/p/address-sanitizer/wiki/AddressSanitizerAlgorithm#MappingSigned-off-by:
Andrey Ryabinin <a.ryabinin@samsung.com> Reported-by:
Maria Guseva <m.guseva@samsung.com> Signed-off-by:
Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by:
-
Russell King authored
commit 767bf7e7 upstream. Normally, when a CPU wants to clear a cache line to zero in the external L2 cache, it would generate bus cycles to write each word as it would do with any other data access. However, a Cortex A9 connected to a L2C-310 has a specific feature where the CPU can detect this operation, and signal that it wants to zero an entire cache line. This feature, known as Full Line of Zeros (FLZ), involves a non-standard AXI signalling mechanism which only the L2C-310 can properly interpret. There are separate enable bits in both the L2C-310 and the Cortex A9 - the L2C-310 needs to be enabled and have the FLZ enable bit set in the auxiliary control register before the Cortex A9 has this feature enabled. Unfortunately, the suspend code was not respecting this - it's not obvious from the code: swsusp_arch_suspend() cpu_suspend() /* saves the Cortex A9 auxiliary control register */ arch_save_image() soft_restart() /* turns off FLZ in Cortex A9, and disables L2C */ cpu_resume() /* restores the Cortex A9 registers, inc auxcr */ At this point, we end up with the L2C disabled, but the Cortex A9 with FLZ enabled - which means any memset() or zeroing of a full cache line will fail to take effect. A similar issue exists in the resume path, but it's slightly more complex: swsusp_arch_suspend() cpu_suspend() /* saves the Cortex A9 auxiliary control register */ arch_save_image() /* image with A9 auxcr saved */ ... swsusp_arch_resume() call_with_stack() arch_restore_image() /* restores image with A9 auxcr saved above */ soft_restart() /* turns off FLZ in Cortex A9, and disables L2C */ cpu_resume() /* restores the Cortex A9 registers, inc auxcr */ Again, here we end up with the L2C disabled, but Cortex A9 FLZ enabled. There's no need to turn off the L2C in either of these two paths; there are benefits from not doing so - for example, the page copies will be faster with the L2C enabled. Hence, fix this by providing a variant of soft_restart() which can be used without turning the L2 cache controller off, and use it in both of these paths to keep the L2C enabled across the respective resume transitions. Fixes: 8ef418c7 ("ARM: l2c: trial at enabling some Cortex-A9 optimisations") Reported-by:
Sean Cross <xobs@kosagi.com> Tested-by:
-
Andrew Elble authored
commit c1b8940b upstream. We have observed a BUG() crash in fs/attr.c:notify_change(). The crash occurs during an rsync into a filesystem that is exported via NFS. 1.) fs/attr.c:notify_change() modifies the caller's version of attr. 2.) 6de0ec00 ("VFS: make notify_change pass ATTR_KILL_S*ID to setattr operations") introduced a BUG() restriction such that "no function will ever call notify_change() with both ATTR_MODE and ATTR_KILL_S*ID set". Under some circumstances though, it will have assisted in setting the caller's version of attr to this very combination. 3.) 27ac0ffe ("locks: break delegations on any attribute modification") introduced code to handle breaking delegations. This can result in notify_change() being re-called. attr _must_ be explicitly reset to avoid triggering the BUG() established in #2. 4.) The path that that triggers this is via fs/open.c:chmod_common(). The combination of attr flags set here and in the first call to notify_change() along with a later failed break_deleg_wait() results in notify_change() being called again via retry_deleg without resetting attr. Solution is to move retry_deleg in chmod_common() a bit further up to ensure attr is completely reset. There are other places where this seemingly could occur, such as fs/utimes.c:utimes_common(), but the attr flags are not initially set in such a way to trigger this. Fixes: 27ac0ffe ("locks: break delegations on any attribute modification") Reported-by:
