On Fri, Feb 16, 2024 at 07:46:13PM +0100, Greg Kroah-Hartman wrote:
On Fri, Feb 16, 2024 at 07:54:42AM +0100, Thorsten Leemhuis wrote:
TWIMC, that patch (which is also queued for the next 6.6.y-rc) afaics is causing boot issues on rk3399-roc-pc for Mark [now CCed] with mainline. For details see:
https://lore.kernel.org/lkml/ZcVPHtPt2Dppe_9q@finisterre.sirena.org.uk/https...
Yeah, this is tough, this is a revert to fix a previous regression, so I think we need to stay here, at the "we fixed a regression, but the original problem is back" stage until people can figure it out and provide a working change for everyone.
Given that nobody from the USB side seems to have shown much interest in fixing this since I reported the regression should we not just be reverting whatever it was that triggered the need for the original revert (it's really not clear from the changelog...)? I got no response other than an "I said that would happen" to my initial report, then the revert triggered a half done patch which I can't really judge given my lack of familiarity with the code here but that's not been submitted as an actual patch.
The board has been working for years (at least three prior to the initial revert of "usb: typec: tcpm: fix cc role at port reset" which was only applied last June), presumably whatever the original revert was intended to fix has been there for a while?
This getting backported to older stables is breaking at least this board in those stables, and I would tend to rate a "remove all power from the system" bug at the very high end of the severity scale while the reverted patch was there for six months and several kernel releases.