Re: [PATCH] wifi: rtl8xxxu: retry firmware download on error

[Soeren Moch]

On 23.01.25 01:21, Ping-Ke Shih wrote:
> Soeren Moch <smoch@xxxxxx> wrote:
> > Occasionally there is an EPROTO error during firmware download.
> > This error is converted to EAGAIN in the download function.
> > But nobody tries again and so device probe fails.
> >
> > Implement download retry to fix this.
> >
> > Signed-off-by: Soeren Moch <smoch@xxxxxx>
> > ---
> > Cc: Jes Sorensen <Jes.Sorensen@xxxxxxxxx>
> > Cc: Kalle Valo <kvalo@xxxxxxxxxx>
> > Cc: linux-wireless@xxxxxxxxxxxxxxx
> > Cc: linux-kernel@xxxxxxxxxxxxxxx
> >
> > This error was observed (and fix tested) on a tbs2910 board [1]
> > with an embedded RTL8188EU (0bda:8179) device behind a USB hub.
> >
> > [1] arch/arm/boot/dts/nxp/imx/imx6q-tbs2910.dts
> These can be in formal commit message.
OK, I'm happy to move this to the formal commit message if you
think this would be helpful.
> > ---
> >   drivers/net/wireless/realtek/rtl8xxxu/core.c | 10 ++++++++--
> >   1 file changed, 8 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/net/wireless/realtek/rtl8xxxu/core.c
> > b/drivers/net/wireless/realtek/rtl8xxxu/core.c
> > index 7891c988dd5f..cd7d904eae62 100644
> > --- a/drivers/net/wireless/realtek/rtl8xxxu/core.c
> > +++ b/drivers/net/wireless/realtek/rtl8xxxu/core.c
> > @@ -4064,8 +4064,14 @@ static int rtl8xxxu_init_device(struct ieee80211_hw *hw)
> >           */
> >          rtl8xxxu_write16(priv, REG_TRXFF_BNDY + 2, fops->trxff_boundary);
> >
> > -       ret = rtl8xxxu_download_firmware(priv);
> > -       dev_dbg(dev, "%s: download_firmware %i\n", __func__, ret);
> > +       for (int retry = 5; retry ; retry--) {
> > +               ret = rtl8xxxu_download_firmware(priv);
> > +               dev_dbg(dev, "%s: download_firmware %i\n", __func__, ret);
> > +               if (ret != -EAGAIN)
> > +                       break;
> > +               if (retry)
> > +                       dev_info(dev, "retry firmware download\n");
> It looks like 'if (retry)' is always true and unnecessary.
Correct. After seeing the patch in my Inbox I already realized that
the for loop should be

for (int retry = 5; retry >= 0; retry--)

This would make clearer to have a maximum of 5 retries (6 attempts in total)
and also avoid this message after the last possible retry.

> Also, this message
> isn't so important to user, so dev_dbg() is more suitable, but already printing
> "%s: download_firmware %i\n" for every retry is enough? Or move it out of loop?
I think this message is important since there is the "failed to write block"
error message before:

[    6.239205] usb 2-1.1: RTL8188EU rev D (TSMC) romver 0, 1T1R, TX
queues 2, WiFi=1, BT=0, GPS=0, HI PA=0
[    6.248793] usb 2-1.1: RTL8188EU MAC: 00:33:0e:00:15:8e
[    6.254218] usb 2-1.1: rtl8xxxu: Loading firmware rtlwifi/rtl8188eufw.bin
[    6.343753] usb 2-1.1: Firmware revision 11.1 (signature 0x88e1)
[    6.374659] usb 2-1.1: rtl8xxxu_writeN: Failed to write block at
addr: 1c80 size: 0080
[    6.382894] usb 2-1.1: retry firmware download
[    7.221505] usbcore: registered new interface driver rtl8xxxu

So the retry message after the error makes clear that the error will
hopefully
be corrected.

Since rtl8xxxu_writeN is only used for firmware download, an other option
would be to downgrade both the error and the retry message to dev_dbg().

What do you think?

Regards,
Soeren

> > +       }
> >          if (ret)
> >                  goto exit;
> >          ret = rtl8xxxu_start_firmware(priv);
> > --
> > 2.43.0