From patchwork Wed Jan 16 18:01:01 2013 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Christoffer Dall X-Patchwork-Id: 1993041 Return-Path: X-Original-To: patchwork-linux-arm@patchwork.kernel.org Delivered-To: patchwork-process-083081@patchwork1.kernel.org Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) by patchwork1.kernel.org (Postfix) with ESMTP id B77C53FDD1 for ; Wed, 16 Jan 2013 18:16:15 +0000 (UTC) Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.76 #1 (Red Hat Linux)) id 1TvXSY-0002tB-HH; Wed, 16 Jan 2013 18:11:35 +0000 Received: from mail-vb0-f54.google.com ([209.85.212.54]) by merlin.infradead.org with esmtps (Exim 4.76 #1 (Red Hat Linux)) id 1TvXIO-0004m9-S2 for linux-arm-kernel@lists.infradead.org; Wed, 16 Jan 2013 18:01:09 +0000 Received: by mail-vb0-f54.google.com with SMTP id l1so1614346vba.41 for ; Wed, 16 Jan 2013 10:01:04 -0800 (PST) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=x-received:subject:to:from:cc:date:message-id:in-reply-to :references:user-agent:mime-version:content-type :content-transfer-encoding:x-gm-message-state; bh=ku1/JS1Z8tjFt75LfdX6Up6+p4n3Y2edJTZ11f7G6m4=; b=GjzfkeTAQDh7uQzGOzagqLtUVjAZ3dlbb4TzcPR43HBFYnN4xuTmjxEqioJVQjtVA5 VyjJY1v6BW+m+M/TI7lbwkYAskgsTm50MX1/b4a6RD26Xe3RvTmiiEqy2A0V++Du25Zn ukG4cvsNARqfA4lyffG0BrFtThzTKjjBXuObS3Bm3qoq5/FqPvhHv62a8zl+PmytB7lG bd6YqnFte9qlgG8UwRzHV3ECY7aCyDyfKbnb3OM6iRCjlRq9+JKql2qKxtd8dIaEEfLy wuxL4PDhQQOkZC67O/hZ9KhtXxjuJFeua25oxpBRL0niVUU2Nl+e48NFScyxCz08CGZC HP4A== X-Received: by 10.220.238.146 with SMTP id ks18mr2163345vcb.58.1358359263976; Wed, 16 Jan 2013 10:01:03 -0800 (PST) Received: from [127.0.1.1] (pool-72-80-83-148.nycmny.fios.verizon.net. [72.80.83.148]) by mx.google.com with ESMTPS id ly6sm9734424veb.3.2013.01.16.10.01.01 (version=TLSv1.1 cipher=ECDHE-RSA-RC4-SHA bits=128/128); Wed, 16 Jan 2013 10:01:02 -0800 (PST) Subject: [PATCH v6 07/13] ARM: KVM: VGIC distributor handling To: kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, kvmarm@lists.cs.columbia.edu From: Christoffer Dall Date: Wed, 16 Jan 2013 13:01:01 -0500 Message-ID: <20130116180101.29393.48572.stgit@ubuntu> In-Reply-To: <20130116180013.29393.49165.stgit@ubuntu> References: <20130116180013.29393.49165.stgit@ubuntu> User-Agent: StGit/0.15 MIME-Version: 1.0 X-Gm-Message-State: ALoCoQm/eUoW7zy9U2Y+JvTzQ/fNUHheEOL+12ouUnQQ1xu1Jb/PB2dvJI71HfuHi6FoB8nOytUt X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20130116_130106_562128_23DF91DE X-CRM114-Status: GOOD ( 27.05 ) X-Spam-Score: -2.6 (--) X-Spam-Report: SpamAssassin version 3.3.2 on merlin.infradead.org summary: Content analysis details: (-2.6 points) pts rule name description ---- ---------------------- -------------------------------------------------- -0.7 RCVD_IN_DNSWL_LOW RBL: Sender listed at http://www.dnswl.org/, low trust [209.85.212.54 listed in list.dnswl.org] -1.9 BAYES_00 BODY: Bayes spam probability is 0 to 1% [score: 0.0000] Cc: Marc Zyngier , Will Deacon X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: linux-arm-kernel-bounces@lists.infradead.org Errors-To: linux-arm-kernel-bounces+patchwork-linux-arm=patchwork.kernel.org@lists.infradead.org