Re: [PATCH] brcmfmac: Use standard SKB list accessors in brcmf_sdiod_sglist_rw.

On 11/14/2018 11:57 AM, Arend van Spriel wrote:

On 11/14/2018 11:54 AM, Andy Duan wrote:

quoted

From: Arend van Spriel [mailto:arend.vanspriel-dY08KVG/lbpWk0Htik3J/w@public.gmane.org] Sent:
2018年11月14日 16:40

quoted

To: Andy Duan <redacted>; David Miller
[off-list ref]; netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Cc: linux-wireless-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Subject: Re: [PATCH] brcmfmac: Use standard SKB list accessors in
brcmf_sdiod_sglist_rw.

On 11/14/2018 4:28 AM, Andy Duan wrote:

quoted

From: David Miller <davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org> Sent: 2018年11月11日 8:34

quoted

[ As I am trying to remove direct SKB list pointer accesses I am
  committing this to net-next.  If this causes a lot of grief I
  can and will revert, just let me know. ]

[...]

quoted

I just have bcm4339 in hands, test the patch on i.MX7D sdb board with

bcm4339, it works fine with iperf testing.

quoted

Tested-by: Fugang Duan <redacted>

Thanks, Andy

Can you do one more check? Please insert brcmfmac with module parameter
debug=2 and let me know if the following log message is seen:

brcmfmac: brcmf_sdiod_sgtable_alloc: nents=X

If not seen, the driver does not go through the patched code.

My kernel don't enable debug and DEBUG and CONFIG_BRCM_TRACING, I add
the debug info in the brcmf_sdiod_sgtable_alloc(), and the log show
the driver go through the sg path:
Log: brcmf_sdiod_sgtable_alloc: max_segs:128, sg_support:1, nents=35

Thanks, Andy

Works for me ;-)

I should better read the patch email. I tried to apply the patch to 
wireless-testing, but it failed simply because the patch is already in 
place through net-next as Dave mentioned. Anyway, it is good that it has 
been tested to some extent.

Regards,
Arend