From owner-svn-src-projects@FreeBSD.ORG Wed Dec 18 23:39:43 2013 Return-Path: Delivered-To: svn-src-projects@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [8.8.178.115]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id CD131D01; Wed, 18 Dec 2013 23:39:43 +0000 (UTC) Received: from svn.freebsd.org (svn.freebsd.org [IPv6:2001:1900:2254:2068::e6a:0]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mx1.freebsd.org (Postfix) with ESMTPS id AD32B127D; Wed, 18 Dec 2013 23:39:43 +0000 (UTC) Received: from svn.freebsd.org ([127.0.1.70]) by svn.freebsd.org (8.14.7/8.14.7) with ESMTP id rBINdh8r027040; Wed, 18 Dec 2013 23:39:43 GMT (envelope-from grehan@svn.freebsd.org) Received: (from grehan@localhost) by svn.freebsd.org (8.14.7/8.14.7/Submit) id rBINdg0w027032; Wed, 18 Dec 2013 23:39:42 GMT (envelope-from grehan@svn.freebsd.org) Message-Id: <201312182339.rBINdg0w027032@svn.freebsd.org> From: Peter Grehan Date: Wed, 18 Dec 2013 23:39:42 +0000 (UTC) To: src-committers@freebsd.org, svn-src-projects@freebsd.org Subject: svn commit: r259579 - in projects/bhyve_svm/sys/amd64: amd64 vmm vmm/amd X-SVN-Group: projects MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-BeenThere: svn-src-projects@freebsd.org X-Mailman-Version: 2.1.17 Precedence: list List-Id: "SVN commit messages for the src " projects" tree" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 18 Dec 2013 23:39:43 -0000 Author: grehan Date: Wed Dec 18 23:39:42 2013 New Revision: 259579 URL: http://svnweb.freebsd.org/changeset/base/259579 Log: Enable memory overcommit for AMD processors. - No emulation of A/D bits is required since AMD-V RVI supports A/D bits. - Enable pmap PT_RVI support(w/o PAT) which is required for memory over-commit support. - Other minor fixes: * Make use of VMCB EXITINTINFO field. If a #VMEXIT happens while delivering an interrupt, EXITINTINFO has all the details that bhyve needs to inject the same interrupt. * SVM h/w decode assist code was incomplete - removed for now. * Some minor code clean-up (more coming). Submitted by: Anish Gupta (akgupt3@gmail.com) Modified: projects/bhyve_svm/sys/amd64/amd64/pmap.c projects/bhyve_svm/sys/amd64/vmm/amd/amdv.c projects/bhyve_svm/sys/amd64/vmm/amd/npt.c projects/bhyve_svm/sys/amd64/vmm/amd/npt.h projects/bhyve_svm/sys/amd64/vmm/amd/svm.c projects/bhyve_svm/sys/amd64/vmm/amd/svm_softc.h projects/bhyve_svm/sys/amd64/vmm/amd/vmcb.c projects/bhyve_svm/sys/amd64/vmm/amd/vmcb.h projects/bhyve_svm/sys/amd64/vmm/vmm.c projects/bhyve_svm/sys/amd64/vmm/vmm_instruction_emul.c projects/bhyve_svm/sys/amd64/vmm/vmm_msr.c Modified: projects/bhyve_svm/sys/amd64/amd64/pmap.c ============================================================================== --- projects/bhyve_svm/sys/amd64/amd64/pmap.c Wed Dec 18 22:31:53 2013 (r259578) +++ projects/bhyve_svm/sys/amd64/amd64/pmap.c Wed Dec 18 23:39:42 2013 (r259579) @@ -146,6 +146,13 @@ __FBSDID("$FreeBSD$"); #endif static __inline boolean_t +pmap_type_guest(pmap_t pmap) +{ + + return ((pmap->pm_type == PT_EPT) || (pmap->pm_type == PT_RVI)); +} + +static __inline boolean_t pmap_emulate_ad_bits(pmap_t pmap) { @@ -159,6 +166,7 @@ pmap_valid_bit(pmap_t pmap) switch (pmap->pm_type) { case PT_X86: + case PT_RVI: mask = X86_PG_V; break; case PT_EPT: @@ -181,6 +189,7 @@ pmap_rw_bit(pmap_t pmap) switch (pmap->pm_type) { case PT_X86: + case PT_RVI: mask = X86_PG_RW; break; case PT_EPT: @@ -205,6 +214,7 @@ pmap_global_bit(pmap_t pmap) case PT_X86: mask = X86_PG_G; break; + case PT_RVI: case PT_EPT: mask = 0; break; @@ -222,6 +232,7 @@ pmap_accessed_bit(pmap_t pmap) switch (pmap->pm_type) { case PT_X86: + case PT_RVI: mask = X86_PG_A; break; case PT_EPT: @@ -244,6 +255,7 @@ pmap_modified_bit(pmap_t pmap) switch (pmap->pm_type) { case PT_X86: + case PT_RVI: mask = X86_PG_M; break; case PT_EPT: @@ -1094,6 +1106,9 @@ pmap_swap_pat(pmap_t pmap, pt_entry_t en if ((entry & x86_pat_bits) != 0) entry ^= x86_pat_bits; break; + case PT_RVI: + /* XXX: PAT support. */ + break; case PT_EPT: /* * Nothing to do - the memory attributes are represented @@ -1137,6 +1152,11 @@ pmap_cache_bits(pmap_t pmap, int mode, b cache_bits |= PG_NC_PWT; break; + case PT_RVI: + /* XXX: PAT support. */ + cache_bits = 0; + break; + case PT_EPT: cache_bits = EPT_PG_IGNORE_PAT | EPT_PG_MEMORY_TYPE(mode); break; @@ -1157,6 +1177,10 @@ pmap_cache_mask(pmap_t pmap, boolean_t i case PT_X86: mask = is_pde ? X86_PG_PDE_CACHE : X86_PG_PTE_CACHE; break; + case PT_RVI: + /* XXX: PAT support. */ + mask = 0; + break; case PT_EPT: mask = EPT_PG_IGNORE_PAT | EPT_PG_MEMORY_TYPE(0x7); break; @@ -1181,6 +1205,7 @@ pmap_update_pde_store(pmap_t pmap, pd_en switch (pmap->pm_type) { case PT_X86: break; + case PT_RVI: case PT_EPT: /* * XXX @@ -1216,9 +1241,9 @@ pmap_update_pde_invalidate(pmap_t pmap, { pt_entry_t PG_G; - if (pmap->pm_type == PT_EPT) + if (pmap_type_guest(pmap)) return; - + KASSERT(pmap->pm_type == PT_X86, ("pmap_update_pde_invalidate: invalid type %d", pmap->pm_type)); @@ -1331,11 +1356,11 @@ pmap_invalidate_page(pmap_t pmap, vm_off cpuset_t other_cpus; u_int cpuid; - if (pmap->pm_type == PT_EPT) { + if (pmap_type_guest(pmap)) { pmap_invalidate_ept(pmap); return; } - + KASSERT(pmap->pm_type == PT_X86, ("pmap_invalidate_page: invalid type %d", pmap->pm_type)); @@ -1409,7 +1434,7 @@ pmap_invalidate_range(pmap_t pmap, vm_of vm_offset_t addr; u_int cpuid; - if (pmap->pm_type == PT_EPT) { + if (pmap_type_guest(pmap)) { pmap_invalidate_ept(pmap); return; } @@ -1468,7 +1493,7 @@ pmap_invalidate_all(pmap_t pmap) uint64_t cr3; u_int cpuid; - if (pmap->pm_type == PT_EPT) { + if (pmap_type_guest(pmap)) { pmap_invalidate_ept(pmap); return; } @@ -1588,7 +1613,7 @@ pmap_update_pde(pmap_t pmap, vm_offset_t cpuid = PCPU_GET(cpuid); other_cpus = all_cpus; CPU_CLR(cpuid, &other_cpus); - if (pmap == kernel_pmap || pmap->pm_type == PT_EPT) + if (pmap == kernel_pmap || pmap_type_guest(pmap)) active = all_cpus; else { active = pmap->pm_active; @@ -1626,6 +1651,7 @@ pmap_invalidate_page(pmap_t pmap, vm_off if (pmap == kernel_pmap || !CPU_EMPTY(&pmap->pm_active)) invlpg(va); break; + case PT_RVI: case PT_EPT: pmap->pm_eptgen++; break; @@ -1645,6 +1671,7 @@ pmap_invalidate_range(pmap_t pmap, vm_of for (addr = sva; addr < eva; addr += PAGE_SIZE) invlpg(addr); break; + case PT_RVI: case PT_EPT: pmap->pm_eptgen++; break; @@ -1662,6 +1689,7 @@ pmap_invalidate_all(pmap_t pmap) if (pmap == kernel_pmap || !CPU_EMPTY(&pmap->pm_active)) invltlb(); break; + case PT_RVI: