Re: [PATCH] dump_stack: convert generic dump_stack into a weak symbol

From: Arnd Bergmann <arnd@arndb.de>
To: Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Cc: Petr Mladek <pmladek@suse.com>, Tejun Heo <tj@kernel.org>,
	Steven Rostedt <rostedt@goodmis.org>,
	Dave Young <dyoung@redhat.com>, Andi Kleen <ak@linux.intel.com>,
	Greentime Hu <green.hu@gmail.com>,
	Vincent Chen <deanbo422@gmail.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	Stephen Rothwell <sfr@canb.auug.org.au>,
	adi-buildroot-devel@lists.sourceforge.net,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Subject: Re: [PATCH] dump_stack: convert generic dump_stack into a weak symbol
Date: Wed, 7 Mar 2018 11:57:35 +0100	[thread overview]
Message-ID: <CAK8P3a3eK4iS9rEWvoRfH75zjj7KHpsPnofoGdWH28Dqbz-UyQ@mail.gmail.com> (raw)
In-Reply-To: <20180307104002.GA505@jagdpanzerIV>

On Wed, Mar 7, 2018 at 11:40 AM, Sergey Senozhatsky
<sergey.senozhatsky.work@gmail.com> wrote:
> On (03/07/18 09:46), Arnd Bergmann wrote:
>> >
>> > When is this going to happen? 4.17?
>>
>> Originally I planned to wait a few more releases, but the last maintainer
>> has commented that he will now send a patch for immediate removal,
>> so 4.17 is almost certain at this point.
>
> Would be great to get it removed as soon as possible then. Otherwise we
> will get broken blackfin build errors from Stephen (or would need to hold
> off Dave's patch).

You could also add a patch to your tree that removes the blackfin
dump_stack() function, or we could ask Stephen and the other
people operating build bots to stop building blackfin right now
(they will have to do that anyway once the arch gets removed).

>> With your patch, I would actually expect the lib/dump_stack.o file
>> to still not be picked up, so now you have a missing EXPORT_SYMBOL()
>> on the two unusual architectures until the point when you add another
>> (referenced) symbol to it.
>
> Interesting point. Didn't check it. But I checked that we have at least
> one reference to lib/dump_stack from every arch so __weak could work its
> magic. The function is show_regs_print_info(). AFAICT, every arch calls
> it (we have it in lib/dump_stack now, so we will link with lib/dump_stack).
> Anyway, I'll be happy to drop my patch. Thanks for taking a look.

Ah, right, that is after your second patch. So after the first one, it might
be broken, but the follow-up patch fixes it.

Since lib/dump_stack.c is mandatory then, I would suggest making it
obj-y and moving it out of lib/ into kernel/printk/.

      Arnd