From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mout.gmx.net (mout.gmx.net [212.227.15.15]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id C16483B29E for ; Fri, 10 Jan 2025 12:02:26 -0500 (EST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmx.de; s=s31663417; t=1736528543; x=1737133343; i=moeller0@gmx.de; bh=flwZz/Np7T5YRXO3X3wv7g+wcpUspO6NzPpLXcnd8YA=; h=X-UI-Sender-Class:Date:From:To:CC:Subject:In-Reply-To:References: Message-ID:MIME-Version:Content-Type:Content-Transfer-Encoding:cc: content-transfer-encoding:content-type:date:from:message-id: mime-version:reply-to:subject:to; b=oy1C9f/+Zm8YOVG9gOUGW89bxk3cCw9n3J5E+BRQvs8JIAjuhs6/S+X8q4iVcR3+ UQE2770e505czTGTswOPoPv4DBECDmYWBEnkbVQs53UO07kjs9662F1Za2oqQL5v8 4GpX9d23FyRcHeDNekaff0uHMw2UPOWLd4p/IFSfkFhQ2kGjY9h8erm30Ao+R4HvE +DBLe+XbpDRAzi/+GszQIc7WXZwUQLCxAQbcSoM7ntHlrPGB63YNhQDuEw3Vc4xkq 5LmcmXdpEkXN0FQu0e3ElyqZ0bIsjsIRV2KwxsVHU3vryGUux07SmGpGIfGzDAMO5 +e3JfXvHCVtDKEHsWw== X-UI-Sender-Class: 724b4f7f-cbec-4199-ad4e-598c01a50d3a Received: from [127.0.0.1] ([77.8.251.184]) by mail.gmx.net (mrgmx005 [212.227.17.190]) with ESMTPSA (Nemesis) id 1MbRfv-1szAZS2Xjl-00pQlN; Fri, 10 Jan 2025 18:02:23 +0100 Date: Fri, 10 Jan 2025 18:02:18 +0100 From: Sebastian Moeller To: Dave Taht CC: =?ISO-8859-1?Q?Toke_H=F8iland-J=F8rgensen?= , =?ISO-8859-1?Q?Toke_H=F8iland-J=F8rgensen?= , cake@lists.bufferbloat.net User-Agent: K-9 Mail for Android In-Reply-To: References: <20250110155531.300303-1-toke@redhat.com> <2FE6B2E2-9A34-47C1-B203-F0C424F00704@gmx.de> Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable X-Provags-ID: V03:K1:FBaXPZALMfGvpyCFfoq61cr7SaPTTurIfC+OfyYS8eKvvZjzJzZ UNQXAmnZNHP+3dBAq7L7uD1d31W+KJwYYQ9dL9O8d0BJd22DLu//htd2uQhs2DzdGhMgqRP Zw9Fmra6LcN9+pO59iopGQNV6YsMzUtKUF37LK85GEzn0ZOLva+SyxkBaBluuaFfwrOTly/ 0//wWV2NVTZdGrVN1SpMQ== X-Spam-Flag: NO UI-OutboundReport: notjunk:1;M01:P0:aZ/PCdJohwg=;CqwZjZ/omfgNteuhcdVJN82z6np tEVc4Gu+gEUtBTEuqFiXIGnU0oOjd2lr6yGbzu4OGcn7A6rm5kbh9qb7sc0ogcUf2eh1q2XGA jU732NKKbxdZWEt7q2vCZBdWrNgTQCgw4dD1bTlb7Q26yBrqDuLdxIUb5e5lldf4XGuETZq98 a6+YK6kWBuEDpiLSK8xaGGQVKCzqueUi6cNojHEaE8eZPmC6feIdq0eRgonnHa3VfoiXK02zZ +utNpnzg2gWFMruxbuiRpzq23pIIqFDca/0leCo4i6/A/Nu/zTQ5KYzL7J4zue0Va/oM64jJB JznP2kVbRBPbOFj7sSppC5axafeurvDaCOcMKJCYgd2bgBqZaeHBbIL2MVURW5YQ4/j48KHEU kSfo93WrjfShWFJgW/7fJDtPVPvTry10O/vdY8jhsHzCzj3OeDwP4QyCl/2Fs3LM7NXZ+IkO4 VCdlCcGsWwWU2EdZNktS15epPmD2ue/KeJh00ABBqoO7KpOK2dxdbgSoP4VK/bfEMK1xrnsih Nkm8A2H2R53XgzxQ6TLCBVxXIJ64VcmM3xjLY7fHzucfcOSmFp6GzMtusG00ZlR+97OkHVzrJ EwnIFuBAxc48Sez599w31AuUhz5cqKAZ1VlVoWXlt9Nun+iMPc+IQD9Jk7E8vAmWITbQGaWbt XmN2S05wVNIkbFpXqQqOTmDIUHHmXTMoZETmlcehrYRFU0S5m4ZqoP6DpHpGmobZ3jsJkQohq 1ebfShy2pOYxW4vhA9x4N761MpcdawBNSNwzPsogiuyDBpcTHxw7caDyjpAqAdcCAlfiiHdpe Dx1xTwwzV1WAj94abCwaWZAep0oi1qNPGZp/UqP6vUJ8CcON4sM7SM2+G4TAPngiGzY+1Woen SxT3Ss5tXsTFBBdKMXbq+UklPqj8GhqAQb1agFma99m9OapsskE5KsxOygb3qnVNsqwMUkes2 5FwBc71m8EYKAY/xymS0B7Ez38Ucyr297wmg83tOXWCzXFProtgryKOS4qrVN4xI0sOSsIui0 RJ7OrYdYST/cB/D57lFwtUhiMgirS1gQPzq5u6ZYpdWNEfGCMdhR88+C2T+SvYFLcGZVENWWF VCBmhDbtFK1w/1EhWxMLdil9BxLvgX Subject: Re: [Cake] [PATCH net-next] sched: sch_cake: Align QoS treatment to Windows and Zoom X-BeenThere: cake@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: Cake - FQ_codel the next generation List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 10 Jan 2025 17:02:27 -0000 On 10 January 2025 17:34:45 CET, Dave Taht wrote= : >actually in re-reading that, I think it leaves nqb in 0 (44, not 45, >is now VI), which wasn't my intent=2E > >More below=2E=2E=2E > >On Fri, Jan 10, 2025 at 8:24=E2=80=AFAM Sebastian Moeller wrote: >> >> Mmmh=2E off the Linux lists/folks for discussion: >> >> can I propose to drop NQB down to BestEffort? The supposed requirements= for NQB marking