summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* vboot: fix a typo in CL:2112321stabilize-13070.BJoel Kitching2020-04-301-1/+1
| | | | | | | | | | | | | | | | | CL:2112321 inadvertently removed the escape on a \n newline in a VB2_DEBUG string. Put it back. BUG=b:124141368, chromium:968464 TEST=make clean && make runtests BRANCH=none Change-Id: I19ee4a77d3cc58d6995212151b381ae76c602a42 Signed-off-by: Joel Kitching <kitching@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2174039 Reviewed-by: Hsuan Ting Chen <roccochen@chromium.org> Reviewed-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org> Tested-by: Joel Kitching <kitching@chromium.org>
* 2crypto: Force correct section flags for .rodata arraysJulius Werner2020-04-291-3/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | This is a follow-up fix to CL:2157900. That CL put the problematic arrays into sections with the right name, but they still didn't set the right ELF flags for them (specifically the READONLY flag). This made the assembler unhappy in coreboot, although it still seems to have built things correctly in the end: {standard input}: Assembler messages: {standard input}:359: Warning: setting incorrect section attributes for .rodata.vb2_hash_names {standard input}:369: Warning: setting incorrect section attributes for .rodata.vb2_sig_names This patch throws more ugly hacks at the problem to make it disappear. BRANCH=None BUG=None TEST=Build coreboot for both Arm and x86 boards, confirmed that error messages are gone and objdump shows intended section flags. Signed-off-by: Julius Werner <jwerner@chromium.org> Change-Id: Iea93788a13e9fd14d6b5a16626294d5a4b0e5411 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2169480 Reviewed-by: Joel Kitching <kitching@chromium.org>
* vboot: Convert reboot-related errors to vboot2-styleYu-Ping Wu2020-04-2814-199/+220
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Error codes are renamed as follows: VBERROR_SHUTDOWN_REQUESTED --> VB2_REQUEST_SHUTDOWN VBERROR_REBOOT_REQUIRED --> VB2_REQUEST_REBOOT VBERROR_EC_REBOOT_TO_SWITCH_RW --> VB2_REQUEST_REBOOT_EC_SWITCH_RW VBERROR_EC_REBOOT_TO_RO_REQUIRED --> VB2_REQUEST_REBOOT_EC_TO_RO VBERROR_KEEP_LOOPING --> VB2_REQUEST_UI_CONTINUE Since these are not really "errors", they're renamed to VB2_REQUEST_* instead, with larger values to separate from real errors. The old vboot1-style errors will be removed once the errors are renamed in coreboot and depthcharge. BRANCH=none BUG=b:124141368, chromium:988410 TEST=make runtests Change-Id: Ide1328f9c0f2803534fb3f1f1ecd16ad8ea36a4e Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2143030 Reviewed-by: Julius Werner <jwerner@chromium.org> Reviewed-by: Joel Kitching <kitching@chromium.org>
* crossystem: add support for ADL gpiochipSubrata Banik2020-04-281-0/+3
| | | | | | | | | | | | | | On Alderlake platform, the pinctrl (gpiochip) driver label is "INTC105x:00", hence declare it properly. TEST=run 'crossystem wpsw_cur' and see '0' rather than an error on ADLRVP platform. Change-Id: I7c8ead83717b7be35ce68026768f6f0f5b1efa1c Signed-off-by: Subrata Banik <subrata.banik@intel.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2167671 Reviewed-by: Furquan Shaikh <furquan@chromium.org> Tested-by: Subrata Banik <subrata.banik@intel.corp-partner.google.com>
* crossystem: Add quirk reading wpsw_curLukasz Majczak2020-04-231-1/+46
| | | | | | | | | | | | | | | | | | | | | | | | As GPIO numbering has changed with kernel 4.16, so we have to change the finall offset when reading wpsw_cur. Before there was a gap (3.18): 0: INT33FF:02 GPIOS [373 - 384] PINS [0 - 11] and 12: INT33FF:02 GPIOS [385 - 396] PINS [15 - 26]a with 4.19 it looks like below: 0: INT33FF:02 GPIOS [314 - 325] PINS [0 - 11] and 12: INT33FF:02 GPIOS [329 - 340] PINS [15 - 26] BUG=b:143174998 BRANCH=none TEST=crossystem wpsw_cur (on Kefka) 1. with wp present it should return 1 2. without present wp it should return 0 Change-Id: I7c8ead83717b7be35ce68026768f6f0f5b0ff91c Signed-off-by: Lukasz Majczak <lma@semihalf.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2153155 Reviewed-by: Brian Norris <briannorris@chromium.org> Commit-Queue: Brian Norris <briannorris@chromium.org>
* vboot: Add screens for recovery using diskYu-Ping Wu2020-04-221-0/+4
| | | | | | | | | | | | | Add VB2_SCREEN_RECOVERY_DISK_STEP* to enum vb2_screen. BRANCH=none BUG=b:146399181 TEST=USE="menu_ui" emerge-nami depthcharge Change-Id: Ibfaeb25d00ee340adfb9f7de9188f20a947bae24 Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2156701 Reviewed-by: Joel Kitching <kitching@chromium.org>