case PT_EPT: pmap->pm_eptgen++; break; Modified: projects/bhyve_svm/sys/amd64/vmm/amd/amdv.c ============================================================================== --- projects/bhyve_svm/sys/amd64/vmm/amd/amdv.c Wed Dec 18 22:31:53 2013 (r259578) +++ projects/bhyve_svm/sys/amd64/vmm/amd/amdv.c Wed Dec 18 23:39:42 2013 (r259579) @@ -38,136 +38,6 @@ __FBSDID("$FreeBSD$"); #include "io/iommu.h" static int -amdv_init(void) -{ - - printf("amdv_init: not implemented\n"); - return (ENXIO); -} - -static int -amdv_cleanup(void) -{ - - printf("amdv_cleanup: not implemented\n"); - return (ENXIO); -} - -static void * -amdv_vminit(struct vm *vm, struct pmap *pmap) -{ - - printf("amdv_vminit: not implemented\n"); - return (NULL); -} - -static int -amdv_vmrun(void *arg, int vcpu, register_t rip, struct pmap *pmap) -{ - - printf("amdv_vmrun: not implemented\n"); - return (ENXIO); -} - -static void -amdv_vmcleanup(void *arg) -{ - - printf("amdv_vmcleanup: not implemented\n"); - return; -} - -static int -amdv_getreg(void *arg, int vcpu, int regnum, uint64_t *retval) -{ - - printf("amdv_getreg: not implemented\n"); - return (EINVAL); -} - -static int -amdv_setreg(void *arg, int vcpu, int regnum, uint64_t val) -{ - - printf("amdv_setreg: not implemented\n"); - return (EINVAL); -} - -static int -amdv_getdesc(void *vmi, int vcpu, int num, struct seg_desc *desc) -{ - - printf("amdv_get_desc: not implemented\n"); - return (EINVAL); -} - -static int -amdv_setdesc(void *vmi, int vcpu, int num, struct seg_desc *desc) -{ - - printf("amdv_get_desc: not implemented\n"); - return (EINVAL); -} - -static int -amdv_inject_event(void *vmi, int vcpu, int type, int vector, - uint32_t error_code, int error_code_valid) -{ - - printf("amdv_inject_event: not implemented\n"); - return (EINVAL); -} - -static int -amdv_getcap(void *arg, int vcpu, int type, int *retval) -{ - - printf("amdv_getcap: not implemented\n"); - return (EINVAL); -} - -static int -amdv_setcap(void *arg, int vcpu, int type, int val) -{ - - printf("amdv_setcap: not implemented\n"); - return (EINVAL); -} - -static struct vmspace * -amdv_vmspace_alloc(vm_offset_t min, vm_offset_t max) -{ - - printf("amdv_vmspace_alloc: not implemented\n"); - return (NULL); -} - -static void -amdv_vmspace_free(struct vmspace *vmspace) -{ - - printf("amdv_vmspace_free: not implemented\n"); - return; -} - -struct vmm_ops vmm_ops_amd = { - amdv_init, - amdv_cleanup, - amdv_vminit, - amdv_vmrun, - amdv_vmcleanup, - amdv_getreg, - amdv_setreg, - amdv_getdesc, - amdv_setdesc, - amdv_inject_event, - amdv_getcap, - amdv_setcap, - amdv_vmspace_alloc, - amdv_vmspace_free, -}; - -static int amd_iommu_init(void) { Modified: projects/bhyve_svm/sys/amd64/vmm/amd/npt.c ============================================================================== --- projects/bhyve_svm/sys/amd64/vmm/amd/npt.c Wed Dec 18 22:31:53 2013 (r259578) +++ projects/bhyve_svm/sys/amd64/vmm/amd/npt.c Wed Dec 18 23:39:42 2013 (r259579) @@ -28,11 +28,13 @@ __FBSDID("$FreeBSD$"); #include +#include #include -#include +#include #include #include +#include #include #include @@ -44,268 +46,46 @@ __FBSDID("$FreeBSD$"); #include "svm_softc.h" #include "npt.h" -/* - * "Nested Paging" is an optional SVM feature that