factory: HWID DB builder improvements & bugs |
|||
Issue description
Background: during Eve build, we took the chance to try HWID DB builder -- which is pretty good and helps us to create the HWID file from scratch in a very short time.
However, we do have seen few issues:
1. storage probe result will create a \n in compact_str like
compact_str: 'MMC XXXXXX 0x0 0x0100 0x000015 150100XXXXXXXXXXXX0XXXXXXXXXXX00
0x4 #24XXXXXX8'
well, technically this is a probe.py bug.
2. HWID db will add compact_str, even if the dict style values exist.
I think we should (1) don't put compact_str in HWID file if there are other values, and
(2) fix probe so every component would have at least one non-compact-str value
And in future we can deprecate compact_str.
3. most components start to encode with 0 bits. These are non-zero bit fields in my
result:
- region_field: 8
- firmware_keys_field: 3
- board_version_field: 3
- ro_main_firmware_field: 3
- ro_ec_firmware_field: 3
- storage_field: 1
- cpu_field: 1
However, it's slightly weird that it's not really mapped to 5-3-5 fields.
Was that because I didn't have customization_id in VPD?
I think we should always reserve 5 bits for it after region_field.
In fact, I'd recommend to always preserve following fields with given order:
[- image_id: 5]
- board_version_field: 3
- region_field: 5
- customization_id: 5
- firmware_keys_field: 3
- ro_main_firmware: 3
- ro_ec_firmware: 2
- cpu_field: 3
- dram_field: 2
- storage_field: 3
BTW, do we already have the way to write "version checking rules" for getting firmware version?
Can we write in HWID evaluate function, or using the new probing language?
,
Jan 7 2017
Hmmm, not all Chromebooks would have EC, and we probably will remove ro_main_fw in future, so we should migrate the table to: [- image_id: 5] - board_version_field: 3 - region_field: 5 - customization_id: 5 - firmware_keys_field: 3 - cpu_field: 5 - storage_field: 5 - dram_field: 3 - ro_main_firmware: 3 - ro_ec_firmware: 2
,
Jan 7 2017
Re #1. Reply 3: "not really mapped to 5-3-5 fields" I only let important fields to follow the 5-3-5 rule, including region and customization_id currently. I can add the fields you list into the important fields. But if I map the rule to all fields, the length of the HWID string will increase really fast. So I think we can ignore the rule for the unimportant fields.
,
Jan 9 2017
Re#3 I thought in DefaultBitSize you do try to keep the fields 5-3-5 ? Anyway, I think we definitely will have multiple firmware keys (dev, premp, mp, and more if the device has LOEM / zerg), so we should add firmware_keys_field a 3-bit one. Most Chromebooks have multiple cpu, storage, and dram, so it's fine to give them 5/5/3. For the firmware: I'm sure we'll have more then 1, but since in the future we may move them to "checking rules", fine with not having default bits for them. Now, region & customization are virtual fields so you'll handle them explicitly. Firmware_keys is almost platform independent so almost all chromebooks will have it. No problem. Board_version was supported only on modern systems so I wonder if we should list it explicitly, even if probe founds nothing (or let probe to always return something, for example 'UNKNOWN' if not found). So... maybe we do need an explicit list on "default (initial) fields". BTW I changed my mind again. Let's make them [- image_id: 5] - board_version_field: 3 - region_field: 5 - customization_id: 5 - cpu_field: 3 - storage_field: 5 - dram_field: 5 ---------------------------------------------- Fields above: always create & preserve, even if they don't exist in probe values. - firmware_keys_field: 3 - ro_main_firmware: 3 - ro_ec_firmware: 2 ---------------------------------------------- Fields above: add in order if they exist, and give right fields. Can skip if they are not in probe results. And one more thing, currently the HWID config file does not help people recognizing 5-3-5. Can you group them and and blank line between the 5-3-5 groups?
,
Jan 10 2017
One more bug found. If a new field was added, builder will say "enable XXX_field" and try to start a new image pattern. However, "adding new field" does not need new pattern - we can simply add that before the location of last non-zero-bit field. HWID v3 was designed to accept additional fields.
,
Jan 11 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/096c1b8306c56790022a38183f3708117136424a commit 096c1b8306c56790022a38183f3708117136424a Author: Hung-Te Lin <hungte@chromium.org> Date: Thu Jan 05 19:08:11 2017 gooftool: Remove components not in AVL. There are few components not in AVL and won't help us even if they live in HWID: - chipset (only makes sense on x86 as chipset) - display_converter (HDMI converter, only on ZGB era, deprecated by Type-C) - power_mgmt_chip (only on Spring, deprecated by Type-C) - vga (most Chromebooks are all using SOC today) BUG= chromium:678690 TEST=make test Change-Id: I2b56edc0a94da494574dd4094eee74bf69dbede4 Reviewed-on: https://chromium-review.googlesource.com/425470 Commit-Ready: Hung-Te Lin <hungte@chromium.org> Tested-by: Hung-Te Lin <hungte@chromium.org> Reviewed-by: Youcheng Syu <youcheng@google.com> [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/testdata/new_test_probe_result_hwid_utils.yaml [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/tools/testdata/hwid_v3_bundle_X86-GENERIC.sh [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/testdata/test_db.yaml [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/testdata/test_probe_result_hwid_utils.yaml [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/shopfloor/testdata/DEVICE [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/hwid_utils_unittest.py [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/factory_flow/templates/FAKE_HWID [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v2/hwid_unittest.py [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/gooftool/probe.py [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/testdata/test_db_wrong_checksum_field.yaml [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/test/test_lists/generic_smt.py [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/testdata/test_db_builder.yaml [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/testdata/TEST_BOARD [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/gooftool/testdata/test_data.yaml [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/gooftool/testdata/component_db [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/testdata/test_probe_result.yaml [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/database_unittest.py [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/gooftool/testdata/BENDER [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/builder_unittest.py [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/gooftool/gooftool_unittest.py [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/decoder_unittest.py [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/testdata/NEW_TEST_BOARD [modify] https://crrev.com/096c1b8306c56790022a38183f3708117136424a/py/hwid/v3/testdata/test_db_no_checksum.yaml
,
Jan 11 2017
Regarding the issue "compact_str has \n', that looks like YAML's internal format and it'll truncate long strings. Ideally YAML will remove the \n and leading spaces on next line, but for some reason our hwid builder will still generate a very long name with \n. Anyway this should not happen again since we've removed cid.
,
Jan 11 2017
BTW I think we probably want to rename board_version to board, and set its values to version: rev1
,
Jan 11 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/2878bbfe2f7cbdccc78f421a2793a295549bbf64 commit 2878bbfe2f7cbdccc78f421a2793a295549bbf64 Author: Hung-Te Lin <hungte@chromium.org> Date: Thu Jan 05 20:10:59 2017 gooftool: Add 'name' to components with only compact_str. We are going to deprecate compact_str in the future, so fields with only compact_str should be named with some more descriptive key name. BUG= chromium:678690 TEST=make test Change-Id: Iacdc74431cc16f1f3c4de35587f8f0387785772f Reviewed-on: https://chromium-review.googlesource.com/425637 Commit-Ready: Hung-Te Lin <hungte@chromium.org> Tested-by: Hung-Te Lin <hungte@chromium.org> Reviewed-by: Youcheng Syu <youcheng@google.com> [modify] https://crrev.com/2878bbfe2f7cbdccc78f421a2793a295549bbf64/py/gooftool/probe.py
,
Jan 11 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/00993ae196078e8836c5f0816e270ff278796726 commit 00993ae196078e8836c5f0816e270ff278796726 Author: Hung-Te Lin <hungte@chromium.org> Date: Sat Jan 07 08:56:41 2017 hwid: Try to prevent using 'compact_str' in HWID builder. HWID v3 config file supports dict values so we should abandon compact_str. BUG= chromium:678690 TEST=make test Change-Id: I91c810b96cd53ad9661d99a3d5c6c7848fcec978 Reviewed-on: https://chromium-review.googlesource.com/425737 Commit-Ready: Hung-Te Lin <hungte@chromium.org> Tested-by: Hung-Te Lin <hungte@chromium.org> Reviewed-by: Youcheng Syu <youcheng@google.com> [modify] https://crrev.com/00993ae196078e8836c5f0816e270ff278796726/py/hwid/v3/builder.py
,
Jan 13 2017
Re #5: The reason to create new pattern while adding/deleting field is to avoid ambiguity. If we allow to add the new kind of field into pattern, we will decode the old HWID string to index 0 for the field. But actually we don't know if all old devices have the component what maps to index 0. The detail is described here: https://docs.google.com/document/d/1Cxy9KJkysZTAPQzTTFJGSxVP37BSLVdVyE2nvRTwT2g/edit#heading=h.q2cweossgusl So the standard way to generate HWID config file in my original thought is: 1. Confirm the categories of the components that should be encoded into HWID. 2. Generate the probe result by "gooftool probe" 3. Check if there are missing components in the probe result. If so, fill a default item first (replacement of placeholder) 4. Create the HWID with the argument "--add-default-comp" If we cannot confirm the set of encoded components before build, or forgot to add them, we still can add the new field with new pattern. But consider the scenario of proto build, creating new pattern seems too heavy. I think we can add a "--force" argument to append the new component into the same pattern if the user is sure that **all old devices have the same kind of this component**.
,
Jan 16 2017
Re#11: Thanks, I agree maybe adding an option with good error message (tell people why they see this error, what they should check and how to add option) is the proper solution.
,
Jan 19 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/ff5d88f7b4531f7e3039533cdb9bcd1f721a9096 commit ff5d88f7b4531f7e3039533cdb9bcd1f721a9096 Author: Chih-Yu Huang <akahuang@google.com> Date: Wed Jan 18 06:57:09 2017 hwid builder: allow to add extra field into the pattern. Originally we do not allow to add extra field into the pattern because it would cause ambiguity. After adding a new kind of field into the pattern, we will decode the old HWID string to index 0 for the field, but we don't know if all old devices have the component what maps to