* vboot: Add broken screen to vb2_screenYu-Ping Wu2020-04-221-0/+2
| | | | | | | | | | | | | Add VB2_SCREEN_RECOVERY_BROKEN to enum vb2_screen for broken screen. BRANCH=none BUG=b:146399181 TEST=USE="menu_ui" emerge-nami depthcharge Change-Id: I7f535094de216157591eab9d17a8e4aa2fdd772e Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2156699 Reviewed-by: Joel Kitching <kitching@chromium.org>
* vboot: Add recovery select screen to vb2_screenYu-Ping Wu2020-04-223-10/+23
| | | | | | | | | | | | | | | | | | | Add VB2_SCREEN_RECOVERY_SELECT to enum vb2_screen for the recovery select screen. Also add arguments 'selected_item' and 'disabled_item_mask' to vb2ex_display_ui() for menu items, and rename argument 'locale' to 'locale_id'. BRANCH=none BUG=b:146399181 TEST=USE="menu_ui" emerge-nami depthcharge Cq-Depend: chromium:2124025 Change-Id: Ifff67ded8196288cdb3c1558c64c09b7ffc5011e Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2138876 Reviewed-by: Joel Kitching <kitching@chromium.org>
* 2crypto: Force exported const arrays to .rodataJulius Werner2020-04-221-1/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | Unfortunately, there is no way to define really immutable data in the C language. Making something 'const' is just a sign, not a cop, because in theory any code can just cast the const away. Compilers instead use sophisticated static analysis to determine what data is never written to by the code and put that into the .rodata (instead of .data) section. But for externally available global variables, they cannot do that (because the analyzer doesn't cross compilation units). Unfortunately some platforms in coreboot really care about everything being in .rodata (because they're accessing it in-place on flash), and some "benign" refactoring like making a const variable externally available can break that. There is no easy fix for that. So in this case, let's just go for the ugly fix of forcing the offending arrays into the right section. BRANCH=None BUG=None TEST=Built SIEMENS_MC_APL2 with CB:40503 and this patch. Signed-off-by: Julius Werner <jwerner@chromium.org> Change-Id: Ife508f98e3f8ed40a4488b5fe1967d00a62f347c Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2157900 Reviewed-by: Daisuke Nojiri <dnojiri@chromium.org>
* 2api: Add kernel version getter and make dev-signed check fail softJulius Werner2020-04-213-1/+15
| | | | | | | | | | | | | | | | | | | | | | | This patch adds a function to retrieve the kernel rollback version from secdata, which may be interesting to callers that have more advanced ways of retrieving bootable images and want to be able to spot check whether an image can be booted without passing the full thing to vboot. Also reduce the penalty from calling vb2api_is_developer_signed() out of turn from an immediate DIE() to an angry error message, to support a case in depthcharge where for all practical purposes the call should never happen too early, but the framework can't quite guarantee it. BRANCH=None BUG=b:153758197 TEST=None Signed-off-by: Julius Werner <jwerner@chromium.org> Change-Id: Ic7c2fc62e1ba80f69f70421907b9686f0b3dae77 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2153592 Reviewed-by: Jes Klinke <jbk@chromium.org> Reviewed-by: Joel Kitching <kitching@chromium.org> Tested-by: Jes Klinke <jbk@chromium.org>
* vboot: Add vb2api_phone_recovery_enabled() to vboot APIYu-Ping Wu2020-04-209-12/+111
| | | | | | | | | | | | | | | | | | | In order to implement a kill switch for phone recovery in recovery UI, add VB2_SECDATA_KERNEL_FLAGS to enum vb2_secdata_kernel_flags, and add VB2_SECDATA_KERNEL_FLAG_PHONE_RECOVERY_DISABLED to the flags to indicate whether to disable phone recovery. Set this flag to 1 in vb2api_kernel_phase1(). Also add vb2api_phone_recovery_enabled() to vboot API. BRANCH=none BUG=b:146399181, b:147744345 TEST=make runtests Change-Id: I65c8faa593f5f90dd762f0092058048983c5b3a2 Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2131236 Reviewed-by: Joel Kitching <kitching@chromium.org>
* vboot: move recovery reason strings to vboot2 namespaceJoel Kitching2020-04-208-128/+257
