Hi Adam, thanks for the report and offering help. Mikal has already made some tests. It turned out that this isn't only a problem of protocol handling. It seems pms(4) receives input that hasn't the faintest resemblance to the protocol as I know it, and I'm afraid that more email debugging may be too time-consuming and lead to nothing. If Mikal's case is no exception, then it might be best for now to revert the pms patch from August, at least partially.
Regards, Ulf On 9/16/19 8:45 PM, Adam Wolk wrote: > > I have a lenovo G50-70 that exposes the exact same issue. > Last snapshot I tested this on is: > kern.version=OpenBSD 6.6-beta (GENERIC.MP) #313: Tue Sep 10 23:30:52 MDT 2019 > [email protected]:/usr/src/sys/arch/amd64/compile/GENERIC.MP > > After which I compiled from source and resteded with no patches, which is > GENERIC #0 in my dmesgs. > > On Mon, Sep 09, 2019 at 04:53:46AM +0200, Mikal Villa wrote: >> Hi Ulf, >> >> Thanks for the reply. Yea I'm more than happy to test some patches, and >> I already tested the one attached in your reply and unfortunately it did >> not seem to work, mouse/pointer is back to beeing mainly unresponsive. >> > > I tested the same diff. Unfortunately with similar results, the issue > still persists with no noticable changes. This is GENERIC #1 in my dmesgs. > >> I can try describe the issue a bit better; >> >> Lets say I try moving my finger from left to right over the touchpad, >> maybe one of five times, the pointer changes position - but not in a >> controlled way, it can be all from moving half the screen to just move >> about the same length as the default input field for username in xenodm. >> Both with your patch, and cvs HEAD(last commit) give me that behavior >> while pms0 spams dmesg with the not in sync yet message. >> >> Re email: I'll check the spam issue, and I'm following bugs@ :) >> > > That is also what I am experiencing. Additionally, USB mouse doesn't > work when attached after that pms change. > >> >> On 08/09/2019, 21:30, Ulf Brosziewski wrote: >>> Hi Mikal, >>> >>> without the pms patch 1.89, your touchpad isn't recognized as touchpad, >>> and it will emulate a PS2 mouse. That's better than nothing, of course, >>> but only a workaround. >>> > > I tried reverting 1.89 and this leads to a working touchpad and working USB > mouse. This is GENERIC #2 in my dmesgs. > >>> Would you mind to test one or two patches for pms? I've noticed that the >>> packet validation code is not complete. Below is a quick shot that should >>> correct it. >>> >>> Regards, >>> Ulf >>> > > > I am willing to test more patches. > > Regards, > Adam >
