Commit Graph

9606 Commits

Author SHA1 Message Date
Richard Hughes
5fcfe7f0fc Only count the Microsoft hashes when getting the dbx version
HP include extra keys that means the version is higher than expected.
2022-10-17 16:31:37 +01:00
Richard Hughes
3a0f187a16 Move the AMT functionality to the Intel MEI plugin 2022-10-13 18:34:00 +01:00
Richard Hughes
0f8ec55f46 Add a new plugin to get more Intel ME MCA data
This allows us to get the OEM Public Key BootGuard hashes.

Also add a new HSI test for leaked bootguard keys.
2022-10-13 18:34:00 +01:00
Richard Hughes
ef04d0578f trivial: Fix a critical warning when using old kernel versions 2022-10-13 18:34:00 +01:00
Mario Limonciello
d25d4f124e trivial: snap: re-enable UEFI capsule splash images 2022-10-12 11:25:34 -05:00
Mario Limonciello
7a003b597a trivial: snap: drop some of the libraries from the snap that aren't needed 2022-10-12 11:25:34 -05:00
Mario Limonciello
886ec33f79 trivial: snap: disable flashrom by default
When later this is converted to confined we won't want to support
it.  So drop it now.
2022-10-12 11:25:34 -05:00
Mario Limonciello
ea6ac08b7a trivial: snap: add some packages identified as missing 2022-10-12 11:25:34 -05:00
Mario Limonciello
b922ba1bb6 trivial: snap: add libpcre.so.3 into snap package (Fixes: #4195) 2022-10-12 10:42:19 -05:00
Mario Limonciello
95e014a844 trivial: drop snap promotion from RELEASE 2022-10-12 10:35:27 -05:00
Mario Limonciello
0230e03562 trivial: snap: ci: pass file name of snap between jobs 2022-10-12 10:22:57 -05:00
Mario Limonciello
1963f7db4e trivial: snap: ci: display path of the artifacts 2022-10-12 10:02:36 -05:00
Richard Hughes
6894f67dc5 Do not use GKeyFile to parse quirk files
There are two reasons for this. First is that GKeyFile is quite inefficient,
using a large amount of heap memory when loading. Given we don't actually use
the merge and replace functionality of GKeyFile and we only need line-by-line
access we can parse this ourselves and reduce the peak RSS considerably.

This also accidentally fixes another bug. Moving from multiple quirk files to
a single builtin.quirk meant that multiple subsystem plugins were deduped --
which isn't really what we wanted. For example, this now works:

    [MEI]
    Plugin = one
    [MEI]
    Plugin = two
2022-10-12 15:12:56 +01:00
Mario Limonciello
c583d7a2da trivial: snap: ci: upload snap not a directory 2022-10-12 09:10:29 -05:00
Mario Limonciello
6e715143f6 trivial: snap: ci: try to fix deploy to store 2022-10-12 09:00:37 -05:00
Mario Limonciello
652bd65102 trivial: snap: ci: set fetch depth for checkout
To make `git describe` work we need to have all the history.
2022-10-12 09:00:37 -05:00
Mario Limonciello
f68a35e664 trivial: snap: ci: remove checkout step from upload job
This is unnecessary.
2022-10-12 09:00:37 -05:00
Mario Limonciello
abb279fcbe trivial: fix deploy string 2022-10-12 08:11:19 -05:00
Mario Limonciello
4abbf13bef CI: split up different jobs into reusable workflows
We don't need to run the same jobs all the time, some of them make
more sense in PR context and not when commited to the branch.
2022-10-12 08:03:09 -05:00
Mario Limonciello
32994d236c trivial: use github actions to build snaps 2022-10-12 08:03:09 -05:00
Richard Hughes
2dea742004 Match the MEI UUID in quirk files
This allows us to be more specific when matching devices, and also means we get
more attributes 'for free' from the FuUdevDevice->probe().

This would allow us to have multiple device GTypes handling multiple MEI
interfaces in the same plugin., for instance, PTHI and MKHI.