| | | | | | | | | | | | | | | | | | | | Make recovery reason strings more widely available, and print them in kernel verification console messages. Note that we purposely do not print strings in firmware verification in order to avoid linking recovery reason strings into coreboot stages. BUG=b:124141368, chromium:968464 TEST=make clean && make runtests TEST=trigger recovery, check console logs BRANCH=none Change-Id: Ica82a60075da53e4ad0a0403eb8a88685c3bdbd6 Signed-off-by: Joel Kitching <kitching@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2112321 Reviewed-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org> Tested-by: Joel Kitching <kitching@chromium.org>
* secdata: Initialize secdata kernel in vb2api_fw_phase1dnojiri2020-04-186-36/+24
| | | | | | | | | | | | | | | | | | | | | | | Currently, secdata kernel is initialized in vb2api_kernel_phase1. Since we'll be using secdata kernel in romstage for EFS2, it needs to be initialized earlier. This patch makes vb2api_fw_phase1 call vb2_secdata_kernel_init. Signed-off-by: dnojiri <dnojiri@chromium.org> BUG=b:147298634, chromium:1045217, b:148259137 BRANCH=none TEST=Verify Hmir is updated in romstage on Puff. Cq-Depend: chromium:2155377 Change-Id: I8a537126a952ecccc9cc496ce60fc6dc81541a8f Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2153852 Reviewed-by: Julius Werner <jwerner@chromium.org> Reviewed-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Daisuke Nojiri <dnojiri@chromium.org> Tested-by: Daisuke Nojiri <dnojiri@chromium.org> Auto-Submit: Daisuke Nojiri <dnojiri@chromium.org>
* vboot: remove some trailing spacesJoel Kitching2020-04-173-4/+4
| | | | | | | | | | | | | | BUG=b:124141368 TEST=none BRANCH=none Change-Id: I054e668d96e1c76d6ce464e182fd05d17ef3fca0 Signed-off-by: Joel Kitching <kitching@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2154342 Reviewed-by: Joel Kitching <kitching@chromium.org> Reviewed-by: Yu-Ping Wu <yupingso@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org> Tested-by: Joel Kitching <kitching@chromium.org>
* vboot: decouple EC/AUXFW sync from UIYu-Ping Wu2020-04-177-176/+81
| | | | | | | | | | | | | | | | | | | | | | | | | Since we don't always want to show a UI on EC sync (for example, in coreboot, where display hasn't been initialized), decouple vb2api_ec_sync() from VbDisplayScreen() by leaving screen display out of vboot and letting the caller (such as depthcharge) handle it. Similarly, stop calling screen display function from vb2api_auxfw_sync(). In order to display screen from depthcharge, it needs to know the locale. Therefore, add vb2api_get_locale() to vboot API, which returns the locale from nvdata. After this change, the constant EC_SLOW_UPDATE is no longer used, so remove it from Makefile. BRANCH=none BUG=chromium:1055125 TEST=make runtests Cq-Depend: chromium:2117776 Change-Id: I0e2e8ebdd26d48a2e94d36495c2e45a5734cdc5d Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2087016 Reviewed-by: Joel Kitching <kitching@chromium.org>
* vboot: remove VBERROR_TPM_SET_BOOT_MODE_STATE error codeJoel Kitching2020-04-178-40/+48
| | | | | | | | | | | | | | | | | | | | | | Since secdata and nvdata get/set functions no longer return error codes, and instead use VB2_ASSERT and VB2_DIE to abort on failure, vb2_enable_developer_mode no longer has any error code to return. Change the function return type to void, and remove checks around the function call. As a result, VBERROR_TPM_SET_BOOT_MODE_STATE becomes unused and we may remove it. Finally, move the USB_BOOT_ON_DEV logic (enable USB boot when on transition to dev mode) into vb2_enable_developer_mode. Also add unit tests. BUG=b:124141368, chromium:988410 TEST=make clean && make runtests BRANCH=none Change-Id: I286d9343c4c751ff24bf4c149a26fbe5306e383a Signed-off-by: Joel Kitching <kitching@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2152212 Reviewed-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org> Tested-by: Joel Kitching <kitching@chromium.org>
* updater_utils.h: stop using 'dev=1' in flashrom commandNikolai Artemiev2020-04-171-1/+1
| | | | | | | | | | | | | | | Replace 'ec:dev=1' with equivalent 'ec:type=pd'. BUG=b:153024532 BRANCH=none TEST=emerge vboot_reference Change-Id: Ic8df99fcdc3769bec19054b1936b97b14bade0e7 Signed-off-by: Nikolai Artemiev <nartemiev@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2151492 Reviewed-by: Hung-Te Lin <hungte@chromium.org> Reviewed-by: Edward O'Callaghan <quasisec@chromium.org> Tested-by: Edward O'Callaghan <quasisec@chromium.org>
* vboot: Add VB2_ERROR_UI_MEMORY_ALLOC to vb2_error_tYu-Ping Wu2020-04-161-0/+3