provides two levels of - * address translation, thus eliminating the need for the VMM to maintain - * shadow page tables. - * - * Documented in APMv2, section 15.25, Nested Paging. - */ - -#define PAGE_4KB (4 * 1024) -#define PAGE_2MB (2 * 1024 * 1024UL) -#define PAGE_1GB (1024 * 1024 * 1024UL) - -#define GPA_UNMAPPED ((vm_paddr_t)~0) - -/* Get page entry to physical address. */ -#define PTE2PA(x) ((uint64_t)(x) & ~PAGE_MASK) - -MALLOC_DECLARE(M_SVM); - -static uint64_t svm_npt_create(pml4_entry_t *pml4, vm_paddr_t gpa, - vm_paddr_t hpa, vm_memattr_t attr, - int prot, uint64_t size); - -static const int PT_INDEX_MASK = 0x1FF; -static const int PT_SHIFT = 9; - -/* - * Helper function to create nested page table entries for a page - * of size 1GB, 2MB or 4KB. - * - * Starting from PML4 create a PDPTE, PDE or PTE depending on 'pg_size' - * value of 1GB, 2MB or 4KB respectively. - * - * Return size of the mapping created on success and 0 on failure. - * - * XXX: NPT PAT settings. - */ -static uint64_t -svm_npt_create(pml4_entry_t * pml4, vm_paddr_t gpa, vm_paddr_t hpa, - vm_memattr_t attr, int prot, uint64_t pg_size) -{ - uint64_t *pt, *page, pa; - pt_entry_t mode; - int shift, index; - - KASSERT(pg_size, ("Size of page must be 1GB, 2MB or 4KB")); - if (hpa & (pg_size - 1)) { - ERR("HPA(0x%lx) is not aligned, size:0x%lx\n", hpa, pg_size); - return (0); - } - - if (gpa & (pg_size - 1)) { - ERR("GPA(0x%lx) is not aligned, size (0x%lx)\n", gpa, pg_size); - return (0); - } - - /* Find out mode bits for PTE */ - mode = PG_U | PG_V; - if (prot & VM_PROT_WRITE) - mode |= PG_RW; - if ((prot & VM_PROT_EXECUTE) == 0) - mode |= pg_nx; - - pt = (uint64_t *)pml4; - shift = PML4SHIFT; - - while ((shift > PAGE_SHIFT) && (pg_size < (1UL << shift))) { - /* Get PDP, PD or PT index from guest physical address. */ - index = (gpa >> shift) & PT_INDEX_MASK; - - /* If page entry is missing, allocate new page for table.*/ - if (pt[index] == 0) { - page = malloc(PAGE_SIZE, M_SVM, M_WAITOK | M_ZERO); - pt[index] = vtophys(page) | mode; - } - - pa = PTE2PA(pt[index]);; - pt = (uint64_t *)PHYS_TO_DMAP(pa); - shift -= PT_SHIFT; - } - - /* Create leaf entry mapping. */ - index = (gpa >> shift) & PT_INDEX_MASK; - - if (prot != VM_PROT_NONE) { - pt[index] = hpa | mode; - pt[index] |= (pg_size > PAGE_SIZE) ? PG_PS : 0; - } else - pt[index] = 0; - - return (1UL << shift); -} - -/* - * Map guest physical address to host physical address. - */ -int -svm_npt_vmmap_set(void *arg, vm_paddr_t gpa, vm_paddr_t hpa, - size_t size, vm_memattr_t attr, int prot, boolean_t spok) -{ - pml4_entry_t *pml4; - struct svm_softc *svm_sc; - uint64_t len, mapped, pg_size; - - svm_sc = arg; - pml4 = svm_sc->np_pml4; - - pg_size = PAGE_4KB; - if (spok) { - pg_size = PAGE_2MB; - if (amd_feature & AMDID_PAGE1GB) - pg_size = PAGE_1GB; - } - - /* Compute the largest page mapping that can be used */ - while (pg_size > PAGE_4KB) { - if (size >= pg_size && - (gpa & (pg_size - 1)) == 0 && - (hpa & (pg_size - 1)) == 0) { - break; - } - pg_size >>= PT_SHIFT; - } - - len = 0; - while (len < size) { - mapped = svm_npt_create(pml4, gpa + len, hpa + len, attr, prot, - pg_size); - len += mapped; - } - - return (0); -} - -/* - * Get HPA for a given