The slight fly in the ointment is that the kernel does not set the 'dev' for
the mei_me devices, but it's always going to be just /dev/mei0, so hardcode it.
2022-10-12 13:17:42 +01:00
Richard Hughes
48557a428e Add fwupd_device_has_checksum() for future usage 2022-10-12 13:02:44 +01:00
Mario Limonciello
a12a5053b6 trivial: snap: add some missing deps for classic snap
Fixes: #5138
2022-10-11 11:39:54 -05:00
Richard Hughes
366e3a5668 trivial: Define the HECI GUIDs in a central place 2022-10-11 16:17:20 +01:00
Richard Hughes
14e7e1b587 Use EFI-style GUIDs when connecting the MEI interface
This means the GUID matches what the kernel exports in sysfs.
2022-10-11 16:17:20 +01:00
Aaron Skomra
413e33748d wacom_usb: Update quirks with new PID for v2 of 2nd-gen Intuos Pro Small
Wacom has added a new PID for a change in the second gen IPS
2022-10-10 16:56:34 -05:00
Mario Limonciello
bbf62261d3 trivial: snap: move to craftctl
the guidance for core22 says to migrate to this
2022-10-10 11:29:30 -05:00
Mario Limonciello
99d449fb10 trivial: snap: fetch the EFI package from the Ubuntu archive
This will let us track the same version in core22's release.
2022-10-10 11:29:30 -05:00
Mario Limonciello
6613710d4d trivial: snap: re-enable the modem manager plugin
It was accidentaly disabled when moving to core22.
2022-10-10 11:29:30 -05:00
Mario Limonciello
598e31ded5 trivial: snap: move fix-bash-completion step into build step 2022-10-10 11:29:30 -05:00
Mario Limonciello
9475a849e6 trivial: snap: add a workaround to the build process to ignore rpath
snapcraft overrides the rpath for the snap to make it most compatible.
Our rpath decisions make this not work properly though.  Since everything
is in /snap/fwupd anyway, just move the libraries to this place.

Fixes: #5138

Link: https://snapcraft.io/blog/the-new-classic-confinement-in-snaps-even-the-classics-need-a-change
2022-10-10 11:29:30 -05:00
Dylan Van Assche
8243034311 plugins: modem-manager: assume firmware is unsigned
fwupd does not know if the firmware is signed or unsigned unless
the Quectel secureboot commands set this flag. Assume that the firmware
is unsigned by default, which is the case for most firmware unless they
have they support the secureboot AT commands. If that's the case, the
right flag will be set anyway.
2022-10-10 12:44:10 +01:00
Peter
9fe9d7d7c7 Fix millisecond formatting
Fixes #5140
2022-10-09 20:33:19 +01:00
Richard Hughes
a92a33f522 Add the correct instance ID for HUB devices 2022-10-08 08:47:08 +01:00
Mario Limonciello
40c8132dbb trivial: ci: don't add gi-docgen to debian/control
It causes FTBFS due to Build-Profiles nodoc
2022-10-07 23:07:40 -05:00
Richard Hughes
c621c72367 vli: Set auto-parent-children flag for all USB hubs 2022-10-07 16:25:27 -07:00
Richard Hughes
2a2dddfdf0 vli: Remove the incorrect ParentGuid to prevent confusion
The daemon wants to auto-add the parent relationship from the analogix device
to the VLI device automatically, which is arguably more correct anyway.

