Re: [PATCH v1 1/1] mtd: mtdpart: Do not supply NULL to printf()

[Zhihao Cheng]

在 2025/3/13 17:09, Andy Shevchenko 写道:
> On Thu, Mar 13, 2025 at 09:24:21AM +0800, Zhihao Cheng wrote:
> > 在 2025/3/13 4:16, Andy Shevchenko 写道:
> > > Compiler is not happy about NULL being supplied as printf() parameter:
> >
> > printf -> printk? The title has the same issue.
> > > drivers/mtd/mtdpart.c:693:34: error: ‘%s’ directive argument is null [-Werror=format-overflow=]
> > >
> > > Replace that with "(null)" to fix compilation error.
>
> > Reviewed-by: Zhihao Cheng <chengzhihao1@xxxxxxxxxx>
>
> Thank you!
>
> But I think my approach is a hack, the best is to move this message to the
> after the follow up conditional and drop that ternary completely as we have
> already another debug message before that. So, the parser == NULL can be
> deducted from the appearance of the one and not the other one.
>
> I'll send a v2.
>
> > >    			if (!parser && !request_module("%s", *types))
> > >    				parser = mtd_part_parser_get(*types);
> > >    			pr_debug("%s: got parser %s\n", master->name,
> > > -				parser ? parser->name : NULL);
> > > +				parser ? parser->name : "(null)");
> > >    			if (!parser)
> > >    				continue;
>
> (move it here)

After looking through 8e2c992b59fc("mtd: mtdpart: add debug prints to 
partition parser.") and 01f9c7240a90("mtd: partitions: factor out code 
calling parser"), I think we'd better keep the debug message before the 
condition 'if (!parser)', it is used to inform us whether we get a 
parser and which name it is. And the debug message in mtd_part_do_parse 
informs us the result of the 'parser->parse_fn'.
> > >    			ret = mtd_part_do_parse(parser, master, &pparts, data);