GPA. - */ -vm_paddr_t -svm_npt_vmmap_get(void *arg, vm_paddr_t gpa) -{ - struct svm_softc *svm_sc; - pml4_entry_t *pml4; - uint64_t *pt, pa, hpa, pgmask; - int shift, index; - - svm_sc = arg; - pml4 = svm_sc->np_pml4; - - pt = (uint64_t *)pml4; - shift = PML4SHIFT; - - while (shift > PAGE_SHIFT) { - /* Get PDP, PD or PT index from GPA */ - index = (gpa >> shift) & PT_INDEX_MASK; - if (pt[index] == 0) { - ERR("No entry for GPA:0x%lx.", gpa); - return (GPA_UNMAPPED); - } - - if (pt[index] & PG_PS) { - break; - } - - pa = PTE2PA(pt[index]);; - pt = (uint64_t *)PHYS_TO_DMAP(pa); - shift -= PT_SHIFT; - } - - index = (gpa >> shift) & PT_INDEX_MASK; - if (pt[index] == 0) { - ERR("No mapping for GPA:0x%lx.\n", gpa); - return (GPA_UNMAPPED); - } - - /* Add GPA offset to HPA */ - pgmask = (1UL << shift) - 1; - hpa = (PTE2PA(pt[index]) & ~pgmask) | (gpa & pgmask); - - return (hpa); -} +SYSCTL_DECL(_hw_vmm); +SYSCTL_NODE(_hw_vmm, OID_AUTO, npt, CTLFLAG_RW, NULL, NULL); +static int npt_flags; +SYSCTL_INT(_hw_vmm_npt, OID_AUTO, pmap_flags, CTLFLAG_RD, + &npt_flags, 0, NULL); /* * AMD nested page table init. */ int svm_npt_init(void) { + int enable_superpage = 1; + + TUNABLE_INT_FETCH("hw.vmm.npt.enable_superpage", &enable_superpage); + if (enable_superpage) + npt_flags |= PMAP_PDE_SUPERPAGE; return (0); } -/* - * Free Page Table page. - */ -static void -free_pt(pd_entry_t pde) -{ - pt_entry_t *pt; - pt = (pt_entry_t *)PHYS_TO_DMAP(PTE2PA(pde)); - free(pt, M_SVM); -} -/* - * Free Page Directory page. - */ -static void -free_pd(pdp_entry_t pdpe) +static int +npt_pinit(pmap_t pmap) { - pd_entry_t *pd; - int i; - - pd = (pd_entry_t *)PHYS_TO_DMAP(PTE2PA(pdpe)); - for (i = 0; i < NPDEPG; i++) { - /* Skip not-present or superpage entries */ - if ((pd[i] == 0) || (pd[i] & PG_PS)) - continue; - free_pt(pd[i]); - } - - free(pd, M_SVM); + return (pmap_pinit_type(pmap, PT_RVI, npt_flags)); } -/* - * Free Page Directory Pointer page. - */ -static void -free_pdp(pml4_entry_t pml4e) +struct vmspace * +svm_npt_alloc(vm_offset_t min, vm_offset_t max) { - pdp_entry_t *pdp; - int i; - - pdp = (pdp_entry_t *)PHYS_TO_DMAP(PTE2PA(pml4e)); - for (i = 0; i < NPDPEPG; i++) { - /* Skip not-present or superpage entries */ - if ((pdp[i] == 0) || (pdp[i] & PG_PS)) - continue; - - free_pd(pdp[i]); - } - - free(pdp, M_SVM); + + return (vmspace_alloc(min, max, npt_pinit)); } -/* - * Free the guest's nested page table. - */ -int -svm_npt_cleanup(struct svm_softc *svm_sc) +void +svm_npt_free(struct vmspace *vmspace) { - pml4_entry_t *pml4; - int i; - pml4 = svm_sc->np_pml4; - - for (i = 0; i < NPML4EPG; i++) { - if (pml4[i] != 0) { - free_pdp(pml4[i]); - pml4[i] = 0; - } - } - - return (0); + vmspace_free(vmspace); } Modified: projects/bhyve_svm/sys/amd64/vmm/amd/npt.h ============================================================================== --- projects/bhyve_svm/sys/amd64/vmm/amd/npt.h Wed Dec 18 22:31:53 2013 (r259578) +++ projects/bhyve_svm/sys/amd64/vmm/amd/npt.h Wed Dec 18 23:39:42 2013 (r259579) @@ -31,10 +31,7 @@ struct svm_softc; -int svm_npt_init(void); -int svm_npt_cleanup(struct svm_softc *sc); -vm_paddr_t svm_npt_vmmap_get(void *arg, vm_paddr_t gpa); -int svm_npt_vmmap_set(void *arg, vm_paddr_t gpa, vm_paddr_t