[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aHE7RfMy2St7kPma@smile.fi.intel.com>
Date: Fri, 11 Jul 2025 19:26:45 +0300
From: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To: Feng Tang <feng.tang@...ux.alibaba.com>
Cc: Petr Mladek <pmladek@...e.com>, linux-kernel@...r.kernel.org,
Andrew Morton <akpm@...ux-foundation.org>,
Nathan Chancellor <nathan@...nel.org>
Subject: Re: [PATCH v1 4/7] panic: sys_info: Replace struct sys_info_name
with plain array of strings
On Fri, Jul 11, 2025 at 11:21:20PM +0800, Feng Tang wrote:
> On Fri, Jul 11, 2025 at 12:51:10PM +0300, Andy Shevchenko wrote:
> > There is no need to keep a custom structure just for the need of
> > a plain array of strings. Replace struct sys_info_name with plain
> > array of strings.
> >
> > With that done, simplify the code, in particular, naturally use
> > for_each_set_bit() when iterating over si_bits_global bitmap.
>
> IIUC, this will not work.
>
> Actually there is a hole in the bitmap definition:
>
> #define SYS_INFO_PANIC_CONSOLE_REPLAY 0x00000020
>
> Ad Petr pointed in his review, it's only for panic use, that's why
> we make it invisible in sys_info.c. Eventually, we plan to deprecate
> 'panic_print', and use 'panic_sys_info' and 'panic_console_replay' to
> replace it. After that happens that user only see string interface,
> we can change these bitmap definition and remove the hole, and use
> your cleanup here.
Yeah, the hole should be move to the end of the bitmap, so we can have all
visible entries be sequential.
--
With Best Regards,
Andy Shevchenko
Powered by blists - more mailing lists