On 03/10/2022 16:12, Jes Sorensen wrote: > On 10/1/22 17:42, Bitterblue Smith wrote: >> Considering there is a driver for RTL8188EU already in staging, >> which people have been whittling down, should I put any effort >> into adding support to rtl8xxxu? >> >> It's actually mostly working already [0], but rtl8xxxu doesn't >> have AP mode, adhoc mode, power saving, etc so it wouldn't be >> necessarily better than the driver in staging. >> >> [0] https://git.kernel.org/pub/scm/linux/kernel/git/jes/linux.git/log/drivers/net/wireless/realtek/rtl8xxxu?h=rtl8xxxu-8188eu > > The staging driver is based on the vendor driver which is rather > dreadful. Given the similarity of the chips, it feels really silly to > have a second driver for just one chip. > > I see no reason to abandon 8188eu support in rtl8xxxu, but given my > total lack of time, I am also not able to invest in it. > > Cheers, > Jes > So now I actually tested r8188eu and it's so bad. Why are people spending so much time cleaning up a driver that doesn't work well? With the default module parameters it had 35% packet loss with the router 3 meters away in the same room. With rtw_power_mgnt=0 the packet loss is gone but the speed is still disappointing: less than 50 megabits/second download, 65-70 upload, and it's unsteady. The v5.2.2.4 driver from Larry Finger's repo works much better: no packet loss and a steady 95 and 85 megabits/second in the same conditions. I'll get started then.