| | | | | | | | | | | | | | Add VB2_ERROR_UI_MEMORY_ALLOC for memory allocation failure in UI. BRANCH=none BUG=b:146399181 TEST=emerge-nami vboot_reference Change-Id: Ib6d8ccbfe73f90297f400430b4ae1204adbf7ee4 Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2152331 Reviewed-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org>
* vboot: remove more unused pre-vboot1 and vboot1 error codesJoel Kitching2020-04-161-25/+0
| | | | | | | | | | | | | | BUG=b:124141368, chromium:988410 TEST=make clean && make runtests BRANCH=none Change-Id: I24c867af25726f84a4ada18477de66579975a383 Signed-off-by: Joel Kitching <kitching@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2150182 Tested-by: Joel Kitching <kitching@chromium.org> Reviewed-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org>
* vboot: Remove VB2_SCREEN_TOUCHPAD_SYNC from vb2_screenYu-Ping Wu2020-04-141-2/+0
| | | | | | | | | | | | | | | The touchpad sync screen is rendered in kernel instead of firmware, so remove it from enum vb2_screen. BRANCH=none BUG=none TEST=USE="menu_ui" emerge-nami depthcharge Change-Id: I5929ecec51bee1aa36d6abf04bb733fde46ac29c Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2148552 Reviewed-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org>
* vboot: clear recovery request in all boot modesJoel Kitching2020-04-143-11/+33
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, recovery requests are only cleared when user initiates a manual recovery. This causes problems with two cases specifically: * Transient failures - The recovery request remains in the subcode field for some unknown period of time, and then erroneously gets promoted to the "recovery reason" the next time the user initiates a manual recovery request. * TRAIN_AND_REBOOT - The recovery request remains in the subcode field after training has completed. The next time a manual recovery request is initiated, the subcode is promoted and training occurs yet again. When finished, a reboot occurs and the user ends up back in the OS. Make two changes to deal with these cases: * Clear recovery request (including subcode) unconditionally for non-recovery boot modes. * Stop promoting TRAIN_AND_REBOOT subcodes. BUG=b:153157134, b:35576380 TEST=make clean && make runtests BRANCH=none Change-Id: I79f8fbed72a9d052b5ed5f70e9a2515136b6ef10 Signed-off-by: Joel Kitching <kitching@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2139335 Tested-by: Joel Kitching <kitching@chromium.org> Tested-by: Frank Wu <frank_wu@compal.corp-partner.google.com> Reviewed-by: Julius Werner <jwerner@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org>
* vboot: Add a leading space before labelsYu-Ping Wu2020-04-092-2/+2
| | | | | | | | | | | | | | The following command is run: grep -rlE '^\w+:' firmware | xargs -l sed -i -E 's/^\w+:/ \0/g' BRANCH=none BUG=none TEST=emerge-nami vboot_reference Change-Id: I4dedd1107e71add07f51e32925757680e780c882 Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2143017 Reviewed-by: Joel Kitching <kitching@chromium.org>
* vboot: Add VB2_ERROR_UI_INVALID_LOCALE to vb2_error_tYu-Ping Wu2020-04-071-0/+3
| | | | | | | | | | BRANCH=none BUG=b:146399181 TEST=USE="menu_ui" emerge-nami depthcharge Change-Id: I8c0e37adc4f37900a5e426f24a581befb07c6a34 Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2137152
* vboot: Add firmware sync screens to vb2_screenYu-Ping Wu2020-04-071-1/+5
| | | | | | | | | | | | | Add VB2_SCREEN_FIRMWARE_SYNC and VB2_SCREEN_TOUCHPAD_SYNC screens to enum vb2_screen. BRANCH=none BUG=b:146399181 TEST=USE="menu_ui" emerge-nami depthcharge Change-Id: Iddcd77bb79be7a29346d82b70a8dbb82d0b33fcb Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2137151
* Rewrite algorithm type parsers and make them globally availableJulius Werner2020-04-0618-353/+460
| | | | | | | | | | | | | | | | | | | | | There is some code strewn around between futility and the vb21-specific part of hostlib to allow parsing of textual algorithm names to vboot enums, but it is somewhat disorganized and not written in a super efficient way. This patch rewrites it and centralizes all the algorithm mapping stuff under 2crypto.c so it can be a single source of truth for all of vboot. (String parsing routines still need to stay in hostlib since not all firmware targets support things like stroul() and strcasecmp().) BRANCH=None BUG=None TEST=make runtests Signed-off-by: Julius Werner <jwerner@chromium.org> Change-Id: I719b2499992a6e4395a29231bc8b9a7680c5b174 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2099447 Reviewed-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org>
