kernel/linux.git/drivers/firmware, branch v4.19.39

efi/arm/arm64: Allow SetVirtualAddressMap() to be omitted

2019-04-05T20:33:12+00:00

[ Upstream commit 4e46c2a956215482418d7b315749fb1b6c6bc224 ] The UEFI spec revision 2.7 errata A section 8.4 has the following to say about the virtual memory runtime services: "This section contains function definitions for the virtual memory support that may be optionally used by an operating system at runtime. If an operating system chooses to make EFI runtime service calls in a virtual addressing mode instead of the flat physical mode, then the operating system must use the services in this section to switch the EFI runtime services from flat physical addressing to virtual addressing." So it is pretty clear that calling SetVirtualAddressMap() is entirely optional, and so there is no point in doing so unless it achieves anything useful for us. This is not the case for 64-bit ARM. The identity mapping used by the firmware is arbitrarily converted into another permutation of userland addresses (i.e., bits [63:48] cleared), and the runtime code could easily deal with the original layout in exactly the same way as it deals with the converted layout. However, due to constraints related to page size differences if the OS is not running with 4k pages, and related to systems that may expose the individual sections of PE/COFF runtime modules as different memory regions, creating the virtual layout is a bit fiddly, and requires us to sort the memory map and reason about adjacent regions with identical memory types etc etc. So the obvious fix is to stop calling SetVirtualAddressMap() altogether on arm64 systems. However, to avoid surprises, which are notoriously hard to diagnose when it comes to OS<->firmware interactions, let's start by making it an opt-out feature, and implement support for the 'efi=novamap' kernel command line parameter on ARM and arm64 systems. ( Note that 32-bit ARM generally does require SetVirtualAddressMap() to be used, given that the physical memory map and the kernel virtual address map are not guaranteed to be non-overlapping like on arm64. However, having support for efi=novamap,noruntime on 32-bit ARM, combined with the recently proposed support for earlycon=efifb, is likely to be useful to diagnose boot issues on such systems if they have no accessible serial port. ) Tested-by: Jeffrey Hugo Tested-by: Bjorn Andersson Tested-by: Lee Jones Signed-off-by: Ard Biesheuvel Cc: AKASHI Takahiro Cc: Alexander Graf Cc: Borislav Petkov Cc: Heinrich Schuchardt Cc: Leif Lindholm Cc: Linus Torvalds Cc: Matt Fleming Cc: Peter Jones Cc: Peter Zijlstra Cc: Sai Praneeth Prakhya Cc: Thomas Gleixner Cc: linux-efi@vger.kernel.org Link: http://lkml.kernel.org/r/20190202094119.13230-8-ard.biesheuvel@linaro.org Signed-off-by: Ingo Molnar Signed-off-by: Sasha Levin

efi/memattr: Don't bail on zero VA if it equals the region's PA

2019-04-05T20:33:11+00:00

[ Upstream commit 5de0fef0230f3c8d75cff450a71740a7bf2db866 ] The EFI memory attributes code cross-references the EFI memory map with the more granular EFI memory attributes table to ensure that they are in sync before applying the strict permissions to the regions it describes. Since we always install virtual mappings for the EFI runtime regions to which these strict permissions apply, we currently perform a sanity check on the EFI memory descriptor, and ensure that the EFI_MEMORY_RUNTIME bit is set, and that the virtual address has been assigned. However, in cases where a runtime region exists at physical address 0x0, and the virtual mapping equals the physical mapping, e.g., when running in mixed mode on x86, we encounter a memory descriptor with the runtime attribute and virtual address 0x0, and incorrectly draw the conclusion that a runtime region exists for which no virtual mapping was installed, and give up altogether. The consequence of this is that firmware mappings retain their read-write-execute permissions, making the system more vulnerable to attacks. So let's only bail if the virtual address of 0x0 has been assigned to a physical region that does not reside at address 0x0. Signed-off-by: Ard Biesheuvel Acked-by: Sai Praneeth Prakhya Cc: AKASHI Takahiro Cc: Alexander Graf Cc: Bjorn Andersson Cc: Borislav Petkov Cc: Heinrich Schuchardt Cc: Jeffrey Hugo Cc: Lee Jones Cc: Leif Lindholm Cc: Linus Torvalds Cc: Matt Fleming Cc: Peter Jones Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: linux-efi@vger.kernel.org Fixes: 10f0d2f577053 ("efi: Implement generic support for the Memory ...") Link: http://lkml.kernel.org/r/20190202094119.13230-4-ard.biesheuvel@linaro.org Signed-off-by: Ingo Molnar Signed-off-by: Sasha Levin

