From patchwork Wed Mar 9 19:21:19 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Hector Martin X-Patchwork-Id: 12775500 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 bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 5126AC433EF for ; Wed, 9 Mar 2022 19:23:55 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To: Message-Id:Date:Subject:Cc:To:From:Reply-To:Content-ID:Content-Description: Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID: List-Owner; bh=F1ZPwWZdc40BEeQ/EuAudgsCU+QTKh3DlTSeQvd9nUU=; b=H3QpSkuI/sfLMs lmNFK5Vls2TIsF6ikb8Opbum3FlSh2L3WXv9dE2Yyc8QVh8PMUUvuisqKISfTonomdKXsfKegcWv2 o4JijlEtzW05LZ0V8S4xIgfN46yBeRPNL125NeqwzlsNcLbySAgI9JkXauGQWNRzUT6Go4No9zCLq kF7wF2R+6dDFvVipCePbFPzdZW832QRbsSmVDxRH8+1nj+CPI9/WUZ5pcQSN/fusBXe42c9aJ/FYA m7X045iFG+V6+gZxNXB03O/1O5NxizSSvWsYfSEHqfm8Csi4dh5McAn0kKhQP6TwUv9Wf8M1tabJ5 9TPI8BtbQG6hH3w8wLlQ==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.94.2 #2 (Red Hat Linux)) id 1nS1tV-00AEiL-Qb; Wed, 09 Mar 2022 19:22:45 +0000 Received: from marcansoft.com ([2a01:298:fe:f::2] helo=mail.marcansoft.com) by bombadil.infradead.org with esmtps (Exim 4.94.2 #2 (Red Hat Linux)) id 1nS1sz-00AEXO-SW for linux-arm-kernel@lists.infradead.org; Wed, 09 Mar 2022 19:22:15 +0000 Received: from [127.0.0.1] (localhost [127.0.0.1]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) (Authenticated sender: hector@marcansoft.com) by mail.marcansoft.com (Postfix) with ESMTPSA id 5B51F425C8; Wed, 9 Mar 2022 19:22:09 +0000 (UTC) From: Hector Martin To: Thomas Gleixner , Marc Zyngier , Rob Herring Cc: Hector Martin , Sven Peter , Alyssa Rosenzweig , Mark Kettenis , linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, devicetree@vger.kernel.org Subject: [PATCH v3 3/7] irqchip/apple-aic: Add Fast IPI support Date: Thu, 10 Mar 2022 04:21:19 +0900 Message-Id: <20220309192123.152028-4-marcan@marcan.st> X-Mailer: git-send-email 2.33.0 In-Reply-To: <20220309192123.152028-1-marcan@marcan.st> References: <20220309192123.152028-1-marcan@marcan.st> MIME-Version: 1.0 X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20220309_112214_104086_143E76D5 X-CRM114-Status: GOOD ( 29.43 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org The newer AICv2 present in t600x SoCs does not have legacy IPI support at all. Since t8103 also supports Fast IPIs, implement support for this first. The legacy IPI code is left as a fallback, so it can be potentially used by older SoCs in the future. The vIPI code is shared; only the IPI firing/acking bits change for Fast IPIs. Signed-off-by: Hector Martin --- drivers/irqchip/irq-apple-aic.c | 122 ++++++++++++++++++++++++++++---- 1 file changed, 109 insertions(+), 13 deletions(-) diff --git a/drivers/irqchip/irq-apple-aic.c b/drivers/irqchip/irq-apple-aic.c index 38091ebb9403..613e0ebdabdc 100644 --- a/drivers/irqchip/irq-apple-aic.c +++ b/drivers/irqchip/irq-apple-aic.c @@ -24,7 +24,7 @@ * - Default "this CPU" register view and explicit per-CPU views * * In addition, this driver also handles FIQs, as these are routed to the same - * IRQ vector. These are used for Fast IPIs (TODO), the ARMv8 timer IRQs, and + * IRQ vector. These are used for Fast IPIs, the ARMv8 timer IRQs, and * performance counters (TODO). * * Implementation notes: @@ -52,9 +52,11 @@ #include #include #include +#include #include #include #include +#include #include #include #include @@ -106,7 +108,6 @@ /* * IMP-DEF sysregs that control FIQ sources - * Note: sysreg-based IPIs are not supported yet. */ /* Core PMC control register */ @@ -155,6 +156,10 @@ #define SYS_IMP_APL_UPMSR_EL1 sys_reg(3, 7, 15, 6, 4) #define UPMSR_IACT BIT(0) +/* MPIDR fields */ +#define MPIDR_CPU(x) MPIDR_AFFINITY_LEVEL(x, 0) +#define MPIDR_CLUSTER(x) MPIDR_AFFINITY_LEVEL(x, 1) + #define AIC_NR_FIQ 4 #define AIC_NR_SWIPI 32 @@ -173,11 +178,44 @@ #define AIC_TMR_EL02_PHYS AIC_TMR_GUEST_PHYS #define AIC_TMR_EL02_VIRT AIC_TMR_GUEST_VIRT +DEFINE_STATIC_KEY_TRUE(use_fast_ipi); + +struct aic_info { + int version; + + /* Features */ + bool fast_ipi; +}; + +static const struct aic_info aic1_info = { + .version = 1, +}; + +static const struct aic_info aic1_fipi_info = { + .version = 1, + + .fast_ipi = true, +}; + +static const struct of_device_id aic_info_match[] = { + { + .compatible = "apple,t8103-aic", + .data = &aic1_fipi_info, + }, + { + .compatible = "apple,aic", + .data = &aic1_info, + }, + {} +}; + struct aic_irq_chip { void __iomem *base; struct irq_domain *hw_domain; struct irq_domain *ipi_domain; int nr_hw; + + struct aic_info info; }; static DEFINE_PER_CPU(uint32_t, aic_fiq_unmasked); @@ -386,8 +424,12 @@ static void __exception_irq_entry aic_handle_fiq(struct pt_regs *regs) */ if (read_sysreg_s(SYS_IMP_APL_IPI_SR_EL1) & IPI_SR_PENDING) { - pr_err_ratelimited("Fast IPI fired. Acking.\n"); - write_sysreg_s(IPI_SR_PENDING, SYS_IMP_APL_IPI_SR_EL1); + if (static_branch_likely(&use_fast_ipi)) { + aic_handle_ipi(regs); + } else { + pr_err_ratelimited("Fast IPI fired. Acking.\n"); + write_sysreg_s(IPI_SR_PENDING, SYS_IMP_APL_IPI_SR_EL1); + } } if (TIMER_FIRING(read_sysreg(cntp_ctl_el0))) @@ -563,6 +605,22 @@ static const struct irq_domain_ops aic_irq_domain_ops = { * IPI irqchip */ +static void aic_ipi_send_fast(int cpu) +{ + u64 mpidr = cpu_logical_map(cpu); + u64 my_mpidr = read_cpuid_mpidr(); + u64 cluster = MPIDR_CLUSTER(mpidr); + u64 idx = MPIDR_CPU(mpidr); + + if (MPIDR_CLUSTER(my_mpidr) == cluster) + write_sysreg_s(FIELD_PREP(IPI_RR_CPU, idx), + SYS_IMP_APL_IPI_RR_LOCAL_EL1); + else + write_sysreg_s(FIELD_PREP(IPI_RR_CPU, idx) | FIELD_PREP(IPI_RR_CLUSTER, cluster), + SYS_IMP_APL_IPI_RR_GLOBAL_EL1); + isb(); +} + static void aic_ipi_mask(struct irq_data *d) { u32 irq_bit = BIT(irqd_to_hwirq(d)); @@ -588,8 +646,12 @@ static void aic_ipi_unmask(struct irq_data *d) * If a pending vIPI was unmasked, raise a HW IPI to ourselves. * No barriers needed here since this is a self-IPI. */ - if (atomic_read(this_cpu_ptr(&aic_vipi_flag)) & irq_bit) - aic_ic_write(ic, AIC_IPI_SEND, AIC_IPI_SEND_CPU(smp_processor_id())); + if (atomic_read(this_cpu_ptr(&aic_vipi_flag)) & irq_bit) { + if (static_branch_likely(&use_fast_ipi)) + aic_ipi_send_fast(smp_processor_id()); + else + aic_ic_write(ic, AIC_IPI_SEND, AIC_IPI_SEND_CPU(smp_processor_id())); + } } static void aic_ipi_send_mask(struct irq_data *d, const struct cpumask *mask) @@ -617,8 +679,12 @@ static void aic_ipi_send_mask(struct irq_data *d, const struct cpumask *mask) smp_mb__after_atomic(); if (!