Eric Meddaugh <etmsys@rit.edu> Tested-by:
Andrew Elble <aweits@rit.edu> Signed-off-by:
Al Viro <viro@zeniv.linux.org.uk> Signed-off-by:
-
Krzysztof Kozlowski authored
commit a2c1d531 upstream. The return values of create_singlethread_workqueue() and power_supply_register() calls were not checked and even on error probe() function returned 0. 1. If allocation of workqueue failed (returning NULL) then further accesses could lead to NULL pointer dereference. The queue_delayed_work() expects workqueue to be non-NULL. 2. If registration of power supply failed then during unbind the driver tried to unregister power supply which was not actually registered. This could lead to memory corruption because power_supply_unregister() unconditionally cleans up given power supply. Signed-off-by:
Krzysztof Kozlowski <k.kozlowski@samsung.com> Fixes: 00a588f9 ("power: add driver for battery reading on iPaq h3xxx") Signed-off-by:
Sebastian Reichel <sre@kernel.org> Signed-off-by:
-
Krzysztof Kozlowski authored
commit f852ec46 upstream. Driver allocates singlethread workqueue in probe but it is not destroyed during removal. Signed-off-by:
-
Krzysztof Kozlowski authored
commit a7117f81 upstream. Driver forgot to unregister charger power supply if registering of battery supply failed in probe(). In such case the memory associated with power supply leaked. Signed-off-by:
-
Krzysztof Kozlowski authored
commit 68c3ed6f upstream. The return value of power_supply_register() call was not checked and even on error probe() function returned 0. If registering failed then during unbind the driver tried to unregister power supply which was not actually registered. This could lead to memory corruption because power_supply_unregister() unconditionally cleans up given power supply. Signed-off-by:
-
Steven Rostedt authored
commit 80a9b64e upstream. It has come to my attention that this_cpu_read/write are horrible on architectures other than x86. Worse yet, they actually disable preemption or interrupts! This caused some unexpected tracing results on ARM. 101.356868: preempt_count_add <-ring_buffer_lock_reserve 101.356870: preempt_count_sub <-ring_buffer_lock_reserve The ring_buffer_lock_reserve has recursion protection that requires accessing a per cpu variable. But since preempt_disable() is traced, it too got traced while accessing the variable that is suppose to prevent recursion like this. The generic version of this_cpu_read() and write() are: #define this_cpu_generic_read(pcp) \ ({ typeof(pcp) ret__; \ preempt_disable(); \ ret__ = *this_cpu_ptr(&(pcp)); \ preempt_enable(); \ ret__; \ }) #define this_cpu_generic_to_op(pcp, val, op) \ do { \ unsigned long flags; \ raw_local_irq_save(flags); \ *__this_cpu_ptr(&(pcp)) op val; \ raw_local_irq_restore(flags); \ } while (0) Which is unacceptable for locations that know they are within preempt disabled or interrupt disabled locations. Paul McKenney stated that __this_cpu_() versions produce much better code on other architectures than this_cpu_() does, if we know that the call is done in a preempt disabled location. I also changed the recursive_unlock() to use two local variables instead of accessing the per_cpu variable twice. Link: http://lkml.kernel.org/r/20150317114411.GE3589@linux.vnet.ibm.com Link: http://lkml.kernel.org/r/20150317104038.312e73d1@gandalf.local.homeAcked-by:
Christoph Lameter <cl@linux.com> Reported-by:
Uwe Kleine-Koenig <u.kleine-koenig@pengutronix.de> Tested-by:
Steven Rostedt <rostedt@goodmis.org> Signed-off-by:
-
Krzysztof Kozlowski authored
commit 1915a718 upstream. The return value of power_supply_register() call was not checked and even on error probe() function returned 0. If registering failed then during unbind the driver tried to unregister power supply which was not actually registered. This could lead to memory corruption because power_supply_unregister() unconditionally cleans up given power supply. Fix this by checking return status of power_supply_register() call. In case of failure, clean up sysfs entries and fail the probe. Signed-off-by:
