linux-kernel - Re: [PATCH 1/4] of/device: Add a way to probe drivers by match data

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]

Message-ID: <CAMuHMdUTwUTdbKAWCNxH72-tvJuY-LtRO2Dh8izdEOKrFOSwLQ@mail.gmail.com>
Date:   Fri, 9 Nov 2018 10:56:01 +0100
From:   Geert Uytterhoeven <geert@...ux-m68k.org>
To:     Stephen Boyd <sboyd@...nel.org>
Cc:     Rob Herring <robh+dt@...nel.org>,
        Michael Turquette <mturquette@...libre.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>,
        linux-clk <linux-clk@...r.kernel.org>,
        linux-mediatek@...ts.infradead.org,
        "open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS" 
        <devicetree@...r.kernel.org>,
        Matthias Brugger <matthias.bgg@...il.com>,
        ryder.lee@...iatek.com, Frank Rowand <frowand.list@...il.com>
Subject: Re: [PATCH 1/4] of/device: Add a way to probe drivers by match data

Hi Stephen,

On Wed, Nov 7, 2018 at 7:37 PM Stephen Boyd <sboyd@...nel.org> wrote:
> Quoting Rob Herring (2018-11-06 12:44:52)
> > On Tue, Nov 6, 2018 at 12:36 PM Stephen Boyd <sboyd@...nel.org> wrote:
> > > We have a handful of clk drivers that have a collection of slightly
> > > variant device support keyed off of the compatible string. In each of
> > > these drivers, we demux the variant and then call the "real" probe
> > > function based on whatever is stored in the match data for that
> > > compatible string. Let's generalize this function so that it can be
> > > re-used as the platform_driver probe function directly.
> >
> > This looks really hacky to me. It sounds kind of general, but really
> > only works if we have match data that's a single function and we lose
> > any type checking on the function.
>
> I don't know what this means. Are you saying that we lose the ability to
> type check the function pointer stored in the data member? I don't have
> a good answer for this besides it's not any worse than the status quo
> for the mediatek drivers.

The .data field has always been void *, and is used for storing different
things, depending on the driver.
It's already up to the driver writer to get that right.

> One alternative is to add a DT only array to the platform_driver struct
> that has the platform driver probe function type and matches the index
> in the of_device_id table. Then we can add some logic into
> platform_drv_probe() to look for this table being set and find the probe
> function to call. Then we still have match data for anything that wants
> that (maybe it could be passed in to the probe function) at the cost of
> having another array. I don't have a use-case for this right now so I'm
> not sure this is a great idea.
>
>   struct of_platform_driver_probe_func {
>         int (*probe)(struct platform_device *pdev);
>   };
>
>   struct of_platform_driver_probe_func mtk_probes[] = {
>         mtk_probe1,
>         mtk_probe2,
>         mtk_probe3,
>   };
>
>   struct platform_driver mtk_driver = {
>         .of_probes = &mtk_probes;
>         .driver = {
>                 .name = "mtk-foo";
>                 .of_match_table = mtk_match_table,
>         },
>   };

This looks worse to me: people tend to be very good at keeping multiple
arrays in sync :-(

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...ux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds