From patchwork Fri Sep 28 09:45:34 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Paul Mackerras X-Patchwork-Id: 10619241 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 655D7180E for ; Fri, 28 Sep 2018 09:46:24 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 5888A2AB2B for ; Fri, 28 Sep 2018 09:46:24 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 48CF62AEA3; Fri, 28 Sep 2018 09:46:24 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-8.0 required=2.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,MAILING_LIST_MULTI,RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 16FAA2AB2B for ; Fri, 28 Sep 2018 09:46:23 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1729294AbeI1QJR (ORCPT ); Fri, 28 Sep 2018 12:09:17 -0400 Received: from ozlabs.org ([203.11.71.1]:59999 "EHLO ozlabs.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1729120AbeI1QJQ (ORCPT ); Fri, 28 Sep 2018 12:09:16 -0400 Received: from authenticated.ozlabs.org (localhost [127.0.0.1]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPSA id 42M6J92RfYz9sCR; Fri, 28 Sep 2018 19:46:17 +1000 (AEST) Authentication-Results: ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.org DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=ozlabs.org; s=201707; t=1538127977; bh=bYaGtk9+M+4jP9f5VRWnnhXsqWSJlegMSozLLd9/obs=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=vqugIYQ+6dgHGK+UuihaIRzLmrmNkAWLxlpgjYCi4Px+ZGxqKuRHS8B2Rcef1FGx1 pSs+Wmi5LcghySwOvLbBhSEJe8CEqMZHBd56Am2NdyeZbSOiGuhgTnr0yKTVh86bLj o4Qs3nKK68/RocPcWBmCVjgs5GktEtCTKr3CQgcpjPPJbMzvRrs+3WXqTWcuP8CdEF e0qttry6xqPzW9/h3BruNqDjneDu/6VkCRv5BTbLx3OKI35UA5JYdSB3K8XsdKgqH+ eBm9pFkW5rWZcZzNrHzwh2KD4ppLiMt7ehf+xEQTwjAogevefYamlhqIes7W7aHbTG 4L7EyHlMqIXAQ== From: Paul Mackerras To: kvm@vger.kernel.org, kvm-ppc@vger.kernel.org Cc: linuxppc-dev@ozlabs.org, David Gibson Subject: [PATCH v2 04/33] KVM: PPC: Book3S HV: Extract PMU save/restore operations as C-callable functions Date: Fri, 28 Sep 2018 19:45:34 +1000 Message-Id: <1538127963-15645-5-git-send-email-paulus@ozlabs.org> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1538127963-15645-1-git-send-email-paulus@ozlabs.org> References: <1538127963-15645-1-git-send-email-paulus@ozlabs.org> Sender: kvm-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: kvm@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP This pulls out the assembler code that is responsible for saving and restoring the PMU state for the host and guest into separate functions so they can be used from an alternate entry path. The calling convention is made compatible with C. Reviewed-by: David