linux-kernel - Re: [PATCH v3] media: si2168: Refactor command setup code

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: <b84104ca-6e4e-24ee-91e0-56aebd036a48@free.fr>
Date:   Fri, 12 Jul 2019 11:37:42 +0200
From:   Marc Gonzalez <marc.w.gonzalez@...e.fr>
To:     Uwe Kleine-König <uwe@...ine-koenig.org>,
        Antti Palosaari <crope@....fi>,
        Brad Love <brad@...tdimension.cc>, Sean Young <sean@...s.org>
Cc:     Mauro Carvalho Chehab <mchehab@...nel.org>,
        Jonathan Neuschäfer <j.neuschaefer@....net>,
        linux-media <linux-media@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        MSM <linux-arm-msm@...r.kernel.org>,
        Bjorn Andersson <bjorn.andersson@...aro.org>
Subject: Re: [PATCH v3] media: si2168: Refactor command setup code

+ Sean

On 12/07/2019 10:43, Uwe Kleine-König wrote:

> On Thu, Jul 04, 2019 at 12:33:22PM +0200, Marc Gonzalez wrote:
>
>> Refactor the command setup code, and let the compiler determine
>> the size of each command.
>>
>> Reviewed-by: Jonathan Neuschäfer <j.neuschaefer@....net>
>> Signed-off-by: Marc Gonzalez <marc.w.gonzalez@...e.fr>
>> ---
>> Changes from v1:
>> - Use a real function to populate struct si2168_cmd *cmd, and a trivial
>> macro wrapping it (macro because sizeof).
>> Changes from v2:
>> - Fix header mess
>> - Add Jonathan's tag
>> ---
>>  drivers/media/dvb-frontends/si2168.c | 146 +++++++++------------------
>>  1 file changed, 45 insertions(+), 101 deletions(-)
>>
>> diff --git a/drivers/media/dvb-frontends/si2168.c b/drivers/media/dvb-frontends/si2168.c
>> index c64b360ce6b5..5e81e076369c 100644
>> --- a/drivers/media/dvb-frontends/si2168.c
>> +++ b/drivers/media/dvb-frontends/si2168.c
>> @@ -12,6 +12,16 @@
>>  
>>  static const struct dvb_frontend_ops si2168_ops;
>>  
>> +static void cmd_setup(struct si2168_cmd *cmd, char *args, int wlen, int rlen)
> 
> I'd add an "inline" here. And you could add a const for *args.

I was under the (vague) impression that it's better to let the compiler
decide when to inline, except for trivial alternatives in headers.
David Miller wrote: "Please do not use the inline directive in foo.c
files, let the compiler decide."

Antti, Sean, what do you think?

For my notes: https://gcc.gnu.org/onlinedocs/gcc/Inline.html


>> +{
>> +	memcpy(cmd->args, args, wlen);
>> +	cmd->wlen = wlen;
>> +	cmd->rlen = rlen;
>> +}
>> +
>> +#define CMD_SETUP(cmd, args, rlen) \
>> +	cmd_setup(cmd, args, sizeof(args) - 1, rlen)
> 
> Here is the chance to add some static checking. Also it is a good habit
> to put parens around macro arguments.

Wrt parens around arguments, I figured they are not required here, since they
are used as function arguments. Though you may have a valid point.

Antti, Sean?


> Something like:
> 
> #define CMD_SETUP(cmd, args, rlen) ({ \
> 	BUILD_BUG_ON(sizeof((args)) - 1 > SI2168_ARGLEN);
> 	cmd_setup((cmd), (args), __must_be_array((args)) + sizeof((args)) - 1, (rlen));
> 
> Maybe let this macro live in drivers/media/dvb-frontends/si2168_priv.h
> where struct si2168_cmd is defined?

Antti, Sean?


> I looked over the transformations in the rest of the patch and this
> looks good.

Thanks for taking a look!

Regards.