index 0. But in practice, we face the scenario in early build: There is a component that cannot probed after the first HWID config file is created, and there is no second source for this component. In this case, we can append the field into the pattern directly after it can be probed. BUG= chromium:678690 TEST=make test Change-Id: Ib5d4826b1e3a0b31879b77e749f41c1a6fcff573 Reviewed-on: https://chromium-review.googlesource.com/429750 Commit-Ready: Chih-Yu Huang <akahuang@chromium.org> Tested-by: Chih-Yu Huang <akahuang@chromium.org> Reviewed-by: Hung-Te Lin <hungte@chromium.org> [modify] https://crrev.com/ff5d88f7b4531f7e3039533cdb9bcd1f721a9096/py/hwid/v3/builder_unittest.py [modify] https://crrev.com/ff5d88f7b4531f7e3039533cdb9bcd1f721a9096/py/hwid/v3/builder.py [modify] https://crrev.com/ff5d88f7b4531f7e3039533cdb9bcd1f721a9096/py/hwid/v3/hwid_utils_unittest.py [modify] https://crrev.com/ff5d88f7b4531f7e3039533cdb9bcd1f721a9096/py/hwid/v3/hwid_utils.py
,
Jan 24 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/f0677a0a1665461646f7ea64dc71d19d8d198c9e commit f0677a0a1665461646f7ea64dc71d19d8d198c9e Author: Chih-Yu Huang <akahuang@google.com> Date: Thu Nov 10 06:53:28 2016 hwid: database builder supports to add null component. We added a new argument "--add-null-comp" in database builder that add a NULL item for the assigned component. Also rename "--add-comp" to "--add-default-comp" to distinguish these two arguments. BUG= chromium:678690 TEST=make test Change-Id: Ibcbe91d1d4d3ea5c72020e0e25ebdc4915553ac1 Reviewed-on: https://chromium-review.googlesource.com/409891 Commit-Ready: Chih-Yu Huang <akahuang@chromium.org> Tested-by: Chih-Yu Huang <akahuang@chromium.org> Reviewed-by: Chih-Yu Huang <akahuang@chromium.org> [modify] https://crrev.com/f0677a0a1665461646f7ea64dc71d19d8d198c9e/py/hwid/v3/builder_unittest.py [modify] https://crrev.com/f0677a0a1665461646f7ea64dc71d19d8d198c9e/py/hwid/README.md [modify] https://crrev.com/f0677a0a1665461646f7ea64dc71d19d8d198c9e/py/hwid/v3/hwid_utils.py [modify] https://crrev.com/f0677a0a1665461646f7ea64dc71d19d8d198c9e/py/hwid/v3/testdata/test_builder_probe_results.yaml [modify] https://crrev.com/f0677a0a1665461646f7ea64dc71d19d8d198c9e/py/hwid/v3/hwid_cmdline.py [modify] https://crrev.com/f0677a0a1665461646f7ea64dc71d19d8d198c9e/py/hwid/v3/builder.py [modify] https://crrev.com/f0677a0a1665461646f7ea64dc71d19d8d198c9e/py/hwid/v3/hwid_utils_unittest.py
,
Jan 25 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/ce7003af644202fa71f31481a9a2a184503c4580 commit ce7003af644202fa71f31481a9a2a184503c4580 Author: Chih-Yu Huang <akahuang@google.com> Date: Wed Jan 18 09:54:23 2017 hwid: Set default initial encoded fields. We list the important components and preserve the encoded field for them. The first group of fields are always created. We add a default item if they don't exist in the probe result. [- image_id: 5] - board_version_field: 3 - region_field: 5 - customization_id: 5 - cpu_field: 3 - storage_field: 5 - dram_field: 5 The second group of field are added in order if they exist in the probe result. - firmware_keys_field: 3 - ro_main_firmware: 3 - ro_ec_firmware: 2 BUG= chromium:678690 TEST=make test Change-Id: Icc374ba33859719c3b839847fde86a81a37d7b17 Reviewed-on: https://chromium-review.googlesource.com/429970 Commit-Ready: Chih-Yu Huang <akahuang@chromium.org> Tested-by: Chih-Yu Huang <akahuang@chromium.org> Reviewed-by: Chih-Yu Huang <akahuang@chromium.org> [modify] https://crrev.com/ce7003af644202fa71f31481a9a2a184503c4580/py/hwid/v3/builder_unittest.py [modify] https://crrev.com/ce7003af644202fa71f31481a9a2a184503c4580/py/hwid/v3/testdata/test_builder_probe_results.yaml [modify] https://crrev.com/ce7003af644202fa71f31481a9a2a184503c4580/py/hwid/v3/builder.py [modify] https://crrev.com/ce7003af644202fa71f31481a9a2a184503c4580/py/hwid/v3/testdata/test_db_builder.yaml [modify] https://crrev.com/ce7003af644202fa71f31481a9a2a184503c4580/py/hwid/v3/hwid_utils_unittest.py