* 2secdata_kernel: Fix debug messagesstabilize-sylas-13019.Bdnojiri2020-04-021-3/+3
| | | | | | | | | | | BUG=none BRANCH=none TEST=Verified on Puff. Signed-off-by: dnojiri <dnojiri@chromium.org> Change-Id: I7e33abcf0832f83530e94cc37117c03fe164bbdb Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2132716 Reviewed-by: Julius Werner <jwerner@chromium.org>
* ec_sync: Ignore NO_BOOT when GBB says DISABLE_EC_SOFTWARE_SYNCdnojiri2020-04-012-5/+22
| | | | | | | | | | | | | | | | | | | | | | | | | | Currently, there is a final guard in VbSelectAndLoadKernel to enforce NO_BOOT unconditionally. This makes a board with VB2_GBB_FLAG_DISABLE_EC_SOFTWARE_SYNC unable to boot if Hmir != Heff because Hmir will never be synced. If a system doesn't boot, set_gbb_flags.sh can't be used to enable software sync. This patch skips the final guard if VB2_GBB_FLAG_DISABLE_EC_SOFTWARE_SYNC is set. Signed-off-by: dnojiri <dnojiri@chromium.org> BUG=chromium:1045217 BRANCH=none TEST=make runtests Change-Id: I0c7f8c90cb0a6bb54ce30ef42e4d62554decb872 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2129107 Reviewed-by: Julius Werner <jwerner@chromium.org> Tested-by: Daisuke Nojiri <dnojiri@chromium.org> Commit-Queue: Daisuke Nojiri <dnojiri@chromium.org> Auto-Submit: Daisuke Nojiri <dnojiri@chromium.org>
* ec_sync: Reset EC after updating Hmirdnojiri2020-04-013-0/+27
| | | | | | | | | | | | | | | | | | | | | | This patch makes vb2api_sync_ec reset EC whenever Hmir is update because the verification result should be revoked. When Hmir is corrupted, the system enters NO_BOOT mode (and EC stays in RO). Since Heff == Hexp, EC RW won't be updated but we need to reset EC so that EC can jump to RW by EFS2. Signed-off-by: dnojiri <dnojiri@chromium.org> BUG=chromium:1045217 BRANCH=none TEST=Verify firmware_SoftwareSync passes. Change-Id: I21887e212cfefedeadf4c146ecec290022bb2b34 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2122587 Tested-by: Daisuke Nojiri <dnojiri@chromium.org> Reviewed-by: Julius Werner <jwerner@chromium.org> Commit-Queue: Daisuke Nojiri <dnojiri@chromium.org> Auto-Submit: Daisuke Nojiri <dnojiri@chromium.org>
* vboot: Add a blinking cursor to the vendor promptMathew King2020-03-313-2/+35
| | | | | | | | | | | | | | | Add support to show a blinking cursor after the vendor data entry so that users can tell that an entry is required. BUG=b:138812835 TEST=Test on a Drallion device with an updated Depthcharge BRANCH=drallion Cq-Depend: chromium:2120311 Change-Id: I02073d92d02a4171f337f86f880d660eb5d8b6f1 Signed-off-by: Mathew King <mathewk@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2120310 Reviewed-by: Pranay Shoroff <pshoroff@google.com>
* vboot: stop using StrnAppend and Uint64ToStringJoel Kitching2020-03-3140-393/+66
| | | | | | | | | | | | | | | | | | | Use snprintf instead. Remove utility_string library. Also, prepare VbDisplayDebugInfo to handle 64-byte nvdata. BUG=b:124141368, chromium:968464 TEST=make clean && make runtests TEST=boot with 16-byte nvdata, check <TAB> output (one line) TEST=boot with 64-byte nvdata, check <TAB> output (five lines) BRANCH=none Change-Id: If6c4b3a4e9fa7b71cb2d8ca7ccd37e4f36b97fd6 Signed-off-by: Joel Kitching <kitching@google.com> Cq-Depend: chromium:2122061 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2104880 Tested-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org> Reviewed-by: Julius Werner <jwerner@chromium.org>
* firmware: Add VB2_TRY() helper macroYu-Ping Wu2020-03-3117-284/+189
| | | | | | | | | | | | | | | | | | | | | | | Add variadic macro VB2_TRY() to vboot2 APIs, which supports the following usage: - VB2_TRY(func_call()) - VB2_TRY(func_call(), ctx, recovery_reason) and will return the error code if func_call() failed. Also utilize the macro whenever possible throughout vboot. BRANCH=none BUG=chromium:1049032 TEST=make runtests TEST=emerge-nami coreboot coreboot-utils depthcharge -j Cq-Depend: chromium:2115423, chromium:2125616, chrome-internal:2817320 Change-Id: Ie5532a8beaa9372fa6fde0a68bda5ecb640087c4 Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2084062 Reviewed-by: Julius Werner <jwerner@chromium.org>
* updater: Add Drallion to the wilco quirk listDuncan Laurie2020-03-311-0/+2
