From mboxrd@z Thu Jan 1 00:00:00 1970 From: Will Deacon Subject: Re: [PATCH 3/3] iommu/arm-smmu: Error out only if not enough context interrupts Date: Tue, 24 Jul 2018 09:36:34 +0100 Message-ID: <20180724083633.GA19324@arm.com> References: <20180719175356.14753-1-vivek.gautam@codeaurora.org> <20180719175356.14753-4-vivek.gautam@codeaurora.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20180719175356.14753-4-vivek.gautam@codeaurora.org> Sender: linux-kernel-owner@vger.kernel.org To: Vivek Gautam Cc: joro@8bytes.org, robh+dt@kernel.org, robin.murphy@arm.com, andy.gross@linaro.org, mark.rutland@arm.com, iommu@lists.linux-foundation.org, devicetree@vger.kernel.org, linux-kernel@vger.kernel.org, tfiga@chromium.org, sricharan@codeaurora.org, linux-arm-msm@vger.kernel.org, david.brown@linaro.org, linux-soc@vger.kernel.org, linux-arm-kernel@lists.infradead.org List-Id: devicetree@vger.kernel.org On Thu, Jul 19, 2018 at 11:23:56PM +0530, Vivek Gautam wrote: > Currently we check if the number of context banks is not equal to > num_context_interrupts. However, there are booloaders such as, one > on sdm845 that reserves few context banks and thus kernel views > less than the total available context banks. > So, although the hardware definition in device tree would mention > the correct number of context interrupts, this number can be > greater than the number of context banks visible to smmu in kernel. > We should therefore error out only when the number of context banks > is greater than the available number of context interrupts. > > Signed-off-by: Vivek Gautam > Suggested-by: Tomasz Figa > Cc: Robin Murphy > Cc: Will Deacon > --- > drivers/iommu/arm-smmu.c | 19 +++++++++++++------ > 1 file changed, 13 insertions(+), 6 deletions(-) > > diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c > index 7c69736a30f8..4cb53bf4f423 100644 > --- a/drivers/iommu/arm-smmu.c > +++ b/drivers/iommu/arm-smmu.c > @@ -2229,12 +2229,19 @@ static int arm_smmu_device_probe(struct platform_device *pdev) > if (err) > return err; > > - if (smmu->version == ARM_SMMU_V2 && > - smmu->num_context_banks != smmu->num_context_irqs) { > - dev_err(dev, > - "found only %d context interrupt(s) but %d required\n", > - smmu->num_context_irqs, smmu->num_context_banks); > - return -ENODEV; > + if (smmu->version == ARM_SMMU_V2) { > + if (smmu->num_context_banks > smmu->num_context_irqs) { > + dev_err(dev, > + "found only %d context irq(s) but %d required\n", > + smmu->num_context_irqs, smmu->num_context_banks); > + return -ENODEV; > + } else if (smmu->num_context_banks < smmu->num_context_irqs) { > + /* loose extra context interrupts */ > + dev_notice(dev, > + "found %d context irq(s) but only %d required\n", > + smmu->num_context_irqs, smmu->num_context_banks); > + smmu->num_context_irqs = smmu->num_context_banks; > + } I don't see the utility in the new message. Can you simplify with the patch below on top? It's a bit weird that we only decide to ignore the extra irqs after calling platform_get_irq() on them, but that seems to be harmless. Will --->8 diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c index aa46c1ed5bf9..5349e22b5c78 100644 --- a/drivers/iommu/arm-smmu.c +++ b/drivers/iommu/arm-smmu.c @@ -2109,13 +2109,10 @@ static int arm_smmu_device_probe(struct platform_device *pdev) "found only %d context irq(s) but %d required\n", smmu->num_context_irqs, smmu->num_context_banks); return -ENODEV; - } else if (smmu->num_context_banks < smmu->num_context_irqs) { - /* loose extra context interrupts */ - dev_notice(dev, - "found %d context irq(s) but only %d required\n", - smmu->num_context_irqs, smmu->num_context_banks); - smmu->num_context_irqs = smmu->num_context_banks; } + + /* Ignore superfluous interrupts */ + smmu->num_context_irqs = smmu->num_context_banks; } for (i = 0; i < smmu->num_global_irqs; ++i) {