From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-14.0 required=3.0 tests=BAYES_00,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 9D2FEC43460 for ; Fri, 7 May 2021 10:17:05 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 65A316145E for ; Fri, 7 May 2021 10:17:05 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S235926AbhEGKSC convert rfc822-to-8bit (ORCPT ); Fri, 7 May 2021 06:18:02 -0400 Received: from mail.kernel.org ([198.145.29.99]:38264 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229812AbhEGKSB (ORCPT ); Fri, 7 May 2021 06:18:01 -0400 Received: from disco-boy.misterjones.org (disco-boy.misterjones.org [51.254.78.96]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id 64E906143F; Fri, 7 May 2021 10:17:01 +0000 (UTC) Received: from 78.163-31-62.static.virginmediabusiness.co.uk ([62.31.163.78] helo=why.misterjones.org) by disco-boy.misterjones.org with esmtpsa (TLS1.3) tls TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 (Exim 4.94.2) (envelope-from ) id 1lexXX-00BRMu-1h; Fri, 07 May 2021 11:16:59 +0100 Date: Fri, 07 May 2021 11:16:58 +0100 Message-ID: <87czu2q25h.wl-maz@kernel.org> From: Marc Zyngier To: Pali =?UTF-8?B?Um9ow6Fy?= Cc: Lorenzo Pieralisi , Thomas Petazzoni , Rob Herring , Bjorn Helgaas , Russell King , Marek =?UTF-8?B?QmVow7pu?= , Remi Pommarel , Xogium , Tomasz Maciej Nowak , linux-pci@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH 17/42] PCI: aardvark: Fix support for MSI interrupts In-Reply-To: <20210506153153.30454-18-pali@kernel.org> References: <20210506153153.30454-1-pali@kernel.org> <20210506153153.30454-18-pali@kernel.org> User-Agent: Wanderlust/2.15.9 (Almost Unreal) SEMI-EPG/1.14.7 (Harue) FLIM-LB/1.14.9 (=?UTF-8?B?R29qxY0=?=) APEL-LB/10.8 EasyPG/1.0.0 Emacs/27.1 (x86_64-pc-linux-gnu) MULE/6.0 (HANACHIRUSATO) MIME-Version: 1.0 (generated by SEMI-EPG 1.14.7 - "Harue") Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8BIT X-SA-Exim-Connect-IP: 62.31.163.78 X-SA-Exim-Rcpt-To: pali@kernel.org, lorenzo.pieralisi@arm.com, thomas.petazzoni@bootlin.com, robh@kernel.org, bhelgaas@google.com, rmk+kernel@armlinux.org.uk, kabel@kernel.org, repk@triplefau.lt, contact@xogium.me, tmn505@gmail.com, linux-pci@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org X-SA-Exim-Mail-From: maz@kernel.org X-SA-Exim-Scanned: No (on disco-boy.misterjones.org); SAEximRunCond expanded to false Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 06 May 2021 16:31:28 +0100, Pali Rohár wrote: > > MSI domain callback .alloc (implemented by advk_msi_irq_domain_alloc() > function) should return zero on success. Returning non-zero value indicates > failure. Fix return value of this function as in many cases it now returns > failure while allocating IRQs. > > Aardvark hardware supports Multi-MSI and MSI_FLAG_MULTI_PCI_MSI is already > set. But when allocating MSI interrupt numbers for Multi-MSI, they need to > be properly aligned, otherwise endpoint devices send MSI interrupt with > incorrect numbers. Fix this issue by using function bitmap_find_free_region() > instead of bitmap_find_next_zero_area(). > > To ensure that aligned MSI interrupt numbers are used by endpoint devices, > we cannot use Linux virtual irq numbers (as they are random and not > properly aligned). So use hwirq numbers allocated by the function > bitmap_find_free_region(), which are aligned. This needs an update in > advk_msi_irq_compose_msi_msg() and advk_pcie_handle_msi() functions to do > proper mapping between Linux virtual irq numbers and hwirq MSI inner domain > numbers. > > Also the whole 16-bit MSI number is stored in the PCIE_MSI_PAYLOAD_REG > register, not only lower 8 bits. Fix reading content of this register. > > This change fixes receiving MSI interrupts on Armada 3720 boards and allows > using NVMe disks