| | | | | | | | | | | | | | | Add the Drallion board to the list of sarien/arcada so it gets the workaround to lock the CSME on dogfood units safely. BUG=b:152475734 BRANCH=none TEST=none Signed-off-by: Duncan Laurie <dlaurie@google.com> Change-Id: I2119f64d5935d3462f0f7ad27d255ec259733553 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2127313 Reviewed-by: Tim Wawrzynczak <twawrzynczak@chromium.org> Commit-Queue: Tim Wawrzynczak <twawrzynczak@chromium.org>
* Makefile: Explicitly signal success at the end of 'make runtests'Julius Werner2020-03-271-0/+3
| | | | | | | | | | | | | | | | | | | | | | | If you run 'make runtests' with -j (as you should because it takes forever), it can be hard to see if it passed successfully. A 'make: *** [Makefile:XXX: ...] Error 255' line might hide somewhere in the middle with a lot of successful output from other tests running in parallel hiding it. It's hard (I think?) to have make output something at the end (after joining all processes) when it has encountered an error, but at least it is easy to output something when the 'runtests' target finishes successfully. This patch adds such a highly visible success message, so that its absence will make it clear that there was an error further up. BRANCH=None BUG=None TEST='make runtests', both passing and failing Signed-off-by: Julius Werner <jwerner@chromium.org> Change-Id: If979a7635f5c05c44a92daca12d31e344563794c Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2116750 Reviewed-by: Joel Kitching <kitching@chromium.org>
* Makefile: Pull ebuild special cases into install targetsJulius Werner2020-03-271-14/+17
| | | | | | | | | | | | | | | | | | | | | | The install targets in the vboot Makefile are currently missing a few things that get manually fixed up by the ebuild (e.g. 'make install' installs the package-config files for hostlib but not the hostlib itself, which makes absolutely no sense). This isn't great because we have to update both sides every time we make changes there. This patch cleans up the install targets to handle those cases directly so they can be removed from the ebuild. BRANCH=None BUG=None TEST=Compared output of 'equery f vboot_reference' before and after, for both SDK and board builds. Signed-off-by: Julius Werner <jwerner@chromium.org> Change-Id: Ibacec8f5fd3983fc3e9040da01dcedb5d4d1a61b Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2113712 Reviewed-by: Mike Frysinger <vapier@chromium.org> Reviewed-by: Joel Kitching <kitching@chromium.org>
* vboot: Fixes bug in get_gbb_flags.sh -eRob Barnes2020-03-241-1/+1
| | | | | | | | | | | | | | | | get_gbb_flags.sh outputs incorrect information because the hex number is not parsed correctly. BUG=none TEST=Manual BRANCH=none Change-Id: Ie6428a5c50d48ae5d732b31d7a8e7b314653c2d9 Signed-off-by: Rob Barnes <robbarnes@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2108286 Reviewed-by: Joel Kitching <kitching@chromium.org> Reviewed-by: Jack Rosenthal <jrosenth@chromium.org> Commit-Queue: Jack Rosenthal <jrosenth@chromium.org>
* vboot: Implement dev mode boot options for menu UIstabilize-12997.BHsuan Ting Chen2020-03-235-3/+416
| | | | | | | | | | | | | | | | | | | | | | | Add dev default boot retrieval and dev boot related allowance check api in 2misc along with their tests in vb2_misc_tests: - vb2_get_dev_boot_target - vb2_dev_boot_allowed - vb2_dev_boot_legacy_allowed - vb2_dev_boot_usb_allowed Implement parts of vb2_developer_menu functionalities along with tests: - Retrieve the default boot option. - Try to boot with the specified option. BRANCH=none BUG=b:146399181, chromium:1033815 TEST=USE="menu_ui" emerge-nami vboot_reference depthcharge TEST=make runtests Change-Id: Ie82076f93b86ba5abe26a9e3e25076892684855d Signed-off-by: Hsuan Ting Chen <roccochen@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2094508 Reviewed-by: Joel Kitching <kitching@chromium.org>
* vboot: remove compression_type constantsJoel Kitching2020-03-231-8/+0
| | | | | | | | | | | | | | | | These are no longer needed after CL:367882. BUG=b:124141368, chromium:968464 TEST=make clean && make runtests BRANCH=none Change-Id: I267d20054778a6565dc1ed0687ed0d2afa928463 Signed-off-by: Joel Kitching <kitching@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2108752 Tested-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org> Reviewed-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-by: Julius Werner <jwerner@chromium.org>
* vboot: remove VbExRegionRead vestigesJoel Kitching2020-03-231-7/+0
