lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date:   Fri, 31 Mar 2017 12:09:11 +0100
From:   Joao Pinto <Joao.Pinto@...opsys.com>
To:     Niklas Cassel <niklas.cassel@...s.com>,
        Joao Pinto <Joao.Pinto@...opsys.com>
CC:     netdev <netdev@...r.kernel.org>
Subject: Re: stmmac CBS configuration for TX AVB queue

Às 10:59 AM de 3/31/2017, Niklas Cassel escreveu:
> Hello Joao
> 
> 
> I was looking at
> 
> commit 19d9187317979cf0c25f67017d2676149abc46b2
> Author: Joao Pinto <Joao.Pinto@...opsys.com>
> Date:   Fri Mar 10 18:24:59 2017 +0000
> 
>     net: stmmac: configuration of CBS in case of a TX AVB queue
>     
>     This patch adds the configuration of the AVB Credit-Based Shaper.
> 
> 
> 
> It looks to me that we loop through all txqueues (including 0)
> 
> static void stmmac_configure_cbs(struct stmmac_priv *priv)
> ...
>     for (queue = 0; queue < tx_queues_count; queue++) {
>         ...
>         priv->hw->mac->config_cbs(...)
> 
> 
> then in
> 
> static void dwmac4_config_cbs(..)
> ...
>     value = readl(ioaddr + MTL_ETSX_CTRL_BASE_ADDR(queue));
> 
> 
> The MTL_ETSX_CTRL_BASE_ADDR register only exists for txqueues != 0.
> (Since txqueue0 will always handle legacy traffic.)
> 
> I think that we need to make sure that we do not enter
> config_cbs() for txqueue0, even if the user's device tree
> has txqueue0 set to AVB, since MTL_ETSX_CTRL_BASE_ADDR
> will evaluate to a non-existing register for txqueue0.
> 
> Perhaps also clarify how we expect the user to configure
> his device tree for txqueue0 when using AVB in:
> Documentation/devicetree/bindings/net/stmmac.txt
> 
> 
> Best regards,
> Niklas
> 

You are right, we did not add AVB capabilities to default/Q0, this way we
guarantee 75% bandwidth for AVB. I will send a patch soon to fix this.

Thanks,
Joao

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