linux-kernel - Re: [PATCH v3 2/2] pinctrl: samsung: support a bus clock

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] [day] [month] [year] [list]

Date: Fri, 03 May 2024 10:30:06 +0100
From: André Draszik <andre.draszik@...aro.org>
To: Krzysztof Kozlowski <krzk@...nel.org>, Tudor Ambarus
 <tudor.ambarus@...aro.org>, Sylwester Nawrocki <s.nawrocki@...sung.com>, 
 Alim Akhtar <alim.akhtar@...sung.com>, Linus Walleij
 <linus.walleij@...aro.org>, Rob Herring <robh@...nel.org>, Conor Dooley
 <conor+dt@...nel.org>, Tomasz Figa <tomasz.figa@...il.com>, Peter Griffin
 <peter.griffin@...aro.org>
Cc: Will McVicker <willmcvicker@...gle.com>, Sam Protsenko
	 <semen.protsenko@...aro.org>, kernel-team@...roid.com, 
	linux-arm-kernel@...ts.infradead.org, linux-samsung-soc@...r.kernel.org, 
	linux-gpio@...r.kernel.org, devicetree@...r.kernel.org, 
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 2/2] pinctrl: samsung: support a bus clock

Hi Krzysztof,

On Fri, 2024-05-03 at 11:13 +0200, Krzysztof Kozlowski wrote:
> On 02/05/2024 12:41, André Draszik wrote:
> > I was initially thinking the same, but the clock seems to be required for
> > register access only, interrupts are still being received and triggered
> > with pclk turned off as per my testing.
> 
> Probably we could simplify this all and keep the clock enabled always,
> when device is not suspended. Toggling clock on/off for every pin change
> is also an overhead. Anyway, I merged the patches for now, because it
> addresses real problem and seems like one of reasonable solutions.

I had contemplated a global enable of the clock on driver instantiation
as well, but in the end for me the reasons why I chose the fine-grained
approach here instead were:

* Since the clock is only needed for register access, it seems only
  natural to enable it during register accesses only. (The same would
  happen if we had support for automatic clock gating in Linux).
* If we think about external GPIO interrupts, they are likely to occur
  very rarely (think button press by operator on some external keys or
  I2C interrupts), it seems a waste to have the clock running all the
  time.
* drivers/i2c/busses/i2c-exynos5.c and drivers/soc/samsung/exynos-usi.c
  also kinda do it this way. Bus clocks are only enabled when needed
  (e.g. during transfer) (granted, the IPs (IP clocks) are also fully
  enabled/disabled in those drivers when idle, and there is no such
  thing here)

Cheers,
Andre'