efi: cper: Fix possible out-of-bounds access

2019-04-05T20:33:04+00:00

[ Upstream commit 45b14a4ffcc1e0b5caa246638f942cbe7eaea7ad ] When checking a generic status block, we iterate over all the generic data blocks. The loop condition only checks that the start of the generic data block is valid (within estatus->data_length) but not the whole block. Because the size of data blocks (excluding error data) may vary depending on the revision and the revision is contained within the data block, ensure that enough of the current data block is valid before dereferencing any members otherwise an out-of-bounds access may occur if estatus->data_length is invalid. This relies on the fact that struct acpi_hest_generic_data_v300 is a superset of the earlier version. Also rework the other checks to avoid potential underflow. Signed-off-by: Ross Lagerwall Acked-by: Borislav Petkov Tested-by: Tyler Baicar Signed-off-by: Rafael J. Wysocki Signed-off-by: Sasha Levin

iscsi_ibft: Fix missing break in switch statement

2019-03-13T21:02:39+00:00

commit df997abeebadaa4824271009e2d2b526a70a11cb upstream. Add missing break statement in order to prevent the code from falling through to case ISCSI_BOOT_TGT_NAME, which is unnecessary. This bug was found thanks to the ongoing efforts to enable -Wimplicit-fallthrough. Fixes: b33a84a38477 ("ibft: convert iscsi_ibft module to iscsi boot lib") Cc: stable@vger.kernel.org Signed-off-by: Gustavo A. R. Silva Signed-off-by: Konrad Rzeszutek Wilk Signed-off-by: Greg Kroah-Hartman

x86/platform/UV: Use efi_runtime_lock to serialise BIOS calls

2019-02-20T09:25:48+00:00

commit f331e766c4be33f4338574f3c9f7f77e98ab4571 upstream. Calls into UV firmware must be protected against concurrency, expose the efi_runtime_lock to the UV platform, and use it to serialise UV BIOS calls. Signed-off-by: Hedi Berriche Signed-off-by: Borislav Petkov Reviewed-by: Ard Biesheuvel Reviewed-by: Russ Anderson Reviewed-by: Dimitri Sivanich Reviewed-by: Mike Travis Cc: Andy Shevchenko Cc: Bhupesh Sharma Cc: Darren Hart Cc: "H. Peter Anvin" Cc: Ingo Molnar Cc: linux-efi Cc: platform-driver-x86@vger.kernel.org Cc: stable@vger.kernel.org # v4.9+ Cc: Steve Wahl Cc: Thomas Gleixner Cc: x86-ml Link: https://lkml.kernel.org/r/20190213193413.25560-5-hedi.berriche@hpe.com Signed-off-by: Greg Kroah-Hartman