-
Krzysztof Kozlowski authored
commit ad774702 upstream. The commit c2be45f0 ("compal-laptop: Use devm_hwmon_device_register_with_groups") wanted to change the registering of hwmon device to resource-managed version. It mostly did it except the main thing - it forgot to use devm-like function so the hwmon device leaked after device removal or probe failure. Signed-off-by:
Guenter Roeck <linux@roeck-us.net> Acked-by:
Darren Hart <dvhart@linux.intel.com> Signed-off-by:
-
Ian Abbott authored
commit f20fbaad upstream. `spidev_message()` sums the lengths of the individual SPI transfers to determine the overall SPI message length. It restricts the total length, returning an error if too long, but it does not check for arithmetic overflow. For example, if the SPI message consisted of two transfers and the first has a length of 10 and the second has a length of (__u32)(-1), the total length would be seen as 9, even though the second transfer is actually very long. If the second transfer specifies a null `rx_buf` and a non-null `tx_buf`, the `copy_from_user()` could overrun the spidev's pre-allocated tx buffer before it reaches an invalid user memory address. Fix it by checking that neither the total nor the individual transfer lengths exceed the maximum allowed value. Thanks to Dan Carpenter for reporting the potential integer overflow. Signed-off-by:
Ian Abbott <abbotti@mev.co.uk> Signed-off-by:
Mark Brown <broonie@kernel.org> Signed-off-by:
-
Lucas Stach authored
commit f511ab09 upstream. They are used to decide if the controller can do DMA on a buffer of a specific length and thus are needed before any transfer is attempted. This fixes a memory leak where the SPI core uses the drivers can_dma() callback to determine if a buffer needs to be mapped. As the watermark levels aren't correct at that point the driver falsely claims to be able to DMA the buffer when it fact it isn't. After the transfer has been done the core uses the same callback to determine if it needs to unmap the buffers. As the driver now correctly claims to not being able to DMA the buffer the core doesn't attempt to unmap the buffer which leaves the SGT leaking. Fixes: f62caccd (spi: spi-imx: add DMA support) Signed-off-by:
Lucas Stach <l.stach@pengutronix.de> Signed-off-by:
-
Chen-Yu Tsai authored
commit 9e71c589 upstream. The reset control for the sunxi mmc controller is optional. Some newer platforms (sun6i, sun8i, sun9i) have it, while older ones (sun4i, sun5i, sun7i) don't. Use the properly stubbed _optional version so the driver does not fail to compile when RESET_CONTROLLER=n. This patch also adds a check for deferred probing on the reset control. Signed-off-by:
Chen-Yu Tsai <wens@csie.org> Acked-by:
David Lanzendörfer <david.lanzendoerfer@o2s.ch> Signed-off-by:
Ulf Hansson <ulf.hansson@linaro.org> Signed-off-by:
-
Oliver Neukum authored
commit 323ece54 upstream. Values directly from descriptors given in debug statements must be converted to native endianness. Signed-off-by:
Oliver Neukum <oneukum@suse.de> Signed-off-by:
-
Sudeep Holla authored
commit 8a7d95f9 upstream. On architectures that depend on DT for obtaining cache hierarcy, we need to validate the device node for all the cache indices, failing to do so might result in wrong information being exposed to the userspace. This is quite possible on initial/incomplete versions of the device trees. In such cases, it's better to bail out if all the required device nodes are not present. This patch adds checks for the validation of device node for all the caches and doesn't initialise the cacheinfo if there's any error. Reported-by:
Mark Rutland <mark.rutland@arm.com> Acked-by:
Sudeep Holla <sudeep.holla@arm.com> Signed-off-by:
-
NeilBrown authored