should make these flows thrive in a flow isolating schedul= er like cake without any special treatment=2E=2E=2E moving it to Video or w= orse Voice really only invites abuse=2E=2E=2E > >I would prefer consistent behavior=2E Without proper policing, NQB in Voice, with all/most other traffic in Best= Effort a single NQB flow can easily hog up to 50% (I believe this is the ca= ke default) of capacity=2E=2E=2E=20 first rule of priority schedulers think twice before exposing a priority s= cheduler without admission control to the wider internet=2E=2E=2E > >> >> >> >> > On 10=2E Jan 2025, at 16:55, Toke H=C3=B8iland-J=C3=B8rgensen via Cak= e wrote: >> > >> > From: Dave Taht >> > >> > Cake's diffserv4 mode attempted to follow the IETF webrtc >> > QoS marking standards, RFC8837=2E >> > >> > It turns out Windows QoS can only use CS0, CS1, CS5, and CS7=2E >> >> Nah, applications apparently can only request a small set of DSCPs, but= it is possible to make whole applications use a different DSCP: >> example for making putty use EF >> >> New-NetQosPolicy -Name "putty" -AppPathNameMatchCondition "putty=2Eexe"= -PolicyStore ActiveStore -NetworkProfile All -DSCPAction 46 > >Within an App=2E=2E=2E Yepp, that solves zoom's issue, as I am willing to bet default all CS0 zoo= m does not have fluidity issues due to congesting itself=2E=2E=2E > >> >> >> > >> > Zoom defaults to using CS5 for video and screen sharing traffic=2E >> >> Zoom defaults to using CS0 (my packet captures) unless an adminstrator = enables DSCPs, which I assume will only/mostly happen for business accounts= =2E > >I have seen it enabled=2E=20 Sure, on business accounts that do come with an administrator access to th= e businesses zoom account=2E=2E=2E typical home users are CS0 all the way, = all the time=2E > I have also multiple reports of "stuttering", >which I am hoping this resolves=2E > >> >> > Bump CS4, CS5, and NQB to the video tin (2) in diffserv4 mode, for >> > more bandwidth and lower priority=2E >> >> I do endorse that, but that change will affect all current users of cak= e as parts of their individual QoS hierarchies=2E > >I think windows users are not as clued=2E They are also not that likely to use cake in the first place=2E=2E=2E so f= or the convenience on users we do not have we risk annoying experienced use= rs=2E I am not saying this should stop us but we should consider that, afte= r all this can not easily be overridden via the commandline=2E >> > >> > This also better aligns with how WiFi presently treats CS5 and NQB=2E >> >> Two issues: it is not that default WiFi uses a sane system to begin wit= h, and NQB IMHO should not be elevated in priority unless it is also police= d well, something neither WiFi nor cake will do out of the box=2E >> >> >> > >> > Signed-off-by: Dave Taht >> > Signed-off-by: Toke H=C3=B8iland-J=C3=B8rgensen >> > --- >> > net/sched/sch_cake=2Ec | 4 ++-- >> > 1 file changed, 2 insertions(+), 2 deletions(-) >> > >> > diff --git a/net/sched/sch_cake=2Ec b/net/sched/sch_cake=2Ec >> > index 48dd8c88903f=2E=2E2a9288d4b873 100644 >> > --- a/net/sched/sch_cake=2Ec >> > +++ b/net/sched/sch_cake=2Ec >> > @@ -328,8 +328,8 @@ static const u8 diffserv4[] =3D { >> > 1, 0, 0, 0, 0, 0, 0, 0, >> > 2, 0, 2, 0, 2, 0, 2, 0, >> > 2, 0, 2, 0, 2, 0, 2, 0, >> > - 3, 0, 2, 0, 2, 0, 2, 0, >> > - 3, 0, 0, 0, 3, 0, 3, 0, >> > + 2, 0, 2, 0, 2, 0, 2, 0, >> > + 2, 0, 0, 0, 2, 0, 3, 0, >> > 3, 0, 0, 0, 0, 0, 0, 0, >> > 3, 0, 0, 0, 0, 0, 0, 0, >> > }; >> > -- >> > 2=2E47=2E1 >> > >> > _______________________________________________ >> > Cake mailing list >> > Cake@lists=2Ebufferbloat=2Enet >> > https://lists=2Ebufferbloat=2Enet/listinfo/cake >> > > --=20 Sent from my Android device with K-9 Mail=2E Please excuse my brevity=2E