firmware: arm_scmi: provide the mandatory device release callback

2019-02-15T07:10:12+00:00

commit 46edb8d1322c1763dd04e179992f8e9996085047 upstream. The device/driver model clearly mandates that bus driver that discover and allocate the device must set the release callback. This callback will be used to free the device after all references have gone away. scmi bus driver is missing the obvious callback which will result in the following warning if the device is unregistered: Device 'scmi_dev.1' does not have a release() function, it is broken and must be fixed. See Documentation/kobject.txt. WARNING at drivers/base/core.c:922 device_release+0x8c/0xa0 Hardware name: ARM LTD Juno Development Platform BIOS EDK II Jan 21 2019 Workqueue: events deferred_probe_work_func pstate: 60000005 (nZCv daif -PAN -UAO) pc : device_release+0x8c/0xa0 lr : device_release+0x8c/0xa0 Call trace: device_release+0x8c/0xa0 kobject_put+0x8c/0x208 device_unregister+0x30/0x78 scmi_device_destroy+0x28/0x50 scmi_probe+0x354/0x5b0 platform_drv_probe+0x58/0xa8 really_probe+0x2c4/0x3e8 driver_probe_device+0x12c/0x148 __device_attach_driver+0xac/0x150 bus_for_each_drv+0x78/0xd8 __device_attach+0xe0/0x168 device_initial_probe+0x24/0x30 bus_probe_device+0xa0/0xa8 deferred_probe_work_func+0x8c/0xe0 process_one_work+0x1f0/0x478 worker_thread+0x22c/0x450 kthread+0x134/0x138 ret_from_fork+0x10/0x1c ---[ end trace 420bdb7f6af50937 ]--- Fix the issue by providing scmi_device_release callback. We have everything required for device release already in scmi_device_destroy, so we just need to move freeing of the device to scmi_device_release. Fixes: 933c504424a2 ("firmware: arm_scmi: add scmi protocol bus to enumerate protocol devices") Signed-off-by: Sudeep Holla Cc: stable@vger.kernel.org # 4.17+ Signed-off-by: Arnd Bergmann Signed-off-by: Greg Kroah-Hartman

firmware/efi: Add NULL pointer checks in efivars API functions

2019-02-12T18:47:04+00:00

[ Upstream commit ab2180a15ce54739fed381efb4cb12e78dfb1561 ] Since commit: ce2e6db554fa ("brcmfmac: Add support for getting nvram contents from EFI variables") we have a device driver accessing the efivars API. Several functions in the efivars API assume __efivars is set, i.e., that they will be accessed only after efivars_register() has been called. However, the following NULL pointer access was reported calling efivar_entry_size() from the brcmfmac device driver: Unable to handle kernel NULL pointer dereference at virtual address 00000008 pgd = 60bfa5f1 [00000008] *pgd=00000000 Internal error: Oops: 5 [#1] SMP ARM ... Hardware name: NVIDIA Tegra SoC (Flattened Device Tree) Workqueue: events request_firmware_work_func PC is at efivar_entry_size+0x28/0x90 LR is at brcmf_fw_complete_request+0x3f8/0x8d4 [brcmfmac] pc : [] lr : [] psr: a00d0113 sp : ede7fe28 ip : ee983410 fp : c1787f30 r10: 00000000 r9 : 00000000 r8 : bf2b2258 r7 : ee983000 r6 : c1604c48 r5 : ede7fe88 r4 : edf337c0 r3 : 00000000 r2 : 00000000 r1 : ede7fe88 r0 : c17712c8 Flags: NzCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none Control: 10c5387d Table: ad16804a DAC: 00000051 Disassembly showed that the local static variable __efivars is NULL, which is not entirely unexpected given that it is a non-EFI platform. So add a NULL pointer check to efivar_entry_size(), and to related functions while at it. In efivars_register() a couple of sanity checks are added as well. Reported-by: Jon Hunter Signed-off-by: Arend van Spriel Signed-off-by: Ard Biesheuvel Cc: Andy Lutomirski Cc: Bhupesh Sharma Cc: Borislav Petkov Cc: Dave Hansen Cc: Eric Snowberg Cc: Hans de Goede Cc: Joe Perches Cc: Julien Thierry Cc: Linus Torvalds Cc: Marc Zyngier Cc: Matt Fleming Cc: Nathan Chancellor Cc: Peter Zijlstra Cc: Sai Praneeth Prakhya Cc: Sedat Dilek Cc: Thomas Gleixner Cc: YiFei Zhu Cc: linux-efi@vger.kernel.org Link: http://lkml.kernel.org/r/20181129171230.18699-9-ard.biesheuvel@linaro.org Signed-off-by: Ingo Molnar Signed-off-by: Sasha Levin