No behaviour change, but the tree output in fwupdmgr will be reversed now.
2022-10-07 16:25:27 -07:00
Mario Limonciello
dd0b456a2b releasing package fwupd version 1.8.6-2 2022-10-07 16:07:57 -05:00
Mario Limonciello
02484fd4da - Drop gi-docgen from B-d that was re-added when dependencies
regenerated.  This is only supposed to be present in B-d-i
2022-10-07 16:02:55 -05:00
Mario Limonciello
deb5e437f1 releasing package fwupd version 1.8.6-1 2022-10-07 14:18:49 -05:00
Mario Limonciello
73dffc4a39 * Add a patch to force rpath to fix a false positive dpkg-shlibdeps check
on some binaries.
2022-10-07 14:16:52 -05:00
Mario Limonciello
15c9cac1b0 add an rpath to things not compiled into a library
A harmless error shows up in debian packages at build time:
```
dpkg-shlibdeps: warning: cannot find library libfwupdplugin.so needed by debian/fwupd/usr/lib/x86_64-linux-gnu/fwupd-1.8.6/libfu_plugin_flashrom.so (ELF format: 'elf64-x86-64' abi: '0201003e00000000'; RPATH: '')
```

This doesn't cause a functional problem because libfwupdplugin has already
been loaded by the daemon by the time these libraries are loaded.

In case the `dpkg-shlibdeps` checker becomes more stringent in the future
fix the warning.
2022-10-07 14:16:37 -05:00
Mario Limonciello
0ca8169d4c add an rpath to things not compiled into a library
A harmless error shows up in debian packages at build time:
```
dpkg-shlibdeps: warning: cannot find library libfwupdplugin.so needed by debian/fwupd/usr/lib/x86_64-linux-gnu/fwupd-1.8.6/libfu_plugin_flashrom.so (ELF format: 'elf64-x86-64' abi: '0201003e00000000'; RPATH: '')
```

This doesn't cause a functional problem because libfwupdplugin has already
been loaded by the daemon by the time these libraries are loaded.

In case the `dpkg-shlibdeps` checker becomes more stringent in the future
fix the warning.
2022-10-07 14:08:01 -05:00
Mario Limonciello
fd4287cf39 refresh symbols 2022-10-07 13:59:36 -05:00
Mario Limonciello
1aee161580 * debian/control:
- Refresh for this upstream change.
2022-10-07 13:55:23 -05:00
Mario Limonciello
a7c9af4ab9 Drop all patches, upstream. 2022-10-07 13:47:17 -05:00
Mario Limonciello
e0b7628e27 New upstream version (1.8.6)
* New upstream version (1.8.6)
  - Notable change upstream is that libfwupdplugin is no longer a library
    that is available for developing out-of-tree plugins.  All plugins are
    now linked into an internal library distributed in the fwupd package.
2022-10-07 13:43:40 -05:00
Mario Limonciello
08392505d8 Release fwupd 1.8.6
-----BEGIN PGP SIGNATURE-----
 
 iQFIBAABCAAyFiEEFj61ARkiXbPfj0nqF6y6jfqXDhcFAmNAR9wUHHJpY2hhcmRA
 aHVnaHNpZS5jb20ACgkQF6y6jfqXDheh5gf/YgiXA/fRmdmHXFpfcNxaUxvFBiqi
 pjmVTg8d6AA/Z9qkgq+QMYgA0HFQe4LZ76D/SgBUmrNH1Vd12mibDU3fp4tdDqDv
 E9IyrxxJNvxyzx1FKtPBWjoQi1SCdcPjFwq5fPL19uwIoiOxr56Zobc4emUcYxUn
 AJcsVjBaio9cp461w29HdQE0t70OtAyXGkt33wP9pGP6g4D3SYveu2hNDmU1m3C+
 HeP4CouiN7kiXjKdaI0E+Z0WhAbZfyJFQmtwx8eXtWd16VgrkabqG+yL9vhg11hO
 pf+1bmZaywy3Zxv9gRXB6TwwFnZivi0JlxK7rXtsIAJN8xX7Op33eDJmNQ==
 =nAuA
 -----END PGP SIGNATURE-----

Merge tag '1.8.6' into debian

Release fwupd 1.8.6
2022-10-07 13:42:01 -05:00
Richard Hughes
367b684e0b trivial: post release version bump 2022-10-07 18:00:02 +01:00
Richard Hughes
2627e8ab6a Release fwupd 1.8.6 2022-10-07 16:38:04 +01:00