Message ID | 20210413112955.1885127-1-vicamo.yang@canonical.com |
---|---|
Headers | show |
Series | Add missed firmwares claimed by kernel | expand |
Acked-by: Tim Gardner <tim.gardner@canonical.com> You really should make this a pull request. These binary blobs don't add much value to the mailing list. rtg On 4/13/21 5:29 AM, You-Sheng Yang wrote: > BugLink: https://bugs.launchpad.net/bugs/1922350 > > [Impact] > > Some kernel firmwares might be landed after linux-firmware forked for > each series, or for LTS series, kernels of new versions may claim > additional ones that were not previously available. This leaves devices > fail to function correctly while they ought to. > > With a checking script that enumerates all the blobs available in > upstream repo and in the Ubuntu fork of a given series, e.g. hirsute, > iterate through all the installed kernel modules of a given version to > check if claimed firmware blobs are available in Ubuntu's fork. > Following kernel version/series have been checked: > > * 5.11.0-14-generic/hirsute > > And it gives following files were missing: > > * amdgpu/arcturus_asd.bin > * amdgpu/arcturus_gpu_info.bin > * amdgpu/arcturus_mec.bin > * amdgpu/arcturus_mec2.bin > * amdgpu/arcturus_rlc.bin > * amdgpu/arcturus_sdma.bin > * amdgpu/arcturus_smc.bin > * amdgpu/arcturus_sos.bin > * amdgpu/arcturus_ta.bin > * amdgpu/arcturus_vcn.bin > * cmmb_vega_12mhz.inp > * cmmb_venice_12mhz.inp > * dvb_nova_12mhz.inp > * dvb_nova_12mhz_b0.inp > * isdbt_nova_12mhz.inp > * isdbt_nova_12mhz_b0.inp > * isdbt_rio.inp > * sms1xxx-hcw-55xxx-dvbt-02.fw > * sms1xxx-hcw-55xxx-isdbt-02.fw > * sms1xxx-nova-a-dvbt-01.fw > * sms1xxx-nova-b-dvbt-01.fw > * sms1xxx-stellar-dvbt-01.fw > * tdmb_nova_12mhz.inp > > [Fix] > > * af1ca28f amdgpu: add arcturus firmware > * d2adbb0b linux-firmware: Add firmware files for Siano DTV devices > - SKIPPED: this commit was explicitly reverted in LP: #1401711 > > [Test Case] > > * The same script is used to make sure as many blobs are installed as > possible. > * Also make sure only blob addition were made. No existing blob touched. > > [Where problems could occur] > > While this is to backport firmwares claimed by current kernel drivers, > devices previously failed to be successfully initialized due to the lack > of corresponding firmware will then be enabled. This may introduce new > issues carried by those devices. > > [Other Info] > > Unstable is not tested/nominated, because it should be following upstream > closely and this issue should most like happen after a developing series > being release. > > Alex Deucher (1): > amdgpu: add arcturus firmware > > WHENCE | 10 ++++++++++ > amdgpu/arcturus_asd.bin | Bin 0 -> 180736 bytes > amdgpu/arcturus_gpu_info.bin | Bin 0 -> 316 bytes > amdgpu/arcturus_mec.bin | Bin 0 -> 268560 bytes > amdgpu/arcturus_mec2.bin | Bin 0 -> 268560 bytes > amdgpu/arcturus_rlc.bin | Bin 0 -> 48044 bytes > amdgpu/arcturus_sdma.bin | Bin 0 -> 17664 bytes > amdgpu/arcturus_smc.bin | Bin 0 -> 270698 bytes > amdgpu/arcturus_sos.bin | Bin 0 -> 199248 bytes > amdgpu/arcturus_ta.bin | Bin 0 -> 82688 bytes > amdgpu/arcturus_vcn.bin | Bin 0 -> 420736 bytes > 11 files changed, 10 insertions(+) > create mode 100644 amdgpu/arcturus_asd.bin > create mode 100644 amdgpu/arcturus_gpu_info.bin > create mode 100644 amdgpu/arcturus_mec.bin > create mode 100644 amdgpu/arcturus_mec2.bin > create mode 100644 amdgpu/arcturus_rlc.bin > create mode 100644 amdgpu/arcturus_sdma.bin > create mode 100644 amdgpu/arcturus_smc.bin > create mode 100644 amdgpu/arcturus_sos.bin > create mode 100644 amdgpu/arcturus_ta.bin > create mode 100644 amdgpu/arcturus_vcn.bin >
On Tue, Apr 13, 2021 at 07:29:54PM +0800, You-Sheng Yang wrote: > BugLink: https://bugs.launchpad.net/bugs/1922350 > > [Impact] > > Some kernel firmwares might be landed after linux-firmware forked for > each series, or for LTS series, kernels of new versions may claim > additional ones that were not previously available. This leaves devices > fail to function correctly while they ought to. > > With a checking script that enumerates all the blobs available in > upstream repo and in the Ubuntu fork of a given series, e.g. hirsute, > iterate through all the installed kernel modules of a given version to > check if claimed firmware blobs are available in Ubuntu's fork. > Following kernel version/series have been checked: > > * 5.11.0-14-generic/hirsute > > And it gives following files were missing: > > * amdgpu/arcturus_asd.bin > * amdgpu/arcturus_gpu_info.bin > * amdgpu/arcturus_mec.bin > * amdgpu/arcturus_mec2.bin > * amdgpu/arcturus_rlc.bin > * amdgpu/arcturus_sdma.bin > * amdgpu/arcturus_smc.bin > * amdgpu/arcturus_sos.bin > * amdgpu/arcturus_ta.bin > * amdgpu/arcturus_vcn.bin > * cmmb_vega_12mhz.inp > * cmmb_venice_12mhz.inp > * dvb_nova_12mhz.inp > * dvb_nova_12mhz_b0.inp > * isdbt_nova_12mhz.inp > * isdbt_nova_12mhz_b0.inp > * isdbt_rio.inp > * sms1xxx-hcw-55xxx-dvbt-02.fw > * sms1xxx-hcw-55xxx-isdbt-02.fw > * sms1xxx-nova-a-dvbt-01.fw > * sms1xxx-nova-b-dvbt-01.fw > * sms1xxx-stellar-dvbt-01.fw > * tdmb_nova_12mhz.inp > > [Fix] > > * af1ca28f amdgpu: add arcturus firmware > * d2adbb0b linux-firmware: Add firmware files for Siano DTV devices > - SKIPPED: this commit was explicitly reverted in LP: #1401711 > > [Test Case] > > * The same script is used to make sure as many blobs are installed as > possible. > * Also make sure only blob addition were made. No existing blob touched. > > [Where problems could occur] > > While this is to backport firmwares claimed by current kernel drivers, > devices previously failed to be successfully initialized due to the lack > of corresponding firmware will then be enabled. This may introduce new > issues carried by those devices. > > [Other Info] > > Unstable is not tested/nominated, because it should be following upstream > closely and this issue should most like happen after a developing series > being release. Sorry, I hadn't noticed that this wasn't nacked with the other similar submissions. This one could potentially be okay to apply, if we know that this is enabling specific hardware and have some indication that it works correctly with 5.11. Though the bug used is rather genrically about missing firmware so a bug about enabling this hardware would be more suitable. Seth