Gibson Signed-off-by: Paul Mackerras --- arch/powerpc/include/asm/asm-prototypes.h | 5 + arch/powerpc/kvm/book3s_hv_interrupts.S | 95 ++++---- arch/powerpc/kvm/book3s_hv_rmhandlers.S | 363 ++++++++++++++++-------------- 3 files changed, 253 insertions(+), 210 deletions(-) diff --git a/arch/powerpc/include/asm/asm-prototypes.h b/arch/powerpc/include/asm/asm-prototypes.h index 1f4691c..024e8fc 100644 --- a/arch/powerpc/include/asm/asm-prototypes.h +++ b/arch/powerpc/include/asm/asm-prototypes.h @@ -150,4 +150,9 @@ extern s32 patch__memset_nocache, patch__memcpy_nocache; extern long flush_count_cache; +void kvmhv_save_host_pmu(void); +void kvmhv_load_host_pmu(void); +void kvmhv_save_guest_pmu(struct kvm_vcpu *vcpu, bool pmu_in_use); +void kvmhv_load_guest_pmu(struct kvm_vcpu *vcpu); + #endif /* _ASM_POWERPC_ASM_PROTOTYPES_H */ diff --git a/arch/powerpc/kvm/book3s_hv_interrupts.S b/arch/powerpc/kvm/book3s_hv_interrupts.S index 666b91c..a6d1001 100644 --- a/arch/powerpc/kvm/book3s_hv_interrupts.S +++ b/arch/powerpc/kvm/book3s_hv_interrupts.S @@ -64,52 +64,7 @@ BEGIN_FTR_SECTION END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_207S) /* Save host PMU registers */ -BEGIN_FTR_SECTION - /* Work around P8 PMAE bug */ - li r3, -1 - clrrdi r3, r3, 10 - mfspr r8, SPRN_MMCR2 - mtspr SPRN_MMCR2, r3 /* freeze all counters using MMCR2 */ - isync -END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) - li r3, 1 - sldi r3, r3, 31 /* MMCR0_FC (freeze counters) bit */ - mfspr r7, SPRN_MMCR0 /* save MMCR0 */ - mtspr SPRN_MMCR0, r3 /* freeze all counters, disable interrupts */ - mfspr r6, SPRN_MMCRA - /* Clear MMCRA in order to disable SDAR updates */ - li r5, 0 - mtspr SPRN_MMCRA, r5 - isync - lbz r5, PACA_PMCINUSE(r13) /* is the host using the PMU? */ - cmpwi r5, 0 - beq 31f /* skip if not */ - mfspr r5, SPRN_MMCR1 - mfspr r9, SPRN_SIAR - mfspr r10, SPRN_SDAR - std r7, HSTATE_MMCR0(r13) - std r5, HSTATE_MMCR1(r13) - std r6, HSTATE_MMCRA(r13) - std r9, HSTATE_SIAR(r13) - std r10, HSTATE_SDAR(r13) -BEGIN_FTR_SECTION - mfspr r9, SPRN_SIER - std r8, HSTATE_MMCR2(r13) - std r9, HSTATE_SIER(r13) -END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) - mfspr r3, SPRN_PMC1 - mfspr r5, SPRN_PMC2 - mfspr r6, SPRN_PMC3 - mfspr r7, SPRN_PMC4 - mfspr r8, SPRN_PMC5 - mfspr r9, SPRN_PMC6 - stw r3, HSTATE_PMC1(r13) - stw r5, HSTATE_PMC2(r13) - stw r6, HSTATE_PMC3(r13) - stw r7, HSTATE_PMC4(r13) - stw r8, HSTATE_PMC5(r13) - stw r9, HSTATE_PMC6(r13) -31: + bl kvmhv_save_host_pmu /* * Put whatever is in the decrementer into the @@ -161,3 +116,51 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300) ld r0, PPC_LR_STKOFF(r1) mtlr r0 blr + +_GLOBAL(kvmhv_save_host_pmu) +BEGIN_FTR_SECTION + /* Work around P8 PMAE bug */ + li r3, -1 + clrrdi r3, r3, 10 + mfspr r8, SPRN_MMCR2 + mtspr SPRN_MMCR2, r3 /* freeze all counters using MMCR2 */ + isync +END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) + li r3, 1 + sldi r3, r3, 31 /* MMCR0_FC (freeze counters) bit */ + mfspr r7, SPRN_MMCR0 /* save MMCR0 */ + mtspr SPRN_MMCR0, r3 /* freeze all counters, disable interrupts */ + mfspr r6, SPRN_MMCRA + /* Clear MMCRA in order to disable SDAR updates */ + li r5, 0 + mtspr SPRN_MMCRA, r5 + isync + lbz r5, PACA_PMCINUSE(r13) /* is the host using the PMU? */ + cmpwi r5, 0 + beq 31f /* skip if not */ + mfspr r5, SPRN_MMCR1 + mfspr r9, SPRN_SIAR + mfspr r10, SPRN_SDAR + std r7, HSTATE_MMCR0(r13) + std r5, HSTATE_MMCR1(r13) + std r6, HSTATE_MMCRA(r13) + std r9, HSTATE_SIAR(r13) + std r10, HSTATE_SDAR(r13) +BEGIN_FTR_SECTION + mfspr r9, SPRN_SIER + std r8, HSTATE_MMCR2(r13) + std r9, HSTATE_SIER(r13) +END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) + mfspr r3, SPRN_PMC1 + mfspr r5, SPRN_PMC2 + mfspr r6, SPRN_PMC3 + mfspr r7, SPRN_PMC4 + mfspr r8, SPRN_PMC5 + mfspr r9, SPRN_PMC6 + stw r3, HSTATE_PMC1(r13) + stw r5, HSTATE_PMC2(r13) + stw r6, HSTATE_PMC3(r13) + stw r7, HSTATE_PMC4(r13) + stw r8, HSTATE_PMC5(r13) + stw r9, HSTATE_PMC6(r13) +31: blr diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S index 6752da1..5b2ae34 100644 --- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S +++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S @@ -28,6 +28,7 @@ #include #include #include +#include #include #include #include @@ -113,45 +114,7 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_207S) mtspr SPRN_SPRG_VDSO_WRITE,r3 /* Reload the host's PMU registers */ - lbz r4, PACA_PMCINUSE(r13) /* is the host using the PMU? */ - cmpwi r4, 0 - beq 23f /* skip if not */ -BEGIN_FTR_SECTION - ld r3, HSTATE_MMCR0(r13) - andi. r4, r3, MMCR0_PMAO_SYNC | MMCR0_PMAO - cmpwi r4, MMCR0_PMAO - beql kvmppc_fix_pmao -END_FTR_SECTION_IFSET(CPU_FTR_PMAO_BUG) - lwz r3, HSTATE_PMC1(r13) - lwz r4, HSTATE_PMC2(r13) - lwz r5, HSTATE_PMC3(r13) - lwz r6, HSTATE_PMC4(r13) - lwz r8, HSTATE_PMC5(r13) - lwz r9, HSTATE_PMC6(r13) - mtspr SPRN_PMC1, r3 - mtspr SPRN_PMC2, r4 - mtspr SPRN_PMC3, r5 - mtspr SPRN_PMC4, r6 - mtspr SPRN_PMC5, r8 - mtspr SPRN_PMC6, r9 - ld r3, HSTATE_MMCR0(r13) - ld r4, HSTATE_MMCR1(r13) - ld r5, HSTATE_MMCRA(r13) - ld r6, HSTATE_SIAR(r13) - ld r7, HSTATE_SDAR(r13) - mtspr SPRN_MMCR1, r4 - mtspr SPRN_MMCRA, r5 - mtspr SPRN_SIAR, r6 - mtspr SPRN_SDAR, r7 -BEGIN_FTR_SECTION - ld r8, HSTATE_MMCR2(r13) - ld r9, HSTATE_SIER(r13) - mtspr SPRN_MMCR2, r8 - mtspr SPRN_SIER, r9 -END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) - mtspr SPRN_MMCR0, r3 - isync -23: + bl kvmhv_load_host_pmu /* * Reload DEC. HDEC interrupts were disabled when @@ -805,57 +768,12 @@ END_FTR_SECTION(CPU_FTR_TM | CPU_FTR_P9_TM_HV_ASSIST, 0) 91: #endif - /* Load guest PMU registers */ - /* R4 is live here (vcpu pointer) */ - li r3, 1 - sldi r3, r3, 31 /* MMCR0_FC (freeze counters) bit */ - mtspr SPRN_MMCR0, r3 /* freeze all counters, disable ints */ - isync -BEGIN_FTR_SECTION - ld r3, VCPU_MMCR(r4) - andi. r5, r3, MMCR0_PMAO_SYNC | MMCR0_PMAO - cmpwi r5, MMCR0_PMAO - beql kvmppc_fix_pmao -END_FTR_SECTION_IFSET(CPU_FTR_PMAO_BUG) - lwz r3, VCPU_PMC(r4) /* always load up guest PMU registers */ - lwz r5, VCPU_PMC + 4(r4) /* to prevent information leak */ - lwz r6, VCPU_PMC + 8(r4) - lwz r7, VCPU_PMC + 12(r4) - lwz r8, VCPU_PMC + 16(r4) - lwz r9, VCPU_PMC + 20(r4) - mtspr SPRN_PMC1, r3 - mtspr SPRN_PMC2, r5 - mtspr SPRN_PMC3, r6 - mtspr SPRN_PMC4, r7 - mtspr SPRN_PMC5, r8 - mtspr SPRN_PMC6, r9 - ld r3, VCPU_MMCR(r4) - ld r5, VCPU_MMCR + 8(r4) - ld r6, VCPU_MMCR + 16(r4) - ld r7, VCPU_SIAR(r4) - ld r8, VCPU_SDAR(r4) - mtspr SPRN_MMCR1, r5 - mtspr SPRN_MMCRA, r6 - mtspr SPRN_SIAR, r7 - mtspr SPRN_SDAR, r8 -BEGIN_FTR_SECTION - ld r5, VCPU_MMCR + 24(r4) - ld r6, VCPU_SIER(r4) - mtspr SPRN_MMCR2, r5 - mtspr SPRN_SIER, r6 -BEGIN_FTR_SECTION_NESTED(96) - lwz r7, VCPU_PMC + 24(r4) - lwz r8, VCPU_PMC + 28(r4) - ld r9, VCPU_MMCR + 32(r4) - mtspr SPRN_SPMC1, r7 - mtspr SPRN_SPMC2, r8 - mtspr SPRN_MMCRS, r9 -END_FTR_SECTION_NESTED(CPU_FTR_ARCH_300, 0, 96) -END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) - mtspr SPRN_MMCR0, r3 - isync + /* Load guest PMU registers; r4 = vcpu pointer here */ + mr r3, r4 + bl kvmhv_load_guest_pmu /* Load up FP, VMX and VSX registers */ + ld r4, HSTATE_KVM_VCPU(r13) bl kvmppc_load_fp ld r14, VCPU_GPR(R14)(r4) @@ -1766,83 +1684,12 @@ END_FTR_SECTION(CPU_FTR_TM | CPU_FTR_P9_TM_HV_ASSIST, 0) 25: /* Save PMU registers if requested */ /* r8 and cr0.eq are live here */ -BEGIN_FTR_SECTION - /* - * POWER8 seems to have a hardware bug where setting - * MMCR0[PMAE] along with MMCR0[PMC1CE] and/or MMCR0[PMCjCE] - * when some counters are already negative doesn't seem - * to cause a performance monitor alert (and hence interrupt). - * The effect of this is that when saving the PMU state, - * if there is no PMU alert pending when we read MMCR0 - * before freezing the counters, but one becomes pending - * before we read the counters, we lose it. - * To