(pending & irq_bit) && - (atomic_read(per_cpu_ptr(&aic_vipi_enable, cpu)) & irq_bit)) - send |= AIC_IPI_SEND_CPU(cpu); + (atomic_read(per_cpu_ptr(&aic_vipi_enable, cpu)) & irq_bit)) { + if (static_branch_likely(&use_fast_ipi)) + aic_ipi_send_fast(cpu); + else + send |= AIC_IPI_SEND_CPU(cpu); + } } /* @@ -650,8 +716,16 @@ static void aic_handle_ipi(struct pt_regs *regs) /* * Ack the IPI. We need to order this after the AIC event read, but * that is enforced by normal MMIO ordering guarantees. + * + * For the Fast IPI case, this needs to be ordered before the vIPI + * handling below, so we need to isb(); */ - aic_ic_write(aic_irqc, AIC_IPI_ACK, AIC_IPI_OTHER); + if (static_branch_likely(&use_fast_ipi)) { + write_sysreg_s(IPI_SR_PENDING, SYS_IMP_APL_IPI_SR_EL1); + isb(); + } else { + aic_ic_write(aic_irqc, AIC_IPI_ACK, AIC_IPI_OTHER); + } /* * The mask read does not need to be ordered. Only we can change @@ -679,7 +753,8 @@ static void aic_handle_ipi(struct pt_regs *regs) * No ordering needed here; at worst this just changes the timing of * when the next IPI will be delivered. */ - aic_ic_write(aic_irqc, AIC_IPI_MASK_CLR, AIC_IPI_OTHER); + if (!static_branch_likely(&use_fast_ipi)) + aic_ic_write(aic_irqc, AIC_IPI_MASK_CLR, AIC_IPI_OTHER); } static int aic_ipi_alloc(struct irq_domain *d, unsigned int virq, @@ -776,10 +851,15 @@ static int aic_init_cpu(unsigned int cpu) /* * Always keep IPIs unmasked at the hardware level (except auto-masking * by AIC during processing). We manage masks at the vIPI level. + * These registers only exist on AICv1, AICv2 always uses fast IPIs. */ aic_ic_write(aic_irqc, AIC_IPI_ACK, AIC_IPI_SELF | AIC_IPI_OTHER); - aic_ic_write(aic_irqc, AIC_IPI_MASK_SET, AIC_IPI_SELF); - aic_ic_write(aic_irqc, AIC_IPI_MASK_CLR, AIC_IPI_OTHER); + if (static_branch_likely(&use_fast_ipi)) { + aic_ic_write(aic_irqc, AIC_IPI_MASK_SET, AIC_IPI_SELF | AIC_IPI_OTHER); + } else { + aic_ic_write(aic_irqc, AIC_IPI_MASK_SET, AIC_IPI_SELF); + aic_ic_write(aic_irqc, AIC_IPI_MASK_CLR, AIC_IPI_OTHER); + } /* Initialize the local mask state */ __this_cpu_write(aic_fiq_unmasked, 0); @@ -799,6 +879,7 @@ static int __init aic_of_ic_init(struct device_node *node, struct device_node *p void __iomem *regs; u32 info; struct aic_irq_chip *irqc; + const struct of_device_id *match; regs = of_iomap(node, 0); if (WARN_ON(!regs)) @@ -808,12 +889,24 @@ static int __init aic_of_ic_init(struct device_node *node, struct device_node *p if (!irqc) return -ENOMEM; - aic_irqc = irqc; irqc->base = regs; + match = of_match_node(aic_info_match, node); + if (!match) + return -ENODEV; + + irqc->info = *(struct aic_info *)match->data; + + aic_irqc = irqc; + info = aic_ic_read(irqc, AIC_INFO); irqc->nr_hw = FIELD_GET(AIC_INFO_NR_HW, info); + if (irqc->info.fast_ipi) + static_branch_enable(&use_fast_ipi); + else + static_branch_disable(&use_fast_ipi); + irqc->hw_domain = irq_domain_create_linear(of_node_to_fwnode(node), irqc->nr_hw + AIC_NR_FIQ, &aic_irq_domain_ops, irqc); @@ -845,6 +938,9 @@ static int __init aic_of_ic_init(struct device_node *node, struct device_node *p if (!is_kernel_in_hyp_mode()) pr_info("Kernel running in EL1, mapping interrupts"); + if (static_branch_likely(&use_fast_ipi)) + pr_info("Using Fast IPIs"); + cpuhp_setup_state(CPUHP_AP_IRQ_APPLE_AIC_STARTING, "irqchip/apple-aic/ipi:starting", aic_init_cpu, NULL);