From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [PATCH 17/28] spi: fsl-espi: avoid processing uninitalized data on error Date: Mon, 24 Oct 2016 22:37:53 +0200 Message-ID: <3945489.MQiL9ES5FK@wuerfel> References: <20161017220342.1627073-1-arnd@arndb.de> <20161024184543.GR17252@sirena.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7Bit Cc: Heiner Kallweit , Linus Torvalds , linux-kernel@vger.kernel.org, Heiner Kallweit , Nobuteru Hayashi , linux-spi@vger.kernel.org To: Mark Brown Return-path: In-Reply-To: <20161024184543.GR17252@sirena.org.uk> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-spi.vger.kernel.org On Monday, October 24, 2016 7:45:43 PM CEST Mark Brown wrote: > On Mon, Oct 24, 2016 at 08:36:37PM +0200, Heiner Kallweit wrote: > > Am 24.10.2016 um 19:27 schrieb Mark Brown: > > > > This doesn't apply against current code, please check and resend. > > > The not yet reviewed part of my patch series from Oct 2nd, > > namely "[PATCH 07/11] spi: fsl-espi: fix and improve reading > > from RX FIFO" replaces the code in question. > > There's more to fix like removing polling from the ISR. > > If you prefer to apply Arnd's fix first I'd rebase the open part > > of the patch series and resend it. > > If there are dependencies you should mention them when you resend (in > general you should always mention any unapplied or cross tree > dependencies when sending things). I think my patch (the version I sent) should ideally make it into v4.9 as a bugfix. This was the powerpc warning I saw from Olof's autobuilder with the -Wmaybe-uninitialized warning added back, and it's one of the actual bugs I found (though rather unlikely to hit in practice). Merging with Heiner's patches should be trivial, and I'm pretty sure we want the patch either way. Not sure if we need a backport, it was introduced earlier this year in commit 6319a68011b8 ("spi/fsl-espi: avoid infinite loops on fsl_espi_cpu_irq()") as I now found. Arnd