,
Feb 6 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/e5f32e3b2c9175285b61827dbeb8b48a4f68504f commit e5f32e3b2c9175285b61827dbeb8b48a4f68504f Author: Chih-Yu Huang <akahuang@google.com> Date: Mon Feb 06 07:43:36 2017 hwid: add use case in README.md. We add the yaml file and the command for each use case, and use gitiles syntax to add the table of content (TOC), and show the difference of the yaml file by two-column layout. Also we replace the word "transfer" by "convert" to describe the action of replacing legacy style of region field. BUG= chromium:678690 TEST=none Change-Id: Ibca21e802b6ec7e0aa0b5a4011cf06719b408993 Reviewed-on: https://chromium-review.googlesource.com/431545 Commit-Ready: Chih-Yu Huang <akahuang@chromium.org> Tested-by: Chih-Yu Huang <akahuang@chromium.org> Reviewed-by: Joel Kitching <kitching@chromium.org> [modify] https://crrev.com/e5f32e3b2c9175285b61827dbeb8b48a4f68504f/py/hwid/v3/builder_unittest.py [modify] https://crrev.com/e5f32e3b2c9175285b61827dbeb8b48a4f68504f/py/hwid/v3/testdata/test_db_builder.yaml [modify] https://crrev.com/e5f32e3b2c9175285b61827dbeb8b48a4f68504f/py/hwid/README.md [modify] https://crrev.com/e5f32e3b2c9175285b61827dbeb8b48a4f68504f/py/hwid/v3/builder.py
,
Feb 10 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/61ce7e20990243818b973dc705be1d492b260960 commit 61ce7e20990243818b973dc705be1d492b260960 Author: Hung-Te Lin <hungte@chromium.org> Date: Fri Feb 10 05:10:05 2017 gooftool: Remove components not in AVL. There are few components not in AVL and won't help us even if they live in HWID: - chipset (only makes sense on x86 as chipset) - display_converter (HDMI converter, only on ZGB era, deprecated by Type-C) - power_mgmt_chip (only on Spring, deprecated by Type-C) - vga (most Chromebooks are all using SOC today) BUG= chromium:678690 TEST=make test Change-Id: I2b56edc0a94da494574dd4094eee74bf69dbede4 Reviewed-on: https://chromium-review.googlesource.com/425470 Commit-Ready: Hung-Te Lin <hungte@chromium.org> Tested-by: Hung-Te Lin <hungte@chromium.org> Reviewed-by: Youcheng Syu <youcheng@google.com> Reviewed-on: https://chromium-review.googlesource.com/440709 Reviewed-by: Hung-Te Lin <hungte@chromium.org> Commit-Queue: Hung-Te Lin <hungte@chromium.org> [modify] https://crrev.com/61ce7e20990243818b973dc705be1d492b260960/py/factory_flow/templates/FAKE_HWID [modify] https://crrev.com/61ce7e20990243818b973dc705be1d492b260960/py/test/test_lists/generic_smt.py [modify] https://crrev.com/61ce7e20990243818b973dc705be1d492b260960/py/gooftool/gooftool_unittest.py [modify] https://crrev.com/61ce7e20990243818b973dc705be1d492b260960/py/gooftool/testdata/test_data.yaml [modify] https://crrev.com/61ce7e20990243818b973dc705be1d492b260960/py/gooftool/testdata/component_db [modify] https://crrev.com/61ce7e20990243818b973dc705be1d492b260960/py/hwid/v2/hwid_unittest.py [modify] https://crrev.com/61ce7e20990243818b973dc705be1d492b260960/py/gooftool/probe.py [modify] https://crrev.com/61ce7e20990243818b973dc705be1d492b260960/py/tools/testdata/hwid_v3_bundle_X86-GENERIC.sh [modify] https://crrev.com/61ce7e20990243818b973dc705be1d492b260960/py/shopfloor/testdata/DEVICE [modify] https://crrev.com/61ce7e20990243818b973dc705be1d492b260960/py/gooftool/testdata/BENDER