commit 47d68979 upstream. Since commit 20d0189b in v3.14-rc1 RAID0 has performed incorrect calculations when the chunksize is not a power of 2. This happens because "sector_div()" modifies its first argument, but this wasn't taken into account in the patch. So restore that first arg before re-using the variable. Reported-by:
Joe Landman <joe.landman@gmail.com> Reported-by:
Dave Chinner <david@fromorbit.com> Fixes: 20d0189bSigned-off-by:
NeilBrown <neilb@suse.de> Signed-off-by:
-
Malcolm Priestley authored
commit a6388e68 upstream. Information for packet type is in ieee80211_tx_info band IEEE80211_BAND_5GHZ for PK_TYPE_11A. IEEE80211_TX_RC_USE_CTS_PROTECT via tx_rate flags selects PK_TYPE_11GB This ensures that the packet is always the right type. Signed-off-by:
Malcolm Priestley <tvboxspy@gmail.com> Signed-off-by:
-
Alistair Strachan authored
commit 8e43c9c7 upstream. The android_fence_release() function checks for active sync points by calling list_empty() on the list head embedded on the sync point. However, it is only valid to use list_empty() on nodes that have been initialized with INIT_LIST_HEAD() or list_del_init(). Because the list entry has likely been removed from the active list by sync_timeline_signal(), there is a good chance that this WARN_ON_ONCE() will be hit due to dangling pointers pointing at freed memory (even though the sync drivers did nothing wrong) and memory corruption will ensue as the list entry is removed for a second time, corrupting the active list. This problem can be reproduced quite easily with CONFIG_DEBUG_LIST=y and fences with more than one sync point. Signed-off-by:
Alistair Strachan <alistair.strachan@imgtec.com> Cc: Maarten Lankhorst <maarten.lankhorst@canonical.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Colin Cross <ccross@google.com> Signed-off-by:
-
Sudip Mukherjee authored
commit 2c20d92d upstream. the lcd type as defined in the Kconfig is not matching in the code. as a result the rs, rw and en pins were getting interchanged. Kconfig defines the value of PANEL_LCD to be 1 if we select custom configuration but in the code LCD_TYPE_CUSTOM is defined as 5. my hardware is LCD_TYPE_CUSTOM, but the pins were assigned to it as pins of LCD_TYPE_OLD, and it was not working. Now values are corrected with referenece to the values defined in Kconfig and it is working. checked on JHD204A lcd with LCD_TYPE_CUSTOM configuration. Signed-off-by:
Sudip Mukherjee <sudip@vectorindia.org> Acked-by:
Willy Tarreau <w@1wt.eu> Signed-off-by:
-
Huacai Chen authored
commit a843d00d upstream. We found that TLB mismatch not only happens after kernel resume, but also happens during snapshot restore. So move it to the beginning of swsusp_arch_suspend(). Signed-off-by:
Huacai Chen <chenhc@lemote.com> Cc: Steven J. Hill <Steven.Hill@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Fuxin Zhang <zhangfx@lemote.com> Cc: Zhangjin Wu <wuzhangjin@gmail.com> Patchwork: https://patchwork.linux-mips.org/patch/9621/Signed-off-by:
Ralf Baechle <ralf@linux-mips.org> Signed-off-by:
-
Huacai Chen authored
commit 0add9c2f upstream. HPET irq is routed to i8259 and then to MIPS CPU irq (cascade). After commit a3e6c1ef (MIPS: IRQ: Fix disable_irq on CPU IRQs), if without IRQF_NO_SUSPEND in cascade_irqaction, HPET interrupts will lost during suspend. The result is machine cannot be waken up. Signed-off-by:
-
Markos Chandras authored
commit 60cd7e08 upstream. Introduce new macros for kernel load/store variants which will be used to perform regular kernel space load/store operations in EVA mode. Signed-off-by:
Markos Chandras <markos.chandras@imgtec.com> Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/9500/Signed-off-by:
-
Markos Chandras authored
commit f7f8aea4 upstream. memsize denotes the amount of RAM we can access from kseg{0,1} and that should be up to 256M. In case the bootloader reports a value higher than that (perhaps reporting all the available RAM) it's best if we fix it ourselves and just warn the user about that. This is usually a problem with the bootloader and/or its environment. [ralf@linux-mips.org: Remove useless parens as suggested bei Sergei. Reformat long pr_warn statement to fit into 80 column limit.] Signed-off-by:
