Em Tue, 21 Dec 2021 07:34:27 +0100 Hans Petter Selasky <hps@xxxxxxxxxxx> escreveu: > On 9/26/21 22:51, Mauro Carvalho Chehab wrote: > > As discussed on: > > https://github.com/hselasky/webcamd/issues/16 > > > > the dib0700 had a regression on Kernel 2.6.39. Such regression didn't > > affect most devices, in practice, as it seems to happen only under > > certain circunstances. > > > > Michael came up with a solution for the issue (already submitted to > > the ML) but let's take the opportunity to do a cleanup, as the resulting > > code was still touching both adapters when an stream off command > > was issued to one adapter, turning on the other one. > > > > After the change, each adapter is idependently controlled by > > a separate bit, as can be shown when its debug message > > is turned on (tested on a dual-adapter device: Hauppauge > > WinTV Nova TD): > > > > [608855.124780] adapter 1, streaming ON: 0f 10 12 > > [608868.189827] adapter 0, streaming ON: 0f 10 13 > > [608879.584330] adapter 1, streaming OFF: 0f 00 11 > > [608887.014772] adapter 0, streaming OFF: 0f 00 10 > > > > Mauro Carvalho Chehab (2): > > media: dib0700: cleanup start/stop streaming logic > > media: dib0700: Only touch one bit when start/stop an adapter > > > > Michael Kuron (1): > > media: dib0700: fix undefined behavior in tuner shutdown > > > > drivers/media/usb/dvb-usb/dib0700_core.c | 28 +++++++++++------------- > > 1 file changed, 13 insertions(+), 15 deletions(-) > > > > Were these patches upstreamed yet? Those were merged on media upstream tree. They should likely be merged upstream before v5.17-rc1. > > --HPS Thanks, Mauro