| | | | | | | | | | | | | | | | API was originally removed in CL:852798. BUG=b:124141368, chromium:968464 TEST=make clean && make runtests BRANCH=none Change-Id: I29c0b642dbac80b5c8c7eceaf108843467a90be4 Signed-off-by: Joel Kitching <kitching@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2108751 Tested-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org> Reviewed-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-by: Julius Werner <jwerner@chromium.org>
* vboot: Introduce menu UIHsuan Ting Chen2020-03-216-3/+164
| | | | | | | | | | | | | | | | | | | | | | | | | According to crbug.com/1033815, add files for MENU_UI: - 2ui.{c,h} for entry point functions called by kernel selection. - vb2_ui_tests.c for tests (normal boot included). There are three entry points: vb2_developer_menu(), vb2_broken_recovery_menu(), and vb2_manual_recovery_menu(). Only infinite loop for this CL, others will be added in separate CLs. BRANCH=none BUG=b:146399181,chromium:1033815 TEST=USE="legacy_clamshell_ui" emerge-nami vboot_reference depthcharge TEST=USE="legacy_menu_ui" emerge-nami vboot_reference depthcharge TEST=USE="menu_ui" emerge-nami vboot_reference depthcharge TEST=make runtests Cq-Depend: chromium:2043102 Change-Id: I5a23eb006754b5ff08eb42bcd5021374995eab40 Signed-off-by: Hsuan Ting Chen <roccochen@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2087557 Reviewed-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Yu-Ping Wu <yupingso@chromium.org>
* vboot: Fix fuzzer buildManoj Gupta2020-03-191-0/+1
| | | | | | | | | | | | | | | | | | | | LLD is complaining about the symbol vb2_check_padding present in a discarded section. This is caused by multiple defintions of vb2_check_padding being present (fuzzers & 2rsa.c) and the definition in 2rsa.c also being used in the same file. Fix by marking vb2_check_padding as weak. BUG=chromium:1062665 TEST=Fuzzer builds work BRANCH=none Signed-off-by: Manoj Gupta <manojgupta@google.com> Change-Id: Id1e2b69cdb35cbd127ba57881a70fc352ea2450a Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2108483 Tested-by: Manoj Gupta <manojgupta@chromium.org> Reviewed-by: Julius Werner <jwerner@chromium.org> Commit-Queue: Manoj Gupta <manojgupta@chromium.org>
* Avoid zero-size VLA in cmd_dump_fmap.cinglorion2020-03-181-1/+2
| | | | | | | | | | | | | | | | | Zero-size variable length arrays are undefined behavior. UBSan reported a zero-size VLA in futility/cmd_dump_fmap.c when running tests. This change fixes that by making sure the size of the VLA is at least 1. BUG=chromium:1058086 TEST=FEATURES="test" USE="llvm-next ubsan" emerge -e --nodeps vboot_reference BRANCH=none Change-Id: I3c0eee0359668ce2e1ec3c0724f243788161fb13 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2103436 Tested-by: Bob Haarman <inglorion@chromium.org> Reviewed-by: Manoj Gupta <manojgupta@chromium.org> Reviewed-by: Julius Werner <jwerner@chromium.org> Commit-Queue: Manoj Gupta <manojgupta@chromium.org>
* host: add host library for accessing chromeos-configrelease-R82-12974.BJack Rosenthal2020-03-134-0/+325
| | | | | | | | | | | | | | | | | | | | Host side library for accessing chromeos-config. Initially, this will be used by futility to access the /firmware:image-name property during a firmware update. More background: go/mosys-firmware-name (note: despite the name "mosys" in the design doc, this is an effort to *not* rely on mosys during the firmware update) BUG=chromium:1061192 BRANCH=none TEST=provided unit tests Signed-off-by: Jack Rosenthal <jrosenth@chromium.org> Change-Id: Ib8e5f8f836a93695e3b30731ae227501f37c4633 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2099449 Reviewed-by: Julius Werner <jwerner@chromium.org>
* vboot: stop reading from ACPI for wpsw_bootJoel Kitching2020-03-111-4/+0
| | | | | | | | | | | | | | | | Stop reading wpsw_boot value from ACPI as fallback for wpsw_cur. wpsw_cur is available on all non-EOL'd devices. BUG=b:124141368, chromium:950273, chromium:1059670 TEST=make clean && make runtests BRANCH=none Change-Id: I9c6f21b9f341a61e85db3e6e45b553b1ab4293e3 Signed-off-by: Joel Kitching <kitching@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2094807 Reviewed-by: Julius Werner <jwerner@chromium.org> Tested-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org>
* vboot: Move vb2_normal_boot tests to vb2_kernel_tests.cHsuan Ting Chen2020-03-103-41/+53
| | | | | | | | | | | | | | Move vb2_normal_boot tests from legacy clamshell UI tests to 2kernel tests. BUG=b:124141368, chromium:968464 TEST=make clean && make runtests BRANCH=none Change-Id: Iaf05dfa845222cd5eaf5aa39afcfa23f0ef21425 Signed-off-by: Hsuan Ting Chen <roccochen@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2091152 Reviewed-by: Joel Kitching <kitching@chromium.org>