-
James Hogan authored
commit acaf6a97 upstream. The lose_fpu() function only disables the FPU in CP0_Status.CU1 if the FPU is in use and MSA isn't enabled. This isn't necessarily a problem because KSTK_STATUS(current), the version of CP0_Status stored on the kernel stack on entry from user mode, does always get updated and gets restored when returning to user mode, but I don't think it was intended, and it is inconsistent with the case of only the FPU being in use. Sometimes leaving the FPU enabled may also mask kernel bugs where FPU operations are executed when the FPU might not be enabled. So lets disable the FPU in the MSA case too. Fixes: 33c771ba ("MIPS: save/disable MSA in lose_fpu") Signed-off-by:
James Hogan <james.hogan@imgtec.com> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: Paul Burton <paul.burton@imgtec.com> Cc: linux-mips@linux-mips.org Patchwork: https://patchwork.linux-mips.org/patch/9323/Signed-off-by:
-
James Hogan authored
commit 98119ad5 upstream. Guest user mode can generate a guest MSA Disabled exception on an MSA capable core by simply trying to execute an MSA instruction. Since this exception is unknown to KVM it will be passed on to the guest kernel. However guest Linux kernels prior to v3.15 do not set up an exception handler for the MSA Disabled exception as they don't support any MSA capable cores. This results in a guest OS panic. Since an older processor ID may be being emulated, and MSA support is not advertised to the guest, the correct behaviour is to generate a Reserved Instruction exception in the guest kernel so it can send the guest process an illegal instruction signal (SIGILL), as would happen with a non-MSA-capable core. Fix this as minimally as reasonably possible by preventing kvm_mips_check_privilege() from relaying MSA Disabled exceptions from guest user mode to the guest kernel, and handling the MSA Disabled exception by emulating a Reserved Instruction exception in the guest, via a new handle_msa_disabled() KVM callback. Signed-off-by:
-
Christoffer Dall authored
commit ae705930 upstream. There is an interesting bug in the vgic code, which manifests itself when the KVM run loop has a signal pending or needs a vmid generation rollover after having disabled interrupts but before actually switching to the guest. In this case, we flush the vgic as usual, but we sync back the vgic state and exit to userspace before entering the guest. The consequence is that we will be syncing the list registers back to the software model using the GICH_ELRSR and GICH_EISR from the last execution of the guest, potentially overwriting a list register containing an interrupt. This showed up during migration testing where we would capture a state where the VM has masked the arch timer but there were no interrupts, resulting in a hung test. Cc: Marc Zyngier <marc.zyngier@arm.com> Reported-by:
Alex Bennee <alex.bennee@linaro.org> Signed-off-by:
Christoffer Dall <christoffer.dall@linaro.org> Signed-off-by:
Marc Zyngier <marc.zyngier@arm.com> Signed-off-by:
Shannon Zhao <shannon.zhao@linaro.org> Signed-off-by:
-
Marc Zyngier authored
commit 04b8dc85 upstream. The kernel's pgd_index macro is designed to index a normal, page sized array. KVM is a bit diffferent, as we can use concatenated pages to have a bigger address space (for example 40bit IPA with 4kB pages gives us an 8kB PGD. In the above case, the use of pgd_index will always return an index inside the first 4kB, which makes a guest that has memory above 0x8000000000 rather unhappy, as it spins forever in a page fault, whist the host happilly corrupts the lower pgd. The obvious fix is to get our own kvm_pgd_index that does the right thing(tm). Tested on X-Gene with a hacked kvmtool that put memory at a stupidly high address. Reviewed-by:
-