From: Marc Zyngier Add the GIC distributor emulation code. A number of the GIC features are simply ignored as they are not required to boot a Linux guest. Reviewed-by: Will Deacon Signed-off-by: Marc Zyngier Signed-off-by: Christoffer Dall --- arch/arm/include/asm/kvm_vgic.h | 82 +++++ arch/arm/kvm/vgic.c | 596 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 677 insertions(+), 1 deletion(-) diff --git a/arch/arm/include/asm/kvm_vgic.h b/arch/arm/include/asm/kvm_vgic.h index 6b00735..b1ea1b6 100644 --- a/arch/arm/include/asm/kvm_vgic.h +++ b/arch/arm/include/asm/kvm_vgic.h @@ -19,12 +19,94 @@ #ifndef __ASM_ARM_KVM_VGIC_H #define __ASM_ARM_KVM_VGIC_H +#include +#include +#include +#include +#include +#include #include +#define VGIC_NR_IRQS 128 +#define VGIC_NR_SGIS 16 +#define VGIC_NR_PPIS 16 +#define VGIC_NR_PRIVATE_IRQS (VGIC_NR_SGIS + VGIC_NR_PPIS) +#define VGIC_NR_SHARED_IRQS (VGIC_NR_IRQS - VGIC_NR_PRIVATE_IRQS) +#define VGIC_MAX_CPUS KVM_MAX_VCPUS + +/* Sanity checks... */ +#if (VGIC_MAX_CPUS > 8) +#error Invalid number of CPU interfaces +#endif + +#if (VGIC_NR_IRQS & 31) +#error "VGIC_NR_IRQS must be a multiple of 32" +#endif + +#if (VGIC_NR_IRQS > 1024) +#error "VGIC_NR_IRQS must be <= 1024" +#endif + +/* + * The GIC distributor registers describing interrupts have two parts: + * - 32 per-CPU interrupts (SGI + PPI) + * - a bunch of shared interrupts (SPI) + */ +struct vgic_bitmap { + union { + u32 reg[VGIC_NR_PRIVATE_IRQS / 32]; + DECLARE_BITMAP(reg_ul, VGIC_NR_PRIVATE_IRQS); + } percpu[VGIC_MAX_CPUS]; + union { + u32 reg[VGIC_NR_SHARED_IRQS / 32]; + DECLARE_BITMAP(reg_ul, VGIC_NR_SHARED_IRQS); + } shared; +}; + +struct vgic_bytemap { + u32 percpu[VGIC_MAX_CPUS][VGIC_NR_PRIVATE_IRQS / 4]; + u32 shared[VGIC_NR_SHARED_IRQS / 4]; +}; + struct vgic_dist { +#ifdef CONFIG_KVM_ARM_VGIC + spinlock_t lock; + + /* Virtual control interface mapping */ + void __iomem *vctrl_base; + /* Distributor and vcpu interface mapping in the guest */ phys_addr_t vgic_dist_base; phys_addr_t vgic_cpu_base; + + /* Distributor enabled */ + u32 enabled; + + /* Interrupt enabled (one bit per IRQ) */ + struct vgic_bitmap irq_enabled; + + /* Interrupt 'pin' level */ + struct vgic_bitmap irq_state; + + /* Level-triggered interrupt in progress */ + struct vgic_bitmap irq_active; + + /* Interrupt priority. Not used yet. */ + struct vgic_bytemap irq_priority; + + /* Level/edge triggered */ + struct vgic_bitmap irq_cfg; + + /* Source CPU per SGI and target CPU */ + u8 irq_sgi_sources[VGIC_MAX_CPUS][VGIC_NR_SGIS]; + + /* Target CPU for each IRQ */ + u8 irq_spi_cpu[VGIC_NR_SHARED_IRQS]; + struct vgic_bitmap irq_spi_target[VGIC_MAX_CPUS]; + + /* Bitmap indicating which CPU has something pending */ + unsigned long irq_pending_on_cpu; +#endif }; struct vgic_cpu { diff --git a/arch/arm/kvm/vgic.c b/arch/arm/kvm/vgic.c index b333b58..815069f 100644 --- a/arch/arm/kvm/vgic.c +++ b/arch/arm/kvm/vgic.c @@ -22,6 +22,43 @@ #include #include +/* + * How the whole thing works (courtesy of Christoffer Dall): + * + * - At any time, the dist->irq_pending_on_cpu is the oracle that knows if + * something is pending + * - VGIC