work around this, we need a way to freeze the counters - * before reading MMCR0. Normally, freezing the counters - * is done by writing MMCR0 (to set MMCR0[FC]) which - * unavoidably writes MMCR0[PMA0] as well. On POWER8, - * we can also freeze the counters using MMCR2, by writing - * 1s to all the counter freeze condition bits (there are - * 9 bits each for 6 counters). - */ - li r3, -1 /* set all freeze bits */ - clrrdi r3, r3, 10 - mfspr r10, SPRN_MMCR2 - mtspr SPRN_MMCR2, r3 - isync -END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) - li r3, 1 - sldi r3, r3, 31 /* MMCR0_FC (freeze counters) bit */ - mfspr r4, SPRN_MMCR0 /* save MMCR0 */ - mtspr SPRN_MMCR0, r3 /* freeze all counters, disable ints */ - mfspr r6, SPRN_MMCRA - /* Clear MMCRA in order to disable SDAR updates */ - li r7, 0 - mtspr SPRN_MMCRA, r7 - isync + mr r3, r9 + li r4, 1 beq 21f /* if no VPA, save PMU stuff anyway */ - lbz r7, LPPACA_PMCINUSE(r8) - cmpwi r7, 0 /* did they ask for PMU stuff to be saved? */ - bne 21f - std r3, VCPU_MMCR(r9) /* if not, set saved MMCR0 to FC */ - b 22f -21: mfspr r5, SPRN_MMCR1 - mfspr r7, SPRN_SIAR - mfspr r8, SPRN_SDAR - std r4, VCPU_MMCR(r9) - std r5, VCPU_MMCR + 8(r9) - std r6, VCPU_MMCR + 16(r9) -BEGIN_FTR_SECTION - std r10, VCPU_MMCR + 24(r9) -END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) - std r7, VCPU_SIAR(r9) - std r8, VCPU_SDAR(r9) - mfspr r3, SPRN_PMC1 - mfspr r4, SPRN_PMC2 - mfspr r5, SPRN_PMC3 - mfspr r6, SPRN_PMC4 - mfspr r7, SPRN_PMC5 - mfspr r8, SPRN_PMC6 - stw r3, VCPU_PMC(r9) - stw r4, VCPU_PMC + 4(r9) - stw r5, VCPU_PMC + 8(r9) - stw r6, VCPU_PMC + 12(r9) - stw r7, VCPU_PMC + 16(r9) - stw r8, VCPU_PMC + 20(r9) -BEGIN_FTR_SECTION - mfspr r5, SPRN_SIER - std r5, VCPU_SIER(r9) -BEGIN_FTR_SECTION_NESTED(96) - mfspr r6, SPRN_SPMC1 - mfspr r7, SPRN_SPMC2 - mfspr r8, SPRN_MMCRS - stw r6, VCPU_PMC + 24(r9) - stw r7, VCPU_PMC + 28(r9) - std r8, VCPU_MMCR + 32(r9) - lis r4, 0x8000 - mtspr SPRN_MMCRS, r4 -END_FTR_SECTION_NESTED(CPU_FTR_ARCH_300, 0, 96) -END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) -22: + lbz r4, LPPACA_PMCINUSE(r8) +21: bl kvmhv_save_guest_pmu + ld r9, HSTATE_KVM_VCPU(r13) /* Restore host values of some registers */ BEGIN_FTR_SECTION @@ -3388,6 +3235,194 @@ kvmppc_msr_interrupt: blr /* + * Load up guest PMU state. R3 points to the vcpu struct. + */ +_GLOBAL(kvmhv_load_guest_pmu) +EXPORT_SYMBOL_GPL(kvmhv_load_guest_pmu) + mr r4, r3 + mflr r0 + li r3, 1 + sldi r3, r3, 31 /* MMCR0_FC (freeze counters) bit */ + mtspr SPRN_MMCR0, r3 /* freeze all counters, disable ints */ + isync +BEGIN_FTR_SECTION + ld