From patchwork Sun Jan 16 02:25:48 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Marek Vasut X-Patchwork-Id: 12714449 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 26148C433FE for ; Sun, 16 Jan 2022 02:26:19 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233989AbiAPC0Q (ORCPT ); Sat, 15 Jan 2022 21:26:16 -0500 Received: from phobos.denx.de ([85.214.62.61]:41316 "EHLO phobos.denx.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S232795AbiAPC0Q (ORCPT ); Sat, 15 Jan 2022 21:26:16 -0500 Received: from tr.lan (ip-89-176-112-137.net.upcbroadband.cz [89.176.112.137]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)) (No client certificate requested) (Authenticated sender: marex@denx.de) by phobos.denx.de (Postfix) with ESMTPSA id E66A580F6F; Sun, 16 Jan 2022 03:26:13 +0100 (CET) From: marek.vasut@gmail.com To: linux-pci@vger.kernel.org Cc: Marek Vasut , Arnd Bergmann , Bjorn Helgaas , Geert Uytterhoeven , =?utf-8?q?Krzysztof_Wilczy?= =?utf-8?q?=C5=84ski?= , Lorenzo Pieralisi , Wolfram Sang , Yoshihiro Shimoda , linux-renesas-soc@vger.kernel.org Subject: [PATCH 1/2] PCI: rcar: Finish transition to L1 state in rcar_pcie_config_access() Date: Sun, 16 Jan 2022 03:25:48 +0100 Message-Id: <20220116022549.456486-1-marek.vasut@gmail.com> X-Mailer: git-send-email 2.34.1 MIME-Version: 1.0 X-Virus-Scanned: clamav-milter 0.103.2 at phobos.denx.de X-Virus-Status: Clean Precedence: bulk List-ID: X-Mailing-List: linux-pci@vger.kernel.org From: Marek Vasut In case the controller is transitioning to L1 in rcar_pcie_config_access(), any read/write access to PCIECDR triggers asynchronous external abort. This is because the transition to L1 link state must be manually finished by the driver. The PCIe IP can transition back from L1 state to L0 on its own. Avoid triggering the abort in rcar_pcie_config_access() by checking whether the controller is in the transition state, and if so, finish the transition right away. This prevents a lot of unnecessary exceptions, although not all of them. Signed-off-by: Marek Vasut Cc: Arnd Bergmann Cc: Bjorn Helgaas Cc: Geert Uytterhoeven Cc: Krzysztof Wilczyński Cc: Lorenzo Pieralisi Cc: Wolfram Sang Cc: Yoshihiro Shimoda Cc: linux-renesas-soc@vger.kernel.org Reported-by: kernel test robot --- drivers/pci/controller/pcie-rcar-host.c | 61 ++++++++++++++++--------- 1 file changed, 40 insertions(+), 21 deletions(-) diff --git a/drivers/pci/controller/pcie-rcar-host.c b/drivers/pci/controller/pcie-rcar-host.c index 38b6e02edfa9..0be58a91ddea 100644 --- a/drivers/pci/controller/pcie-rcar-host.c +++ b/drivers/pci/controller/pcie-rcar-host.c @@ -54,6 +54,34 @@ static void __iomem *pcie_base; * device is runtime suspended or not. */ static struct device *pcie_dev; + +static DEFINE_SPINLOCK(pmsr_lock); +static int rcar_pcie_wakeup(struct device *pcie_dev, void __iomem *pcie_base) +{ + u32 pmsr, val; + int ret = 0; + + if (!pcie_base || pm_runtime_suspended(pcie_dev)) + return 1; + + pmsr = readl(pcie_base + PMSR); + + /* + * Test if the PCIe controller received PM_ENTER_L1 DLLP and + * the PCIe controller is not in L1 link state. If true, apply + * fix, which will put the controller into L1 link state, from + * which it can return to L0s/L0 on its own. + */ + if ((pmsr & PMEL1RX) && ((pmsr & PMSTATE) != PMSTATE_L1)) { + writel(L1IATN, pcie_base + PMCTLR); + ret = readl_poll_timeout_atomic(pcie_base + PMSR, val, + val & L1FAEG, 10, 1000); + WARN(ret, "Timeout waiting for L1 link state, ret=%d\n", ret); + writel(L1FAEG | PMEL1RX, pcie_base + PMSR); + } + + return ret; +} #endif /* Structure representing the PCIe interface */ @@ -85,6 +113,15 @@ static int rcar_pcie_config_access(struct rcar_pcie_host *host, { struct rcar_pcie *pcie = &host->pcie; unsigned int dev, func, reg, index; + unsigned long flags; + int ret; + + /* Wake the bus up in case it is in L1 state. */ + spin_lock_irqsave(&pmsr_lock, flags); + ret = rcar_pcie_wakeup(pcie->dev, pcie->base); + spin_unlock_irqrestore(&pmsr_lock, flags); + if (ret) + return ret; dev = PCI_SLOT(devfn); func = PCI_FUNC(devfn); @@ -1050,36 +1087,18 @@ static struct platform_driver rcar_pcie_driver = { }; #ifdef CONFIG_ARM -static DEFINE_SPINLOCK(pmsr_lock); static int rcar_pcie_aarch32_abort_handler(unsigned long addr, unsigned int fsr, struct pt_regs *regs) { unsigned long flags; - u32 pmsr, val; int ret = 0; spin_lock_irqsave(&pmsr_lock, flags); - if (!pcie_base || pm_runtime_suspended(pcie_dev)) { - ret = 1; + ret = rcar_pcie_wakeup(pcie_dev, pcie_base); + spin_unlock_irqrestore(&pmsr_lock, flags); + if (ret) goto unlock_exit; - } - - pmsr = readl(pcie_base + PMSR); - - /* - * Test if the PCIe controller received PM_ENTER_L1 DLLP and - * the PCIe controller is not in L1 link state. If true, apply - * fix, which will put the controller into L1 link state, from - * which it can return to L0s/L0 on its own. - */ - if ((pmsr & PMEL1RX) && ((pmsr & PMSTATE) != PMSTATE_L1)) { - writel(L1IATN, pcie_base + PMCTLR); - ret = readl_poll_timeout_atomic(pcie_base + PMSR, val, - val & L1FAEG, 10, 1000); - WARN(ret, "Timeout waiting for L1 link state, ret=%d\n", ret); - writel(L1FAEG | PMEL1RX, pcie_base + PMSR); - } unlock_exit: spin_unlock_irqrestore(&pmsr_lock, flags);