pending interrupts are stored on the vgic.irq_state vgic + * bitmap (this bitmap is updated by both user land ioctls and guest + * mmio ops, and other in-kernel peripherals such as the + * arch. timers) and indicate the 'wire' state. + * - Every time the bitmap changes, the irq_pending_on_cpu oracle is + * recalculated + * - To calculate the oracle, we need info for each cpu from + * compute_pending_for_cpu, which considers: + * - PPI: dist->irq_state & dist->irq_enable + * - SPI: dist->irq_state & dist->irq_enable & dist->irq_spi_target + * - irq_spi_target is a 'formatted' version of the GICD_ICFGR + * registers, stored on each vcpu. We only keep one bit of + * information per interrupt, making sure that only one vcpu can + * accept the interrupt. + * - The same is true when injecting an interrupt, except that we only + * consider a single interrupt at a time. The irq_spi_cpu array + * contains the target CPU for each SPI. + * + * The handling of level interrupts adds some extra complexity. We + * need to track when the interrupt has been EOIed, so we can sample + * the 'line' again. This is achieved as such: + * + * - When a level interrupt is moved onto a vcpu, the corresponding + * bit in irq_active is set. As long as this bit is set, the line + * will be ignored for further interrupts. The interrupt is injected + * into the vcpu with the GICH_LR_EOI bit set (generate a + * maintenance interrupt on EOI). + * - When the interrupt is EOIed, the maintenance interrupt fires, + * and clears the corresponding bit in irq_active. This allow the + * interrupt line to be sampled again. + */ + #define VGIC_ADDR_UNDEF (-1) #define IS_VGIC_ADDR_UNDEF(_x) ((_x) == VGIC_ADDR_UNDEF) @@ -34,6 +71,119 @@ #define ACCESS_WRITE_VALUE (3 << 1) #define ACCESS_WRITE_MASK(x) ((x) & (3 << 1)) +static void vgic_update_state(struct kvm *kvm); +static void vgic_dispatch_sgi(struct kvm_vcpu *vcpu, u32 reg); + +static u32 *vgic_bitmap_get_reg(struct vgic_bitmap *x, + int cpuid, u32 offset) +{ + offset >>= 2; + if (!offset) + return x->percpu[cpuid].reg; + else + return x->shared.reg + offset - 1; +} + +static int vgic_bitmap_get_irq_val(struct vgic_bitmap *x, + int cpuid, int irq) +{ + if (irq < VGIC_NR_PRIVATE_IRQS) + return test_bit(irq, x->percpu[cpuid].reg_ul); + + return test_bit(irq - VGIC_NR_PRIVATE_IRQS, x->shared.reg_ul); +} + +static void vgic_bitmap_set_irq_val(struct vgic_bitmap *x, int cpuid, + int irq, int val) +{ + unsigned long *reg; + + if (irq < VGIC_NR_PRIVATE_IRQS) { + reg = x->percpu[cpuid].reg_ul; + } else { + reg = x->shared.reg_ul; + irq -= VGIC_NR_PRIVATE_IRQS; + } + + if (val) + set_bit(irq, reg); + else + clear_bit(irq, reg); +} + +static unsigned long *vgic_bitmap_get_cpu_map(struct vgic_bitmap *x, int cpuid) +{ + if (unlikely(cpuid >= VGIC_MAX_CPUS)) + return NULL; + return x->percpu[cpuid].reg_ul; +} + +static unsigned long *vgic_bitmap_get_shared_map(struct vgic_bitmap *x) +{ + return x->shared.reg_ul; +} + +static u32 *vgic_bytemap_get_reg(struct vgic_bytemap *x, int cpuid, u32 offset) +{ + offset >>= 2; + BUG_ON(offset > (VGIC_NR_IRQS / 4)); + if (offset < 4) + return