efi/libstub: Disable some warnings for x86{,_64}

2019-01-26T08:32:36+00:00

[ Upstream commit 3db5e0ba8b8f4aee631d7ee04b7a11c56cfdc213 ] When building the kernel with Clang, some disabled warnings appear because this Makefile overrides KBUILD_CFLAGS for x86{,_64}. Add them to this list so that the build is clean again. -Wpointer-sign was disabled for the whole kernel before the beginning of Git history. -Waddress-of-packed-member was disabled for the whole kernel and for the early boot code in these commits: bfb38988c51e ("kbuild: clang: Disable 'address-of-packed-member' warning") 20c6c1890455 ("x86/boot: Disable the address-of-packed-member compiler warning"). -Wgnu was disabled for the whole kernel and for the early boot code in these commits: 61163efae020 ("kbuild: LLVMLinux: Add Kbuild support for building kernel with Clang") 6c3b56b19730 ("x86/boot: Disable Clang warnings about GNU extensions"). [ mingo: Made the changelog more readable. ] Tested-by: Sedat Dilek Signed-off-by: Nathan Chancellor Signed-off-by: Ard Biesheuvel Reviewed-by: Sedat Dilek Cc: Andy Lutomirski Cc: Arend van Spriel Cc: Bhupesh Sharma Cc: Borislav Petkov Cc: Dave Hansen Cc: Eric Snowberg Cc: Hans de Goede Cc: Joe Perches Cc: Jon Hunter Cc: Julien Thierry Cc: Linus Torvalds Cc: Marc Zyngier Cc: Matt Fleming Cc: Peter Zijlstra Cc: Sai Praneeth Prakhya Cc: Thomas Gleixner Cc: YiFei Zhu Cc: linux-efi@vger.kernel.org Link: http://lkml.kernel.org/r/20181129171230.18699-8-ard.biesheuvel@linaro.org Link: https://github.com/ClangBuiltLinux/linux/issues/112 Signed-off-by: Ingo Molnar Signed-off-by: Sasha Levin

efi/arm: Revert deferred unmap of early memmap mapping

2018-12-01T08:37:33+00:00

[ Upstream commit 33412b8673135b18ea42beb7f5117ed0091798b6 ] Commit: 3ea86495aef2 ("efi/arm: preserve early mapping of UEFI memory map longer for BGRT") deferred the unmap of the early mapping of the UEFI memory map to accommodate the ACPI BGRT code, which looks up the memory type that backs the BGRT table to validate it against the requirements of the UEFI spec. Unfortunately, this causes problems on ARM, which does not permit early mappings to persist after paging_init() is called, resulting in a WARN() splat. Since we don't support the BGRT table on ARM anway, let's revert ARM to the old behaviour, which is to take down the early mapping at the end of efi_init(). Signed-off-by: Ard Biesheuvel Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: linux-efi@vger.kernel.org Fixes: 3ea86495aef2 ("efi/arm: preserve early mapping of UEFI memory ...") Link: http://lkml.kernel.org/r/20181114175544.12860-3-ard.biesheuvel@linaro.org Signed-off-by: Ingo Molnar Signed-off-by: Sasha Levin

efi/libstub: arm: support building with clang

2018-12-01T08:37:24+00:00

commit 41f1c48420709470c51ee0e54b6fb28b956bb4e0 upstream. When building with CONFIG_EFI and CONFIG_EFI_STUB on ARM, the libstub Makefile would use -mno-single-pic-base without checking it was supported by the compiler. As the ARM (32-bit) clang backend does not support this flag, the build would fail. This changes the Makefile to check the compiler's support for -mno-single-pic-base before using it, similar to c1c386681bd7 ("ARM: 8767/1: add support for building ARM kernel with clang"). Signed-off-by: Alistair Strachan Reviewed-by: Stefan Agner Signed-off-by: Ard Biesheuvel Cc: Nick Desaulniers Signed-off-by: Greg Kroah-Hartman