hpa, - size_t len, vm_memattr_t attr, int prot, - boolean_t sp); +int svm_npt_init(void); +struct vmspace *svm_npt_alloc(vm_offset_t min, vm_offset_t max); +void svm_npt_free(struct vmspace *vmspace); #endif /* _SVM_NPT_H_ */ Modified: projects/bhyve_svm/sys/amd64/vmm/amd/svm.c ============================================================================== --- projects/bhyve_svm/sys/amd64/vmm/amd/svm.c Wed Dec 18 22:31:53 2013 (r259578) +++ projects/bhyve_svm/sys/amd64/vmm/amd/svm.c Wed Dec 18 23:39:42 2013 (r259579) @@ -61,7 +61,7 @@ __FBSDID("$FreeBSD$"); #include "npt.h" /* - * SVM CPUID function 0x8000_000Ai, edx bit decoding. + * SVM CPUID function 0x8000_000A, edx bit decoding. */ #define AMD_CPUID_SVM_NP BIT(0) /* Nested paging or RVI */ #define AMD_CPUID_SVM_LBR BIT(1) /* Last branch virtualization */ @@ -79,7 +79,7 @@ MALLOC_DEFINE(M_SVM, "svm", "svm"); /* Per-CPU context area. */ extern struct pcpu __pcpu[]; -static int svm_vmexit(struct svm_softc *svm_sc, int vcpu, +static bool svm_vmexit(struct svm_softc *svm_sc, int vcpu, struct vm_exit *vmexit); static int svm_msr_rw_ok(uint8_t *btmap, uint64_t msr); static int svm_msr_index(uint64_t msr, int *index, int *bit); @@ -98,11 +98,6 @@ static uint32_t guest_asid = 1; */ static int max_asid; -/* - * Statistics - */ -static VMM_STAT_AMD(VMEXIT_NPF_LAPIC, "vm exits due to Local APIC access"); - /* * SVM host state saved area of size 4KB for each core. */ @@ -113,6 +108,8 @@ static uint8_t hsave[MAXCPU][PAGE_SIZE] */ static struct svm_regctx host_ctx[MAXCPU]; +static VMM_STAT_AMD(VCPU_EXITINTINFO, "Valid EXITINTINFO"); + /* * Common function to enable or disabled SVM for a CPU. */ @@ -123,19 +120,13 @@ cpu_svm_enable_disable(boolean_t enable) efer_msr = rdmsr(MSR_EFER); - if (enable) { + if (enable) efer_msr |= EFER_SVM; - } else { + else efer_msr &= ~EFER_SVM; - } wrmsr(MSR_EFER, efer_msr); - if(rdmsr(MSR_EFER) != efer_msr) { - ERR("SVM couldn't be enabled on CPU%d.\n", curcpu); - return (EIO); - } - return(0); } @@ -199,20 +190,16 @@ svm_cpuid_features(void) } /* - * XXX: BHyVe need EPT or RVI to work. + * bhyve need RVI to work. */ if (!(svm_feature & AMD_CPUID_SVM_NP)) { printf("Missing Nested paging or RVI SVM support in processor.\n"); return (EIO); } - if (svm_feature & (AMD_CPUID_SVM_NRIP_SAVE | - AMD_CPUID_SVM_DECODE_ASSIST)) { + if (svm_feature & AMD_CPUID_SVM_NRIP_SAVE) return (0); - } - /* XXX: Should never be here? */ - printf("Processor doesn't support nRIP or decode assist, can't" - "run BhyVe.\n"); + return (EIO); } @@ -267,16 +254,16 @@ svm_init(void) int err; err = is_svm_enabled(); - if (err) { + if (err) return (err); - } + svm_npt_init(); /* Start SVM on all CPUs */ smp_rendezvous(NULL, svm_enable, NULL, NULL); - return(0); + return (0); } /* @@ -383,7 +370,7 @@ svm_init_vcpu(struct svm_vcpu *vcpu, vm_ * Initialise a virtual machine. */ static void * -svm_vminit(struct vm *vm) +svm_vminit(struct vm *vm, pmap_t pmap) { struct svm_softc *svm_sc; vm_paddr_t msrpm_pa, iopm_pa, pml4_pa; @@ -401,10 +388,10 @@ svm_vminit(struct vm *vm) svm_sc->vm = vm; svm_sc->svm_feature = svm_feature; svm_sc->vcpu_cnt = VM_MAXCPU; - + svm_sc->nptp = (vm_offset_t)vtophys(pmap->pm_pml4); /* * Each guest has its own unique