x->percpu[cpuid] + offset; + else + return x->shared + offset - 8; +} + +#define VGIC_CFG_LEVEL 0 +#define VGIC_CFG_EDGE 1 + +static bool vgic_irq_is_edge(struct kvm_vcpu *vcpu, int irq) +{ + struct vgic_dist *dist = &vcpu->kvm->arch.vgic; + int irq_val; + + irq_val = vgic_bitmap_get_irq_val(&dist->irq_cfg, vcpu->vcpu_id, irq); + return irq_val == VGIC_CFG_EDGE; +} + +static int vgic_irq_is_enabled(struct kvm_vcpu *vcpu, int irq) +{ + struct vgic_dist *dist = &vcpu->kvm->arch.vgic; + + return vgic_bitmap_get_irq_val(&dist->irq_enabled, vcpu->vcpu_id, irq); +} + +static void vgic_dist_irq_set(struct kvm_vcpu *vcpu, int irq) +{ + struct vgic_dist *dist = &vcpu->kvm->arch.vgic; + + vgic_bitmap_set_irq_val(&dist->irq_state, vcpu->vcpu_id, irq, 1); +} + +static void vgic_dist_irq_clear(struct kvm_vcpu *vcpu, int irq) +{ + struct vgic_dist *dist = &vcpu->kvm->arch.vgic; + + vgic_bitmap_set_irq_val(&dist->irq_state, vcpu->vcpu_id, irq, 0); +} + +static void vgic_cpu_irq_set(struct kvm_vcpu *vcpu, int irq) +{ + if (irq < VGIC_NR_PRIVATE_IRQS) + set_bit(irq, vcpu->arch.vgic_cpu.pending_percpu); + else + set_bit(irq - VGIC_NR_PRIVATE_IRQS, + vcpu->arch.vgic_cpu.pending_shared); +} + +static void vgic_cpu_irq_clear(struct kvm_vcpu *vcpu, int irq) +{ + if (irq < VGIC_NR_PRIVATE_IRQS) + clear_bit(irq, vcpu->arch.vgic_cpu.pending_percpu); + else + clear_bit(irq - VGIC_NR_PRIVATE_IRQS, + vcpu->arch.vgic_cpu.pending_shared); +} + static u32 mmio_data_read(struct kvm_exit_mmio *mmio, u32 mask) { return *((u32 *)mmio->data) & mask; @@ -105,6 +255,291 @@ static void vgic_reg_access(struct kvm_exit_mmio *mmio, u32 *reg, } } +static bool handle_mmio_misc(struct kvm_vcpu *vcpu, + struct kvm_exit_mmio *mmio, phys_addr_t offset) +{ + u32 reg; + u32 word_offset = offset & 3; + + switch (offset & ~3) { + case 0: /* CTLR */ + reg = vcpu->kvm->arch.vgic.enabled; + vgic_reg_access(mmio, ®, word_offset, + ACCESS_READ_VALUE | ACCESS_WRITE_VALUE); + if (mmio->is_write) { + vcpu->kvm->arch.vgic.enabled = reg & 1; + vgic_update_state(vcpu->kvm); + return true; + } + break; + + case 4: /* TYPER */ + reg = (atomic_read(&vcpu->kvm->online_vcpus) - 1) << 5; + reg |= (VGIC_NR_IRQS >> 5) - 1; + vgic_reg_access(mmio, ®, word_offset, + ACCESS_READ_VALUE | ACCESS_WRITE_IGNORED); + break; + + case 8: /* IIDR */ + reg = 0x4B00043B; + vgic_reg_access(mmio, ®, word_offset, + ACCESS_READ_VALUE | ACCESS_WRITE_IGNORED); + break; + } + + return false; +} + +static bool handle_mmio_raz_wi(struct kvm_vcpu *vcpu, + struct kvm_exit_mmio *mmio, phys_addr_t offset) +{ + vgic_reg_access(mmio, NULL, offset, + ACCESS_READ_RAZ | ACCESS_WRITE_IGNORED); + return false; +} + +static bool handle_mmio_set_enable_reg(struct kvm_vcpu *vcpu, + struct kvm_exit_mmio *mmio, + phys_addr_t offset) +{ + u32 *reg = vgic_bitmap_get_reg(&vcpu->kvm->arch.vgic.irq_enabled, + vcpu->vcpu_id, offset); + vgic_reg_access(mmio, reg, offset, + ACCESS_READ_VALUE | ACCESS_WRITE_SETBIT); + if (mmio->is_write) { + vgic_update_state(vcpu->kvm); + return true; + } + + return false; +} + +static bool handle_mmio_clear_enable_reg(struct kvm_vcpu *vcpu, + struct kvm_exit_mmio *mmio, + phys_addr_t offset) +{ + u32 *reg = vgic_bitmap_get_reg(&vcpu->kvm->arch.vgic.irq_enabled, + vcpu->vcpu_id, offset); + vgic_reg_access(mmio, reg, offset, + ACCESS_READ_VALUE | ACCESS_WRITE_CLEARBIT); + if (mmio->is_write) { + if (offset < 4) /* Force SGI enabled */ + *reg |= 0xffff; + vgic_update_state(vcpu->kvm); + return true; + } + + return false; +} + +static bool handle_mmio_set_pending_reg(struct kvm_vcpu *vcpu, + struct kvm_exit_mmio *mmio, + phys_addr_t offset) +{ + u32 *reg = vgic_bitmap_get_reg(&vcpu->kvm->arch.vgic.irq_state, + vcpu->vcpu_id, offset); + vgic_reg_access(mmio, reg, offset, + ACCESS_READ_VALUE | ACCESS_WRITE_SETBIT); + if (mmio->is_write) { + vgic_update_state(vcpu->kvm); + return true; + } + + return false; +} + +static bool handle_mmio_clear_pending_reg(struct kvm_vcpu *vcpu, + struct kvm_exit_mmio *mmio, + phys_addr_t offset) +{ + u32 *reg = vgic_bitmap_get_reg(&vcpu->kvm->arch.vgic.irq_state, + vcpu->vcpu_id, offset); + vgic_reg_access(mmio, reg, offset, + ACCESS_READ_VALUE | ACCESS_WRITE_CLEARBIT); + if (mmio->is_write) { + vgic_update_state(vcpu->kvm); + return true; + } + + return false; +} + +static bool handle_mmio_priority_reg(struct kvm_vcpu *vcpu, + struct kvm_exit_mmio *mmio, + phys_addr_t offset) +{ + u32 *reg = vgic_bytemap_get_reg(&vcpu->kvm->arch.vgic.irq_priority, + vcpu->vcpu_id, offset); + vgic_reg_access(mmio, reg, offset, + ACCESS_READ_VALUE | ACCESS_WRITE_VALUE); + return false; +} + +#define GICD_ITARGETSR_SIZE 32 +#define GICD_CPUTARGETS_BITS 8 +#define GICD_IRQS_PER_ITARGETSR (GICD_ITARGETSR_SIZE / GICD_CPUTARGETS_BITS) +static u32 vgic_get_target_reg(struct kvm *kvm, int irq) +{ + struct vgic_dist *dist = &kvm->arch.vgic; + struct kvm_vcpu *vcpu; + int i, c; + unsigned long *bmap; + u32 val = 0; + + irq -= VGIC_NR_PRIVATE_IRQS; + + kvm_for_each_vcpu(c, vcpu, kvm) { + bmap = vgic_bitmap_get_shared_map(&dist->irq_spi_target[c]); + for (i = 0; i < GICD_IRQS_PER_ITARGETSR; i++) + if (test_bit(irq + i, bmap)) + val |= 1 << (c + i * 8); + } + + return val; +} + +static void vgic_set_target_reg(struct kvm *kvm, u32 val, int irq) +{ + struct vgic_dist *dist = &kvm->arch.vgic; + struct kvm_vcpu *vcpu; + int i, c; + unsigned long *bmap; + u32 target; + + irq -= VGIC_NR_PRIVATE_IRQS; + + /* + * Pick the LSB in each byte. This ensures we target exactly + * one vcpu per IRQ. If the byte is null, assume we target + * CPU0. + */ + for (i = 0; i < GICD_IRQS_PER_ITARGETSR; i++) { + int shift = i * GICD_CPUTARGETS_BITS; + target = ffs((val >> shift) & 0xffU); + target = target ? (target - 1) : 0; + dist->irq_spi_cpu[irq + i] = target; + kvm_for_each_vcpu(c, vcpu, kvm) { + bmap = vgic_bitmap_get_shared_map(&dist->irq_spi_target[c]); + if (c == target) + set_bit(irq + i, bmap); + else + clear_bit(irq + i, bmap); + } + } +} + +static bool handle_mmio_target_reg(struct kvm_vcpu *vcpu, + struct kvm_exit_mmio *mmio, + phys_addr_t offset) +{ + u32 reg; + + /* We treat the banked interrupts targets as read-only */ + if (offset < 32) { + u32 roreg = 1 << vcpu->vcpu_id; + roreg |= roreg << 8; + roreg |= roreg << 16; + + vgic_reg_access(mmio, &roreg, offset, + ACCESS_READ_VALUE | ACCESS_WRITE_IGNORED); + return false; + } + + reg = vgic_get_target_reg(vcpu->kvm, offset & ~3U); + vgic_reg_access(mmio, ®, offset, + ACCESS_READ_VALUE | ACCESS_WRITE_VALUE); + if (mmio->is_write) { + vgic_set_target_reg(vcpu->kvm, reg, offset & ~3U); + vgic_update_state(vcpu->kvm); + return true; + } + + return false; +} + +static u32 vgic_cfg_expand(u16 val) +{ + u32 res = 0; + int i; + + /* + * Turn a 16bit value like abcd...mnop into a 32bit word + * a0b0c0d0...m0n0o0p0, which is what the HW cfg register is. + */ + for (i = 0; i < 16; i++) + res |= ((val >> i) & VGIC_CFG_EDGE) << (2 * i + 1); + + return res; +} + +static u16 vgic_cfg_compress(u32 val) +{ + u16 res = 0; + int i; + + /* + * Turn a 32bit word a0b0c0d0...m0n0o0p0 into 16bit value like + * abcd...mnop which is what we really care about. + */ + for (i = 0; i < 16; i++) + res |= ((val >> (i * 2 + 1)) & VGIC_CFG_EDGE) << i; + + return res; +} + +/* + * The distributor uses 2 bits per IRQ for the CFG register, but the + * LSB is always 0. As such, we only keep the upper bit, and use the + * two above functions to compress/expand the bits + */ +static bool handle_mmio_cfg_reg(struct kvm_vcpu *vcpu, + struct kvm_exit_mmio *mmio, phys_addr_t offset) +{ + u32 val; + u32 *reg = vgic_bitmap_get_reg(&vcpu->kvm->arch.vgic.irq_cfg, + vcpu->vcpu_id, offset >> 1); + if (offset & 2) + val = *reg >> 16; + else + val = *reg & 0xffff; + + val = vgic_cfg_expand(val); + vgic_reg_access(mmio, &val, offset, + ACCESS_READ_VALUE | ACCESS_WRITE_VALUE); + if (mmio->is_write) { + if (offset < 4) { + *reg = ~0U; /* Force PPIs/SGIs to 1 */ + return false; + } + + val = vgic_cfg_compress(val); + if (offset & 2) { + *reg &= 0xffff; + *reg |= val << 16; + } else { + *reg &= 0xffff << 16; + *reg |= val; + } + } + + return false; +} + +static bool handle_mmio_sgi_reg(struct kvm_vcpu *vcpu, + struct kvm_exit_mmio *mmio, phys_addr_t offset) +{ + u32 reg; + vgic_reg_access(mmio, ®, offset, + ACCESS_READ_RAZ | ACCESS_WRITE_VALUE); + if (mmio->is_write) { + vgic_dispatch_sgi(vcpu, reg); + vgic_update_state(vcpu->kvm); + return true; + } + + return false; +} + /* * I would have liked to use the kvm_bus_io_*() API instead, but it * cannot cope with banked registers (only the VM pointer is passed @@ -119,6 +554,66 @@ struct mmio_range { }; static const struct mmio_range vgic_ranges[] = { + { + .base = GIC_DIST_CTRL, + .len = 12, + .handle_mmio = handle_mmio_misc, + }, + { + .base = GIC_DIST_IGROUP, + .len = VGIC_NR_IRQS / 8, + .handle_mmio = handle_mmio_raz_wi, + }, + { + .base = GIC_DIST_ENABLE_SET, + .len = VGIC_NR_IRQS / 8, + .handle_mmio = handle_mmio_set_enable_reg, + }, + { + .base = GIC_DIST_ENABLE_CLEAR, + .len = VGIC_NR_IRQS / 8, + .handle_mmio = handle_mmio_clear_enable_reg, + }, + { + .base = GIC_DIST_PENDING_SET, + .len = VGIC_NR_IRQS / 8, + .handle_mmio = handle_mmio_set_pending_reg, + }, + { + .base = GIC_DIST_PENDING_CLEAR, + .len = VGIC_NR_IRQS / 8, + .handle_mmio = handle_mmio_clear_pending_reg, + }, + { + .base = GIC_DIST_ACTIVE_SET, + .len = VGIC_NR_IRQS / 8, + .handle_mmio = handle_mmio_raz_wi, + }, + { + .base = GIC_DIST_ACTIVE_CLEAR, + .len = VGIC_NR_IRQS / 8, + .handle_mmio = handle_mmio_raz_wi, + }, + { + .base = GIC_DIST_PRI, + .len = VGIC_NR_IRQS, + .handle_mmio = handle_mmio_priority_reg, + }, + { + .base = GIC_DIST_TARGET, + .len = VGIC_NR_IRQS, + .handle_mmio = handle_mmio_target_reg, + }, + { + .base = GIC_DIST_CONFIG, + .len = VGIC_NR_IRQS / 4, + .handle_mmio = handle_mmio_cfg_reg, + }, + { + .base = GIC_DIST_SOFTINT, + .len = 4, + .handle_mmio = handle_mmio_sgi_reg, + }, {} }; @@ -152,7 +647,106 @@ struct mmio_range *find_matching_range(const struct mmio_range *ranges, bool vgic_handle_mmio(struct kvm_vcpu *vcpu, struct kvm_run *run, struct kvm_exit_mmio *mmio) { - return KVM_EXIT_MMIO; + const struct mmio_range *range; + struct vgic_dist *dist = &vcpu->kvm->arch.vgic; + unsigned long base = dist->vgic_dist_base; + bool updated_state; + unsigned long offset; + + if (!irqchip_in_kernel(vcpu->kvm) || + mmio->phys_addr < base || + (mmio->phys_addr + mmio->len) > (base + KVM_VGIC_V2_DIST_SIZE)) + return false; + + /* We don't support ldrd / strd or ldm / stm to the emulated vgic */ + if (mmio->len > 4) { + kvm_inject_dabt(vcpu, mmio->phys_addr); + return true; + } + + range = find_matching_range(vgic_ranges, mmio, base); + if (unlikely(!range || !range->handle_mmio)) { + pr_warn("Unhandled access %d %08llx %d\n", + mmio->is_write, mmio->phys_addr, mmio->len); + return false; + } + + spin_lock(&vcpu->kvm->arch.vgic.lock); + offset = mmio->phys_addr - range->base - base; + updated_state = range->handle_mmio(vcpu, mmio, offset); + spin_unlock(&vcpu->kvm->arch.vgic.lock); + kvm_prepare_mmio(run, mmio); + kvm_handle_mmio_return(vcpu, run); + + return true; +} + +static void vgic_dispatch_sgi(struct kvm_vcpu *vcpu, u32 reg) +{ + struct kvm *kvm = vcpu->kvm; + struct vgic_dist *dist = &kvm->arch.vgic; + int nrcpus = atomic_read(&kvm->online_vcpus); + u8 target_cpus; + int sgi, mode, c, vcpu_id; + + vcpu_id = vcpu->vcpu_id; + + sgi = reg & 0xf; + target_cpus = (reg >> 16) & 0xff; + mode = (reg >> 24) & 3; + + switch (mode) { + case 0: + if (!target_cpus) + return; + + case 1: + target_cpus = ((1 << nrcpus) - 1) & ~(1 << vcpu_id) & 0xff; + break; + + case 2: + target_cpus = 1 << vcpu_id; + break; + } + + kvm_for_each_vcpu(c, vcpu, kvm) { + if (target_cpus & 1) { + /* Flag the SGI as pending */ + vgic_dist_irq_set(vcpu, sgi); + dist->irq_sgi_sources[c][sgi] |= 1 << vcpu_id; + kvm_debug("SGI%d from CPU%d to CPU%d\n", sgi, vcpu_id, c); + } + + target_cpus >>= 1; + } +} + +static int compute_pending_for_cpu(struct kvm_vcpu *vcpu) +{ + return 0; +} + +/* + * Update the interrupt state and determine which CPUs have pending + * interrupts. Must be called with distributor lock held. + */ +static void vgic_update_state(struct kvm *kvm) +{ + struct vgic_dist *dist = &kvm->arch.vgic; + struct kvm_vcpu *vcpu; + int c; + + if (!dist->enabled) { + set_bit(0, &dist->irq_pending_on_cpu); + return; + } + + kvm_for_each_vcpu(c, vcpu, kvm) { + if (compute_pending_for_cpu(vcpu)) { + pr_debug("CPU%d has pending interrupts\n", c); + set_bit(c, &dist->irq_pending_on_cpu); + } + } } static bool vgic_ioaddr_overlap(struct kvm *kvm)