[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <YxdC7i9F1ayR5icS@smile.fi.intel.com>
Date: Tue, 6 Sep 2022 15:54:06 +0300
From: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To: Dmitry Torokhov <dmitry.torokhov@...il.com>
Cc: linux-input@...r.kernel.org, linux-kernel@...r.kernel.org,
Hans de Goede <hdegoede@...hat.com>
Subject: Re: [PATCH v1 1/1] Input: icn8505 - Utilize acpi_get_subsystem_id()
On Mon, Sep 05, 2022 at 12:35:42PM -0700, Dmitry Torokhov wrote:
> On Mon, Sep 05, 2022 at 08:20:01PM +0300, Andy Shevchenko wrote:
...
> > + subsys = acpi_get_subsystem_id(ACPI_HANDLE(dev));
> > + if (IS_ERR(subsys) && PTR_ERR(subsys) != -ENODATA)
> > + return PTR_ERR(subsys);
> > +
> > + if (IS_ERR(subsys) && PTR_ERR(subsys) == -ENODATA)
> > + subsys = kstrdup_const("unknown", GFP_KERNEL);
>
> Do we really need kstrdup_const() here? This makes me wonder if we need
> to also have error handling here, and if we going to tip some automated
> tools by not having it. Why can't we simply assign the constant here
> (and continue using kfree_const() below)?
Which makes code inconsistent. But okay, no big deal.
> I think this is the case where PTR_ERR_OR_ZERO() might help avoid
> multiple IS_ERR/PTR_ERR:
>
> subsys = acpi_get_subsystem_id(ACPI_HANDLE(dev));
> error = PTR_ERR_OR_ZERO(subsys);
> if (error == -ENODATA)
> subsys = "unknown";
> else if (error)
> return error;
Would it matter? The generated code will be the same in both cases, no?
> > snprintf(icn8505->firmware_name, sizeof(icn8505->firmware_name),
> > "chipone/icn8505-%s.fw", subsys);
> >
> > - kfree(buffer.pointer);
> > + kfree_const(subsys);
> > return 0;
--
With Best Regards,
Andy Shevchenko
Powered by blists - more mailing lists