ASID. - * ASID(Addres Space Identifier) are used by TLB entries. + * ASID(Address Space Identifier) is used by TLB entry. */ svm_sc->asid = guest_asid++; @@ -438,7 +425,7 @@ svm_vminit(struct vm *vm) /* Cache physical address for multiple vcpus. */ iopm_pa = vtophys(svm_sc->iopm_bitmap); msrpm_pa = vtophys(svm_sc->msr_bitmap); - pml4_pa = vtophys(svm_sc->np_pml4); + pml4_pa = svm_sc->nptp; for (i = 0; i < svm_sc->vcpu_cnt; i++) { if (svm_init_vcpu(svm_get_vcpu(svm_sc, i), iopm_pa, msrpm_pa, @@ -458,7 +445,7 @@ cleanup: /* * Handle guest I/O intercept. */ -static int +static bool svm_handle_io(struct svm_softc *svm_sc, int vcpu, struct vm_exit *vmexit) { struct vmcb_ctrl *ctrl; @@ -477,74 +464,39 @@ svm_handle_io(struct svm_softc *svm_sc, vmexit->u.inout.port = (uint16_t)(info1 >> 16); vmexit->u.inout.eax = (uint32_t)(state->rax); - return (1); + return (false); } -/* - * SVM Nested Page(RVI) Fault handler. - * Nested page fault handler used by local APIC emulation. - */ -static int -svm_handle_npf(struct vm *vm, int vcpu, uint64_t gpa, uint64_t rip, - uint64_t exitinfo1, uint64_t cr3, struct vie *vie) +static void +svm_npf_paging(uint64_t exitinfo1, int *type, int *prot) { - int err; - if (exitinfo1 & VMCB_NPF_INFO1_ID) { - VMM_CTR0(vm, vcpu, "SVM:NPF for code access."); - return (0); - } + if (exitinfo1 & VMCB_NPF_INFO1_W) + *type = VM_PROT_WRITE; + else + *type = VM_PROT_READ; - if (exitinfo1 & VMCB_NPF_INFO1_RSV) { - VMM_CTR0(vm, vcpu, "SVM:NPF reserved bits are set."); - return (0); - } - - if (exitinfo1 & VMCB_NPF_INFO1_GPT) { - VMM_CTR0(vm, vcpu, "SVM:NPF during guest page table walk."); - return (0); - } + /* XXX: protection is not used. */ + *prot = 0; +} - /* - * nRIP is NULL for NPF so we don't have the length of instruction, - * we rely on instruction decode s/w to determine the size of - * instruction. - * - * XXX: DecodeAssist can use instruction from buffer. - */ - if (vmm_fetch_instruction(vm, vcpu, rip, VIE_INST_SIZE, - cr3, vie) != 0) { - ERR("SVM:NPF instruction fetch failed, RIP:0x%lx\n", rip); - return (EINVAL); +static bool +svm_npf_emul_fault(uint64_t exitinfo1) +{ + + if (exitinfo1 & VMCB_NPF_INFO1_ID) { + return (false); } - KASSERT(vie->num_valid, ("No instruction to emulate.")); - /* - * SVM doesn't provide GLA unlike Intel VM-x. VIE_INVALID_GLA - * which is a non-cannonical address indicate that GLA is not - * available to instruction emulation. - * - * XXX: Which SVM capability can provided GLA? - */ - if(vmm_decode_instruction(vm, vcpu, VIE_INVALID_GLA, vie)) { - ERR("SVM: Couldn't decode instruction.\n"); - return (0); + if (exitinfo1 & VMCB_NPF_INFO1_GPT) { + return (false); } - /* - * XXX: Decoding for user space(IOAPIC) should be done in - * user space. - */ - if (gpa < DEFAULT_APIC_BASE || gpa >= (DEFAULT_APIC_BASE + PAGE_SIZE)) { - VMM_CTR2(vm, vcpu, "SVM:NPF GPA(0x%lx) outside of local APIC" - " range(0x%x)\n", gpa, DEFAULT_APIC_BASE); - return (0); + if ((exitinfo1 & VMCB_NPF_INFO1_GPA) == 0) { + return (false); } - err = vmm_emulate_instruction(vm, vcpu, gpa, vie, lapic_mmio_read, - lapic_mmio_write, 0); - - return (err ? 