Re: [PATCH 1/2] i2c: rcar: avoid race condition with SMIs

From: Wolfram Sang <wsa+renesas@sang-engineering.com>
To: linux-i2c@vger.kernel.org
Cc: linux-renesas-soc@vger.kernel.org,
	Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Subject: Re: [PATCH 1/2] i2c: rcar: avoid race condition with SMIs
Date: Sat, 21 May 2022 08:37:44 +0200	[thread overview]
Message-ID: <YoiIuBtm2cPcYK00@shikoro> (raw)
In-Reply-To: <20220520103325.81110-2-wsa+renesas@sang-engineering.com>

[-- Attachment #1: Type: text/plain, Size: 1186 bytes --]

On Fri, May 20, 2022 at 12:33:24PM +0200, Wolfram Sang wrote:
> A customer experienced a race condition with 'repeated starts' when a
> System Management Interrupt took over for 30us and more. The problem was
> that during the SMI a new MAT interrupt came in because we set up the
> 'repeated start' condition. But the old one was not acknowledged yet.
> So, when it was acknowledged after the SMI, the new MAT interrupt was
> lost, confusing the state machine of the driver.
> 
> The fix consists of two parts. First, we do not clear the status
> register for 'repeated starts' when preparing the next message anymore.
> The interrupt handlers for sending and receiving data is now solely
> responsible for that and it makes the code easier to follow, in fact.
> Secondly, clearing the status register is now split up to handle MAT
> interrupts independently. This avoids the race condition because the old
> MAT interrupt will be now cleared before we initiate the "repeated
> start" condition.
> 
> Reported-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
> Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>

Applied to for-next, thanks!

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]