On Sat, Apr 16, 2022 at 8:59 AM Ryder Lee <Ryder.Lee@xxxxxxxxxxxx> wrote: > > On Sat, Apr 16, 2022 at 07:43:40AM -0700, Ben Greear wrote: > > It is known to the developers and the mailing list, I told both. > > But there is no fix proposed yet that I have seen. I know how to make > > my system work by both the hostapd change and by removing the > > offending patch from the kernel. That doesn't help other users > > though. > > OK, that is not convenient. > > > So far, there is one driver that supports this feature, and it is broken. > > Makes me think it won't be the last. > > This type of capability should really not be enabled in the driver before it has been properly tested.. I was under the impression that this had been tested, but apparently not sufficiently. > > > In the case bugs are found in features like this, it is a whole lot > > easier to tell the user to enable some obscure feature in their > > hostapd conf file vs have them patch and compile hostapd and/or the > > kernel. That is why I like enable/disable options for this sort of > > thing. The end users can have work arounds while proper fixes are completed and fully tested. > > It would be much more understandable to make this disabled by default rather than providing an obscure configuration parameter that the user would somehow need to figure out when to use to disable it if things don't work. I don't think it would be a good approach to introduce this type of parameters to manually disable something that might not work under some conditions. > > I would be much more likely to accept a change that would disable this recently added capability by default and provide a new option to enable it. Should the driver side functionality become more robust in the future, this new configuration option in hostapd could then be change to default to having this capability enabled or even better, remove that configuration parameter completely once there are no known issues in commonly used kernel versions. > Just to clarify, I don't think this feature is buggy itself. It has been tested on devices that are capable of dedicated chains and works well. The asia-rf device is an exception to me ... it wasn't even on the market at the time this feature was introduced. @Ben, maybe you can resend this https://patchwork.kernel.org/project/linux-wireless/patch/3683c678668c4de0ee849974977cf56e64c7162c.1640014128.git.lorenzo@xxxxxxxxxx/ to make it disabled by default Ryder _______________________________________________ Hostap mailing list Hostap@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/hostap