r3, VCPU_MMCR(r4) + andi. r5, r3, MMCR0_PMAO_SYNC | MMCR0_PMAO + cmpwi r5, MMCR0_PMAO + beql kvmppc_fix_pmao +END_FTR_SECTION_IFSET(CPU_FTR_PMAO_BUG) + lwz r3, VCPU_PMC(r4) /* always load up guest PMU registers */ + lwz r5, VCPU_PMC + 4(r4) /* to prevent information leak */ + lwz r6, VCPU_PMC + 8(r4) + lwz r7, VCPU_PMC + 12(r4) + lwz r8, VCPU_PMC + 16(r4) + lwz r9, VCPU_PMC + 20(r4) + mtspr SPRN_PMC1, r3 + mtspr SPRN_PMC2, r5 + mtspr SPRN_PMC3, r6 + mtspr SPRN_PMC4, r7 + mtspr SPRN_PMC5, r8 + mtspr SPRN_PMC6, r9 + ld r3, VCPU_MMCR(r4) + ld r5, VCPU_MMCR + 8(r4) + ld r6, VCPU_MMCR + 16(r4) + ld r7, VCPU_SIAR(r4) + ld r8, VCPU_SDAR(r4) + mtspr SPRN_MMCR1, r5 + mtspr SPRN_MMCRA, r6 + mtspr SPRN_SIAR, r7 + mtspr SPRN_SDAR, r8 +BEGIN_FTR_SECTION + ld r5, VCPU_MMCR + 24(r4) + ld r6, VCPU_SIER(r4) + mtspr SPRN_MMCR2, r5 + mtspr SPRN_SIER, r6 +BEGIN_FTR_SECTION_NESTED(96) + lwz r7, VCPU_PMC + 24(r4) + lwz r8, VCPU_PMC + 28(r4) + ld r9, VCPU_MMCR + 32(r4) + mtspr SPRN_SPMC1, r7 + mtspr SPRN_SPMC2, r8 + mtspr SPRN_MMCRS, r9 +END_FTR_SECTION_NESTED(CPU_FTR_ARCH_300, 0, 96) +END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) + mtspr SPRN_MMCR0, r3 + isync + mtlr r0 + blr + +/* + * Reload host PMU state saved in the PACA by kvmhv_save_host_pmu. + */ +_GLOBAL(kvmhv_load_host_pmu) +EXPORT_SYMBOL_GPL(kvmhv_load_host_pmu) + mflr r0 + lbz r4, PACA_PMCINUSE(r13) /* is the host using the PMU? */ + cmpwi r4, 0 + beq 23f /* skip if not */ +BEGIN_FTR_SECTION + ld r3, HSTATE_MMCR0(r13) + andi. r4, r3, MMCR0_PMAO_SYNC | MMCR0_PMAO + cmpwi r4, MMCR0_PMAO + beql kvmppc_fix_pmao +END_FTR_SECTION_IFSET(CPU_FTR_PMAO_BUG) + lwz r3, HSTATE_PMC1(r13) + lwz r4, HSTATE_PMC2(r13) + lwz r5, HSTATE_PMC3(r13) + lwz r6, HSTATE_PMC4(r13) + lwz r8, HSTATE_PMC5(r13) + lwz r9, HSTATE_PMC6(r13) + mtspr SPRN_PMC1, r3 + mtspr SPRN_PMC2, r4 + mtspr SPRN_PMC3, r5 + mtspr SPRN_PMC4, r6 + mtspr SPRN_PMC5, r8 + mtspr SPRN_PMC6, r9 + ld r3, HSTATE_MMCR0(r13) + ld r4, HSTATE_MMCR1(r13) + ld r5, HSTATE_MMCRA(r13) + ld r6, HSTATE_SIAR(r13) + ld r7, HSTATE_SDAR(r13) + mtspr SPRN_MMCR1, r4 + mtspr SPRN_MMCRA, r5 + mtspr SPRN_SIAR, r6 + mtspr SPRN_SDAR, r7 +BEGIN_FTR_SECTION + ld r8, HSTATE_MMCR2(r13) + ld r9, HSTATE_SIER(r13) + mtspr SPRN_MMCR2, r8 + mtspr SPRN_SIER, r9 +END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) + mtspr SPRN_MMCR0, r3 + isync + mtlr r0 +23: blr + +/* + * Save guest PMU state into the vcpu struct. + * r3 = vcpu, r4 = full save flag (PMU in use flag set in VPA) + */ +_GLOBAL(kvmhv_save_guest_pmu) +EXPORT_SYMBOL_GPL(kvmhv_save_guest_pmu) + mr r9, r3 + mr r8, r4 +BEGIN_FTR_SECTION + /* + * POWER8 seems