* vboot: create 2kernel.c for kernel-related functionsJoel Kitching2020-03-0910-308/+501
| | | | | | | | | | | | | | | | | | | | | | | Create 2kernel.c for kernel-related functions (including both internal and API-facing). Relocate functions which are currently in active use from lib20/api_kernel.c to 2lib/2kernel.c. Create vb2_kernel_tests.c and move appropriate tests there. Tests for vb2_normal_boot (previously VbBootNormal) will be added in a subsequent CL. BUG=b:124141368, chromium:968464 TEST=make clean && make runtests BRANCH=none Change-Id: I153ea318ba238b7b30bf871105388437f63075d1 Signed-off-by: Joel Kitching <kitching@google.com> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2091126 Reviewed-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org> Tested-by: Joel Kitching <kitching@chromium.org>
* vboot: stop setting VBSD_BOOT_FIRMWARE_WP_ENABLEDJoel Kitching2020-03-062-5/+2
| | | | | | | | | | | | | | | | | | wpsw_boot is being deprecated in favour of using wpsw_cur. Remove the wp_enabled argument from vb2api_export_vbsd and stop setting VBSD_BOOT_FIRMWARE_WP_ENABLED in VBSD. BUG=b:124141368, chromium:950273 TEST=make clean && make runtests BRANCH=none Change-Id: Ib3c7f43b5fb2af519bca93d5bbd95db5eff241a3 Signed-off-by: Joel Kitching <kitching@google.com> Cq-Depend: chromium:2088434 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2087141 Reviewed-by: Julius Werner <jwerner@chromium.org> Tested-by: Joel Kitching <kitching@chromium.org> Commit-Queue: Joel Kitching <kitching@chromium.org>
* Revert "vboot: Add screen types for upcoming UI revamp"Yu-Ping Wu2020-03-061-22/+0
| | | | | | | | | | | | | | | | | This reverts commit 1df08718b7c18a0454240f78bdb80359116fc295. CL:1968268 added screen types to enum VbScreenType_t for menu UI, but we later decided to create a separate enum for that as part of vboot2 APIs. Hence, remove these screen types that are not used in legacy UIs. BRANCH=none BUG=b:146399181 TEST=emerge-nami vboot_reference Change-Id: Id86fcac257894148c82d54d27a8720ecb6b23af6 Signed-off-by: Yu-Ping Wu <yupingso@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2089049 Reviewed-by: Joel Kitching <kitching@chromium.org>
* EFS: Implement EFS2 and NO_BOOT modeDaisuke Nojiri2020-03-0510-48/+196
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | EFS v1 allowed Chromeboxes to verify RW without AP. EFS v2 will bring the benefts to Chromebooks, which are: - Reduce RO dependency and presence. Allow more code to be updated in the fields. - Remove jumptag and workarounds needed for late sysjump. Major imporvements over v1 are: - No A/B slot required. - No signature in RW or public key in RO. - Rollback-attack protection. For battery-equipped devices, additional benefts are: - Immediate boot on drained battery. - Support recovery mode regardless of battery condition. - Faster charge in S5/G3. EC-Cr50 communication is based on the shared UART (go/ec-cr50-comm). EFS2 is documented in go/ec-efs2. BUG=chromium:1020578,chromium:1045217 TEST=Boot Helios in NORMAL/RECOVERY/NO_BOOT mode. TEST=Verify EC is updated by software sync in Depthcharge. Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org> Change-Id: Ie07e6f6ce46c0955a6a0adf595633e65c4ffe724 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/1901868
* Add vb2_secdata_kernel_v10dnojiri2020-03-059-74/+504
| | | | | | | | | | | | | This patch adds vb2_secdata_kernel_v10. It has ec_hash field to support EFS2. The secdata kernel APIs continue to support v0.2. BUG=chromium:1045217 TEST=emerge-hatch depthcharge TEST=make runtests Change-Id: I18d5097ed799e790a2742d54c25c89a7559cbcb2 Signed-off-by: dnojiri <dnojiri@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2041695
* Increase VB2_CONTEXT_MAX_SIZEdnojiri2020-03-051-2/+3
| | | | | | | | | | | | | | | This patch increases VB2_CONTEXT_MAX_SIZE to 384 to accommodate the new secdata_kernel v1.0 in vb2_context. BUG=chromium:1045217 TEST=emerge-hatch depthcharge TEST=make runtests Signed-off-by: dnojiri <dnojiri@chromium.org> Change-Id: I1df3b7a4383a45104a05a06a47cec3677dc44005 Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/vboot_reference/+/2041694 Reviewed-by: Julius Werner <jwerner@chromium.org> Reviewed-by: Joel Kitching <kitching@chromium.org>
View Lorry Controller Status