linux-kernel - Re: [PATCH] Add Multiple TX/RX Queues Support for WWAN Network Device

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] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 16 Dec 2021 15:39:55 +0300
From:   Sergey Ryazanov <ryazanov.s.a@...il.com>
To:     Xiayu Zhang <xiayu.zhang@...iatek.com>
Cc:     Loic Poulain <loic.poulain@...aro.org>,
        David Miller <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Johannes Berg <johannes@...solutions.net>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>,
        Haijun Liu (刘海军) 
        <haijun.liu@...iatek.com>,
        Zhaoping Shu (舒召平) 
        <Zhaoping.Shu@...iatek.com>,
        HW He (何伟) <HW.He@...iatek.com>,
        srv_heupstream <srv_heupstream@...iatek.com>
Subject: Re: [PATCH] Add Multiple TX/RX Queues Support for WWAN Network Device

On Thu, Dec 16, 2021 at 3:13 PM Xiayu Zhang <xiayu.zhang@...iatek.com> wrote:
> Hi Sergey and Loic,
>
> Really thank you for these information.
>
> It seems that I need to submit another patch for discussion.
>
> At the meantime, I have some questions below and hope you could do me a
> favor.
>
> On Wed, 2021-12-15 at 22:16 +0800, Sergey Ryazanov wrote:
>
> > There are two things that trigger the discussion:
> > 1) absence of users of the new API;
>
> Can I choose WWAN device simulator (wwan_hwsim.c) as the in-tree user
> for these new APIs? And, Can I submit new APIs and changes for the user
> driver in a single patch?

This is not a good idea. Simulator is intended to test the API that is
used by other drivers for real hardware. But not for experiments with
an otherwise "userless" API.

If you need to configure the number of queues for an already in-tree
driver, then just submit a patch for it. If you plan to submit a new
driver and you need an infrastructure for it, then include patches
with a new API into a series with the driver.

>> 2) an attempt to silently correct a user choice instead of explicit
>> rejection of a wrong value.
>
> I will try to follow this:
>    a. If user doesn't specify a number, use WWAN driver's default
>    number.
>    b. If user specifies an improper value, reject it explicitly.

Yep, this would be a good solution at the moment.

-- 
Sergey