to have a hardware bug where setting + * MMCR0[PMAE] along with MMCR0[PMC1CE] and/or MMCR0[PMCjCE] + * when some counters are already negative doesn't seem + * to cause a performance monitor alert (and hence interrupt). + * The effect of this is that when saving the PMU state, + * if there is no PMU alert pending when we read MMCR0 + * before freezing the counters, but one becomes pending + * before we read the counters, we lose it. + * To work around this, we need a way to freeze the counters + * before reading MMCR0. Normally, freezing the counters + * is done by writing MMCR0 (to set MMCR0[FC]) which + * unavoidably writes MMCR0[PMA0] as well. On POWER8, + * we can also freeze the counters using MMCR2, by writing + * 1s to all the counter freeze condition bits (there are + * 9 bits each for 6 counters). + */ + li r3, -1 /* set all freeze bits */ + clrrdi r3, r3, 10 + mfspr r10, SPRN_MMCR2 + mtspr SPRN_MMCR2, r3 + isync +END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) + li r3, 1 + sldi r3, r3, 31 /* MMCR0_FC (freeze counters) bit */ + mfspr r4, SPRN_MMCR0 /* save MMCR0 */ + mtspr SPRN_MMCR0, r3 /* freeze all counters, disable ints */ + mfspr r6, SPRN_MMCRA + /* Clear MMCRA in order to disable SDAR updates */ + li r7, 0 + mtspr SPRN_MMCRA, r7 + isync + cmpwi r8, 0 /* did they ask for PMU stuff to be saved? */ + bne 21f + std r3, VCPU_MMCR(r9) /* if not, set saved MMCR0 to FC */ + b 22f +21: mfspr r5, SPRN_MMCR1 + mfspr r7, SPRN_SIAR + mfspr r8, SPRN_SDAR + std r4, VCPU_MMCR(r9) + std r5, VCPU_MMCR + 8(r9) + std r6, VCPU_MMCR + 16(r9) +BEGIN_FTR_SECTION + std r10, VCPU_MMCR + 24(r9) +END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) + std r7, VCPU_SIAR(r9) + std r8, VCPU_SDAR(r9) + mfspr r3, SPRN_PMC1 + mfspr r4, SPRN_PMC2 + mfspr r5, SPRN_PMC3 + mfspr r6, SPRN_PMC4 + mfspr r7, SPRN_PMC5 + mfspr r8, SPRN_PMC6 + stw r3, VCPU_PMC(r9) + stw r4, VCPU_PMC + 4(r9) + stw r5, VCPU_PMC + 8(r9) + stw r6, VCPU_PMC + 12(r9) + stw r7, VCPU_PMC + 16(r9) + stw r8, VCPU_PMC + 20(r9) +BEGIN_FTR_SECTION + mfspr r5, SPRN_SIER + std r5, VCPU_SIER(r9) +BEGIN_FTR_SECTION_NESTED(96) + mfspr r6, SPRN_SPMC1 + mfspr r7, SPRN_SPMC2 + mfspr r8, SPRN_MMCRS + stw r6, VCPU_PMC + 24(r9) + stw r7, VCPU_PMC + 28(r9) + std r8, VCPU_MMCR + 32(r9) + lis r4, 0x8000 + mtspr SPRN_MMCRS, r4 +END_FTR_SECTION_NESTED(CPU_FTR_ARCH_300, 0, 96) +END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S) +22: blr + +/* * This works around a hardware bug on POWER8E processors, where * writing a 1 to the MMCR0[PMAO] bit doesn't generate a * performance monitor interrupt. Instead, when we need to have