linux-kernel - Re: [PATCH] i2c: designware: Suppress error message if platform_get_irq() returns -EPROBE

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: <20150303191153.GA7862@katana>
Date:	Tue, 3 Mar 2015 20:11:53 +0100
From:	Wolfram Sang <wsa@...-dreams.de>
To:	Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Cc:	Christian Ruppert <christian.ruppert@...tech.com>,
	Alexey Brodkin <Alexey.Brodkin@...opsys.com>,
	"christian.ruppert@...lis.com" <christian.ruppert@...lis.com>,
	"linux-i2c@...r.kernel.org" <linux-i2c@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"mika.westerberg@...ux.intel.com" <mika.westerberg@...ux.intel.com>,
	"Vineet.Gupta1@...opsys.com" <Vineet.Gupta1@...opsys.com>
Subject: Re: [PATCH] i2c: designware: Suppress error message if
 platform_get_irq() returns -EPROBE_DEFER


> > > I am all for centralizing printouts. I recommended this at my ELCE
> > > talk last year, too. However, you need to keep in mind that irqs
> > > are sometimes optional and you don't want error messages for those
> > > irqs. IMO worthwhile, but not a low hanging fruit...
> > 
> > There is a lot of truth in that. Thus the initial dev_dbg() suggestion
> > to go half way. I still think that Andy's proposal (or a variation
> > thereof to catch the optional irqs case) should be the ultimate goal
> > but I agree that this is more than a quick patch and that it's
> > probably way out of scope here.
> 
> Yes, I was thinking even about some wrapper on top of platform_get_irq()
> since it seems there are no messaging done inside platform.c, though
> devm_* functions usually have it.

When I had a look a few months ago, the situation with devm_* was messy.
Some rightfully printed errors, some rightfully didn't, some vice versa,
some the other way around, and some did something else...

For driver authors, it is hard to see/remember which devm function does
it and which doesn't. IMO a good cleanup will get rid of this mess. I
started sketching something but especially clks and irqs are basically
everywhere and so it easily grew out of the fun-time project scale,
sadly.


Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)