which use Multi-MSI feature with 3 interrupts. > > Without this change, NVMe disks just freeze booting Linux on Armada 3720 > boards as linux nvme-core.c driver is waiting 60s for an interrupt. > > Signed-off-by: Pali Rohár > Reviewed-by: Marek Behún > Cc: stable@vger.kernel.org # f21a8b1b6837 ("PCI: aardvark: Move to MSI handling using generic MSI support") > --- > drivers/pci/controller/pci-aardvark.c | 32 ++++++++++++++++----------- > 1 file changed, 19 insertions(+), 13 deletions(-) > > diff --git a/drivers/pci/controller/pci-aardvark.c b/drivers/pci/controller/pci-aardvark.c > index 366d7480bc1b..498810c00b6d 100644 > --- a/drivers/pci/controller/pci-aardvark.c > +++ b/drivers/pci/controller/pci-aardvark.c > @@ -118,6 +118,7 @@ > #define PCIE_MSI_STATUS_REG (CONTROL_BASE_ADDR + 0x58) > #define PCIE_MSI_MASK_REG (CONTROL_BASE_ADDR + 0x5C) > #define PCIE_MSI_PAYLOAD_REG (CONTROL_BASE_ADDR + 0x9C) > +#define PCIE_MSI_DATA_MASK GENMASK(15, 0) See my comment below about this addition. > /* LMI registers base address and register offsets */ > #define LMI_BASE_ADDR 0x6000 > @@ -861,7 +862,7 @@ static void advk_msi_irq_compose_msi_msg(struct irq_data *data, > > msg->address_lo = lower_32_bits(msi_msg); > msg->address_hi = upper_32_bits(msi_msg); > - msg->data = data->irq; > + msg->data = data->hwirq; > } > > static int advk_msi_set_affinity(struct irq_data *irq_data, > @@ -878,15 +879,11 @@ static int advk_msi_irq_domain_alloc(struct irq_domain *domain, > int hwirq, i; > > mutex_lock(&pcie->msi_used_lock); > - hwirq = bitmap_find_next_zero_area(pcie->msi_used, MSI_IRQ_NUM, > - 0, nr_irqs, 0); > - if (hwirq >= MSI_IRQ_NUM) { > - mutex_unlock(&pcie->msi_used_lock); > - return -ENOSPC; > - } > - > - bitmap_set(pcie->msi_used, hwirq, nr_irqs); > + hwirq = bitmap_find_free_region(pcie->msi_used, MSI_IRQ_NUM, > + order_base_2(nr_irqs)); > mutex_unlock(&pcie->msi_used_lock); > + if (hwirq < 0) > + return -ENOSPC; > > for (i = 0; i < nr_irqs; i++) > irq_domain_set_info(domain, virq + i, hwirq + i, > @@ -894,7 +891,7 @@ static int advk_msi_irq_domain_alloc(struct irq_domain *domain, > domain->host_data, handle_simple_irq, > NULL, NULL); > > - return hwirq; > + return 0; > } > > static void advk_msi_irq_domain_free(struct irq_domain *domain, > @@ -904,7 +901,7 @@ static void advk_msi_irq_domain_free(struct irq_domain *domain, > struct advk_pcie *pcie = domain->host_data; > > mutex_lock(&pcie->msi_used_lock); > - bitmap_clear(pcie->msi_used, d->hwirq, nr_irqs); > + bitmap_release_region(pcie->msi_used, d->hwirq, order_base_2(nr_irqs)); > mutex_unlock(&pcie->msi_used_lock); > } > > @@ -1048,6 +1045,7 @@ static void advk_pcie_handle_msi(struct advk_pcie *pcie) > { > u32 msi_val, msi_mask, msi_status, msi_idx; > u16 msi_data; > + int virq; > > msi_mask = advk_readl(pcie, PCIE_MSI_MASK_REG); > msi_val = advk_readl(pcie, PCIE_MSI_STATUS_REG); > @@ -1057,9 +1055,17 @@ static void advk_pcie_handle_msi(struct advk_pcie *pcie) > if (!(BIT(msi_idx) & msi_status)) > continue; > > + /* > + * msi_idx contains bits [4:0] of the msi_data and msi_data > + * contains 16bit MSI interrupt number from MSI inner domain > + */ > advk_writel(pcie, BIT(msi_idx), PCIE_MSI_STATUS_REG); > - msi_data = advk_readl(pcie, PCIE_MSI_PAYLOAD_REG) & 0xFF; > - generic_handle_irq(msi_data); > + msi_data = advk_readl(pcie, PCIE_MSI_PAYLOAD_REG) & PCIE_MSI_DATA_MASK; Can this be moved to a separate patch? It seems like this patch should only focus on correctly dealing with the irq/hwirq issues. > + virq = irq_find_mapping(pcie->msi_inner_domain, msi_data); > + if (virq) > + generic_handle_irq(virq); > + else > + dev_err(&pcie->pdev->dev, "unexpected MSI 0x%04hx\n", msi_data); Same concern about the unmitigated screaming. Thanks, M. -- Without deviation from the norm, progress is not possible.