0 : 1); + return (true); } /* @@ -571,12 +523,12 @@ svm_efer(struct svm_softc *svm_sc, int v } /* - * Determine the cause of virtual cpu exit and return to user space if exit - * demand so. - * Return: 1 - Return to user space. - * 0 - Continue vcpu run. + * Determine the cause of virtual cpu exit and handle VMEXIT. + * Return: false - Break vcpu execution loop and handle vmexit + * in kernel or user space. + * true - Continue vcpu run. */ -static int +static bool svm_vmexit(struct svm_softc *svm_sc, int vcpu, struct vm_exit *vmexit) { struct vmcb_state *state; @@ -584,35 +536,27 @@ svm_vmexit(struct svm_softc *svm_sc, int struct svm_regctx *ctx; uint64_t code, info1, info2, val; uint32_t eax, ecx, edx; - int user; /* Flag for user mode */ - int update_rip; /* Flag for updating RIP */ - int inst_len; + bool update_rip, loop; KASSERT(vcpu < svm_sc->vcpu_cnt, ("Guest doesn't have VCPU%d", vcpu)); state = svm_get_vmcb_state(svm_sc, vcpu); ctrl = svm_get_vmcb_ctrl(svm_sc, vcpu); ctx = svm_get_guest_regctx(svm_sc, vcpu); - update_rip = 1; - user = 0; - - vmexit->exitcode = VM_EXITCODE_VMX; - vmexit->u.vmx.error = 0; - code = ctrl->exitcode; + code = ctrl->exitcode; info1 = ctrl->exitinfo1; info2 = ctrl->exitinfo2; - if (ctrl->nrip) { - inst_len = ctrl->nrip - state->rip; - } else { - inst_len = ctrl->inst_decode_size; - } + update_rip = true; + loop = true; + vmexit->exitcode = VM_EXITCODE_VMX; + vmexit->u.vmx.error = 0; switch (code) { case VMCB_EXIT_MC: /* Machine Check. */ vmm_stat_incr(svm_sc->vm, vcpu, VMEXIT_MTRAP, 1); vmexit->exitcode = VM_EXITCODE_MTRAP; - user = 1; + loop = false; break; case VMCB_EXIT_MSR: /* MSR access. */ @@ -628,27 +572,29 @@ svm_vmexit(struct svm_softc *svm_sc, int if (info1) { /* VM exited because of write MSR */ - vmm_stat_incr(svm_sc->vm, vcpu, VMEXIT_WRMSR, 1); + vmm_stat_incr(svm_sc->vm, vcpu, + VMEXIT_WRMSR, 1); vmexit->exitcode = VM_EXITCODE_WRMSR; vmexit->u.msr.code = ecx; val = (uint64_t)edx << 32 | eax; if (emulate_wrmsr(svm_sc->vm, vcpu, ecx, val)) { vmexit->u.msr.wval = val; - user = 1; + loop = false; } VMM_CTR3(svm_sc->vm, vcpu, "VMEXIT WRMSR(%s handling) 0x%lx @0x%x", - user ? "user" : "kernel", val, ecx); + loop ? "kernel" : "user", val, ecx); } else { - vmm_stat_incr(svm_sc->vm, vcpu, VMEXIT_RDMSR, 1); + vmm_stat_incr(svm_sc->vm, vcpu, + VMEXIT_RDMSR, 1); vmexit->exitcode = VM_EXITCODE_RDMSR; vmexit->u.msr.code = ecx; if (emulate_rdmsr(svm_sc->vm, vcpu, ecx)) { - user = 1; + loop = false; } VMM_CTR3(svm_sc->vm, vcpu, "SVM:VMEXIT RDMSR" - " 0x%lx,%lx @0x%x", ctx->e.g.sctx_rdx, - state->rax, ecx); + " MSB=0x%08x, LSB=%08x @0x%x", + ctx->e.g.sctx_rdx, state->rax, ecx); } #define MSR_AMDK8_IPM 0xc0010055 @@ -659,17 +605,16 @@ svm_vmexit(struct svm_softc *svm_sc, int * XXX: special handling of AMD C1E - Ignore. */ if (ecx == MSR_AMDK8_IPM) - user = 0; + loop = true; break; - case VMCB_EXIT_INTR: + case VMCB_EXIT_INTR: /* * Exit on External Interrupt. * Give host interrupt handler to run and if its guest * interrupt, local APIC will inject event in guest. */ - user = 0; - update_rip = 0; + update_rip = false; VMM_CTR1(svm_sc->vm, vcpu, "SVM:VMEXIT ExtInt" " RIP:0x%lx.\n", state->rip); vmm_stat_incr(svm_sc->vm, vcpu, VMEXIT_EXTINT, 1); @@ -677,9 +622,8 @@ svm_vmexit(struct svm_softc *svm_sc, int case VMCB_EXIT_IO: *** DIFF OUTPUT TRUNCATED AT 1000 LINES ***