,
Feb 10 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/b6b91d13b40dcc277dafba56f1311c8171a71d03 commit b6b91d13b40dcc277dafba56f1311c8171a71d03 Author: Hung-Te Lin <hungte@chromium.org> Date: Fri Feb 10 05:10:07 2017 gooftool: Add 'name' to components with only compact_str. We are going to deprecate compact_str in the future, so fields with only compact_str should be named with some more descriptive key name. BUG= chromium:678690 TEST=make test Change-Id: Iacdc74431cc16f1f3c4de35587f8f0387785772f Reviewed-on: https://chromium-review.googlesource.com/425637 Commit-Ready: Hung-Te Lin <hungte@chromium.org> Tested-by: Hung-Te Lin <hungte@chromium.org> Reviewed-by: Youcheng Syu <youcheng@google.com> Reviewed-on: https://chromium-review.googlesource.com/440710 Reviewed-by: Hung-Te Lin <hungte@chromium.org> Commit-Queue: Hung-Te Lin <hungte@chromium.org> [modify] https://crrev.com/b6b91d13b40dcc277dafba56f1311c8171a71d03/py/gooftool/probe.py
,
Apr 14 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/d6ae9e3a12c28cd3cf3d96a8abfe631eaf720584 commit d6ae9e3a12c28cd3cf3d96a8abfe631eaf720584 Author: Chih-Yu Huang <akahuang@google.com> Date: Fri Apr 14 16:30:28 2017 hwid: Change the method of handling name collision. Originally, we add a underscore at the end if the name is conflicted with existed component. So the names would be: FOO FOO_ FOO__ FOO___ The name is confusing. Now we added a suffix number if collision. It becomes: FOO FOO_1 FOO_2 FOO_3 BUG= chromium:678690 TEST=py/hwid/v3/builder_unittest.py Change-Id: Ic6530aff00fba33ca7a4dad731b607a24bab9c5c Reviewed-on: https://chromium-review.googlesource.com/474785 Commit-Ready: Chih-Yu Huang <akahuang@chromium.org> Tested-by: Chih-Yu Huang <akahuang@chromium.org> Reviewed-by: Hung-Te Lin <hungte@chromium.org> Reviewed-by: Yilun Lin <yllin@chromium.org> [modify] https://crrev.com/d6ae9e3a12c28cd3cf3d96a8abfe631eaf720584/py/hwid/v3/builder_unittest.py [modify] https://crrev.com/d6ae9e3a12c28cd3cf3d96a8abfe631eaf720584/py/hwid/v3/builder.py
,
Apr 20 2017
The issues are fixed and the DB builder is verified at one project. Closed this bug.
,
Jun 6 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/platform/factory/+/99918ab4854742be5dfdd87a103e07ccee3287d5 commit 99918ab4854742be5dfdd87a103e07ccee3287d5 Author: Chih-Yu Huang <akahuang@google.com> Date: Tue Jun 06 12:15:55 2017 hwid: database builder chooses better component name. We refine the naming of mainboard, firmware_keys, and dram in this CL. BUG= chromium:678690 TEST=py/hwid/v3/builder_unittest.py Change-Id: I2f97d7d7eb4788c49113c1034441538892f68ace Reviewed-on: https://chromium-review.googlesource.com/474784 Commit-Ready: Hung-Te Lin <hungte@chromium.org> Tested-by: Hung-Te Lin <hungte@chromium.org> Reviewed-by: Hung-Te Lin <hungte@chromium.org> [modify] https://crrev.com/99918ab4854742be5dfdd87a103e07ccee3287d5/py/hwid/v3/builder_unittest.py [modify] https://crrev.com/99918ab4854742be5dfdd87a103e07ccee3287d5/py/hwid/v3/builder.py |
|||
►
Sign in to add a comment |
|||
Comment 1 by hungte@chromium.org
, Jan 7 2017also, according to proposal "Components in ChromeOS HWID", we should remove following components: - chipset (only makes sense on x86 as chipset) - display_converter (HDMI converter, only on ZGB era, deprecated by Type-C) - power_mgmt_chip (only on Spring, deprecated by Type-C) - vga (most Chromebooks are all using SOC today)