Re: [PATCH 0/9] rtw88: prepare locking for SDIO support

From: Martin Blumenstingl
Date: Fri Jan 07 2022 - 16:49:39 EST


Hi Ping-Ke,

On Fri, Jan 7, 2022 at 10:19 AM Pkshih <pkshih@xxxxxxxxxxx> wrote:
>
>
> > -----Original Message-----
> > From: Martin Blumenstingl <martin.blumenstingl@xxxxxxxxxxxxxx>
> > Sent: Wednesday, December 29, 2021 5:15 AM
> > To: linux-wireless@xxxxxxxxxxxxxxx
> > Cc: tony0620emma@xxxxxxxxx; kvalo@xxxxxxxxxxxxxx; johannes@xxxxxxxxxxxxxxxx; netdev@xxxxxxxxxxxxxxx;
> > linux-kernel@xxxxxxxxxxxxxxx; Neo Jou <neojou@xxxxxxxxx>; Jernej Skrabec <jernej.skrabec@xxxxxxxxx>;
> > Pkshih <pkshih@xxxxxxxxxxx>; Martin Blumenstingl <martin.blumenstingl@xxxxxxxxxxxxxx>
> > Subject: [PATCH 0/9] rtw88: prepare locking for SDIO support
> >
> > Hello rtw88 and mac80211 maintainers/contributors,
> >
> > there is an ongoing effort where Jernej and I are working on adding
> > SDIO support to the rtw88 driver [0].
> > The hardware we use at the moment is RTL8822BS and RTL8822CS.
> > We are at a point where scanning, assoc etc. works (though it's not
> > fast yet, in my tests I got ~6Mbit/s in either direction).
>
> Could I know if you have improvement of this throughput issue?
Yes, in the meantime we have made some performance improvements.
Currently the throughput numbers are approx.:
TX: 30Mbit/s
RX: 20Mbit/s

I have seen RX and TX throughputs of up to 50Mbit/s on my RTL8822CS,
but I cannot reliably reproduce this (meaning: if I don't touch my
board and run the same iperf3 test again then in one run it may
achieve 50Mbit/s, but in the next run only 25Mbit/s).
In other words: throughput is much better than what we started with in
summer, but I think it can be improved further.

> I have done simple test of this patchset on RTL8822CE, and it works
> well. But, I think I don't test all flows yet, so I will do more
> test that will take a while. After that, I can give a Tested-by tag.
I also got feedback off-list from a user who used the patches from
this series on top of the out-of-tree rtw88-usb driver. These patches
fix one "scheduling while atomic" issue for him as well.
Maybe you can do your extensive tests after I sent v3 of this series?
Also thanks for offering to test this, I don't have any Realtek PCIe
wifi, so I am unable to verify I broke anything myself.


Best regards,
Martin