On 04.08.23 12:15, Max Chou wrote: >> -----Original Message----- >> [CCing the linux-firmware maintainer, the network maintainers, and the >> regressions lists] >> >> On 04.08.23 07:54, max.chou@xxxxxxxxxxx wrote: >>> From: Max Chou <max.chou@xxxxxxxxxxx> >>> >>> In the commit of linux-firmware project, rtl8852cu_fw.bin is updated >>> as FW v2 format[1]. Consider the case that if driver did not be updated fo> >> FW v2 supported[2], it can not use FW v2. >>> >>> By Canonical's suggestion, older driver should be able to load FW v1, >> >> Well, that's not only Canonical suggestion, that is how things are supposed to >> be handled in general. See >> Documentation/driver-api/firmware/firmware-usage-guidelines.rst (alt: >> https://docs.kernel.org/driver-api/firmware/firmware-usage-guidelines.html >> ) for details. >> >> We had a similar situation in March already with a Wifi driver: >> https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/c >> ommit/?id=416a66cae796939d7d096988c72c641656c81c5a >> >> Makes me wonder: is there a problem here we should try harder to avoid? >> Especially as the problem in this case would have been simple to predict by >> reading the changelog of the linux-firmware change. To quote from >> https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/c >> ommit/?id=55e7448533e7 >> >> ``` >> Note: This firmware patch needs to be used with driver >> "btrtl: Firmware format v2 support". Please update the driver with this commit: >> 9a24ce5e29b15c4c6b0c89c04f9df6ce14addefa >> ``` > > Actually, we thought this note is enough. [...] Yeah, no big deal, error are made and dealt with, and you are working on that. Thx for that. It's just that the episode got got me thinking: It's the second time such error is made within a a few months, hence I wonder if we on the side of the Linux kernel and/or Linux-firmware should change something to ideally prevent this from happening. Ciao, Thorsten
