Fwd: nvsp_rndis_pkt_complete error status and net_ratelimit: callbacks suppressed messages on 6.4.0rc4

From: Bagas Sanjaya <bagasdotme@gmail.com>
To: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Linux Regressions <regressions@lists.linux.dev>,
	Linux Kernel Network Developers <netdev@vger.kernel.org>,
	Linux BPF <bpf@vger.kernel.org>,
	Linux on Hyper-V <linux-hyperv@vger.kernel.org>
Cc: Michael Kelley <mikelley@microsoft.com>,
	Haiyang Zhang <haiyangz@microsoft.com>,
	Paolo Abeni <pabeni@redhat.com>,
	"David S. Miller" <davem@davemloft.net>,
	Jakub Kicinski <kuba@kernel.org>
Subject: Fwd: nvsp_rndis_pkt_complete error status and net_ratelimit: callbacks suppressed messages on 6.4.0rc4
Date: Tue, 30 May 2023 19:25:36 +0700	[thread overview]
Message-ID: <15dd93af-fcd5-5b9a-a6ba-9781768dbae7@gmail.com> (raw)

Hi,

I notice a regression report on Bugzilla [1]. Quoting from it:

> After building 6.4.0rc4 for my VM running on a Windows 10 Hyper-V host, I see the following 
> 
> [  756.697753] net_ratelimit: 34 callbacks suppressed
> [  756.697806] hv_netvsc cd9dd876-2fa9-4764-baa7-b44482f85f9f eth0: nvsp_rndis_pkt_complete error status: 2
> (snipped repeated messages)
> 
> *but*, I'm only able to reliably reproduce this if I am generating garbage on another terminal, e.g. sudo strings /dev/sda
> 
> 
> This doesn't appear to affect latency or bandwidth a huge amount, I ran an iperf3 test between the guest and the host while trying to cause these messages.
> Although you if you take 17-18 gigabit as the "base" speed, you can see it drop a bit to 16 gigabit while the errors happen and "catch up" when I stop spamming the console.
> 
> [  5]  99.00-100.00 sec  1.89 GBytes  16.2 Gbits/sec
> [  5] 100.00-101.00 sec  1.91 GBytes  16.4 Gbits/sec
> [  5] 101.00-102.00 sec  1.91 GBytes  16.4 Gbits/sec
> [  5] 102.00-103.00 sec  1.91 GBytes  16.4 Gbits/sec
> [  5] 103.00-104.00 sec  1.92 GBytes  16.5 Gbits/sec
> [  5] 104.00-105.00 sec  1.94 GBytes  16.6 Gbits/sec
> [  5] 105.00-106.00 sec  1.89 GBytes  16.2 Gbits/sec
> [  5] 106.00-107.00 sec  1.90 GBytes  16.3 Gbits/sec
> [  5] 107.00-108.00 sec  2.23 GBytes  19.2 Gbits/sec
> [  5] 108.00-109.00 sec  2.57 GBytes  22.0 Gbits/sec
> [  5] 109.00-110.00 sec  2.66 GBytes  22.9 Gbits/sec
> [  5] 110.00-111.00 sec  2.64 GBytes  22.7 Gbits/sec
> [  5] 111.00-112.00 sec  2.65 GBytes  22.7 Gbits/sec
> [  5] 112.00-113.00 sec  2.65 GBytes  22.8 Gbits/sec
> [  5] 113.00-114.00 sec  2.65 GBytes  22.8 Gbits/sec
> [  5] 114.00-115.00 sec  2.65 GBytes  22.8 Gbits/sec
> [  5] 115.00-116.00 sec  2.66 GBytes  22.9 Gbits/sec
> [  5] 116.00-117.00 sec  2.63 GBytes  22.6 Gbits/sec
> [  5] 117.00-118.00 sec  2.69 GBytes  23.1 Gbits/sec
> [  5] 118.00-119.00 sec  2.66 GBytes  22.9 Gbits/sec
> [  5] 119.00-120.00 sec  2.67 GBytes  22.9 Gbits/sec
> [  5] 120.00-121.00 sec  2.66 GBytes  22.9 Gbits/sec
> [  5] 121.00-122.00 sec  2.49 GBytes  21.4 Gbits/sec
> [  5] 122.00-123.00 sec  2.15 GBytes  18.5 Gbits/sec
> [  5] 123.00-124.00 sec  2.16 GBytes  18.6 Gbits/sec
> [  5] 124.00-125.00 sec  2.16 GBytes  18.6 Gbits/sec
> 

See bugzilla for the full thread.

Anyway, I'm adding it to regzbot:

#regzbot introduced: dca5161f9bd052 https://bugzilla.kernel.org/show_bug.cgi?id=217503
#regzbot title: net_ratelimit and nvsp_rndis_pkt_complete error due to SEND_RNDIS_PKT status check

Thanks.

[1]: https://bugzilla.kernel.org/show_bug.cgi?id=217503

-- 
An old man doll... just what I always wanted! - Clara