diff options
Diffstat (limited to 'arch/s390/kvm')
| -rw-r--r-- | arch/s390/kvm/Kconfig | 1 | ||||
| -rw-r--r-- | arch/s390/kvm/Makefile | 2 | ||||
| -rw-r--r-- | arch/s390/kvm/gaccess.c | 1 | ||||
| -rw-r--r-- | arch/s390/kvm/intercept.c | 56 | ||||
| -rw-r--r-- | arch/s390/kvm/interrupt.c | 32 | ||||
| -rw-r--r-- | arch/s390/kvm/kvm-s390.c | 5 | ||||
| -rw-r--r-- | arch/s390/kvm/kvm-s390.h | 5 | ||||
| -rw-r--r-- | arch/s390/kvm/sthyi.c | 469 | ||||
| -rw-r--r-- | arch/s390/kvm/trace-s390.h | 1 | ||||
| -rw-r--r-- | arch/s390/kvm/trace.h | 1 | ||||
| -rw-r--r-- | arch/s390/kvm/vsie.c | 50 |
11 files changed, 108 insertions, 515 deletions
diff --git a/arch/s390/kvm/Kconfig b/arch/s390/kvm/Kconfig index b1900239b0ab..9a4594e0a1ff 100644 --- a/arch/s390/kvm/Kconfig +++ b/arch/s390/kvm/Kconfig @@ -1,3 +1,4 @@ +# SPDX-License-Identifier: GPL-2.0 # # KVM configuration # diff --git a/arch/s390/kvm/Makefile b/arch/s390/kvm/Makefile index 09a9e6dfc09f..6048b1c6e580 100644 --- a/arch/s390/kvm/Makefile +++ b/arch/s390/kvm/Makefile @@ -12,6 +12,6 @@ common-objs = $(KVM)/kvm_main.o $(KVM)/eventfd.o $(KVM)/async_pf.o $(KVM)/irqch ccflags-y := -Ivirt/kvm -Iarch/s390/kvm kvm-objs := $(common-objs) kvm-s390.o intercept.o interrupt.o priv.o sigp.o -kvm-objs += diag.o gaccess.o guestdbg.o sthyi.o vsie.o +kvm-objs += diag.o gaccess.o guestdbg.o vsie.o obj-$(CONFIG_KVM) += kvm.o diff --git a/arch/s390/kvm/gaccess.c b/arch/s390/kvm/gaccess.c index 3cc77391a102..c24bfa72baf7 100644 --- a/arch/s390/kvm/gaccess.c +++ b/arch/s390/kvm/gaccess.c @@ -1,3 +1,4 @@ +// SPDX-License-Identifier: GPL-2.0 /* * guest access functions * diff --git a/arch/s390/kvm/intercept.c b/arch/s390/kvm/intercept.c index a4752bf6b526..8fe034beb623 100644 --- a/arch/s390/kvm/intercept.c +++ b/arch/s390/kvm/intercept.c @@ -18,6 +18,7 @@ #include <asm/kvm_host.h> #include <asm/asm-offsets.h> #include <asm/irq.h> +#include <asm/sysinfo.h> #include "kvm-s390.h" #include "gaccess.h" @@ -360,6 +361,61 @@ static int handle_partial_execution(struct kvm_vcpu *vcpu) return -EOPNOTSUPP; } +/* + * Handle the sthyi instruction that provides the guest with system + * information, like current CPU resources available at each level of + * the machine. + */ +int handle_sthyi(struct kvm_vcpu *vcpu) +{ + int reg1, reg2, r = 0; + u64 code, addr, cc = 0, rc = 0; + struct sthyi_sctns *sctns = NULL; + + if (!test_kvm_facility(vcpu->kvm, 74)) + return kvm_s390_inject_program_int(vcpu, PGM_OPERATION); + + kvm_s390_get_regs_rre(vcpu, ®1, ®2); + code = vcpu->run->s.regs.gprs[reg1]; + addr = vcpu->run->s.regs.gprs[reg2]; + + vcpu->stat.instruction_sthyi++; + VCPU_EVENT(vcpu, 3, "STHYI: fc: %llu addr: 0x%016llx", code, addr); + trace_kvm_s390_handle_sthyi(vcpu, code, addr); + + if (reg1 == reg2 || reg1 & 1 || reg2 & 1) + return kvm_s390_inject_program_int(vcpu, PGM_SPECIFICATION); + + if (code & 0xffff) { + cc = 3; + rc = 4; + goto out; + } + + if (addr & ~PAGE_MASK) + return kvm_s390_inject_program_int(vcpu, PGM_SPECIFICATION); + + sctns = (void *)get_zeroed_page(GFP_KERNEL); + if (!sctns) + return -ENOMEM; + + cc = sthyi_fill(sctns, &rc); + +out: + if (!cc) { + r = write_guest(vcpu, addr, reg2, sctns, PAGE_SIZE); + if (r) { + free_page((unsigned long)sctns); + return kvm_s390_inject_prog_cond(vcpu, r); + } + } + + free_page((unsigned long)sctns); + vcpu->run->s.regs.gprs[reg2 + 1] = rc; + kvm_s390_set_psw_cc(vcpu, cc); + return r; +} + static int handle_operexc(struct kvm_vcpu *vcpu) { psw_t oldpsw, newpsw; diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c index a832ad031cee..fa557372d600 100644 --- a/arch/s390/kvm/interrupt.c +++ b/arch/s390/kvm/interrupt.c @@ -213,6 +213,16 @@ static inline unsigned long pending_irqs(struct kvm_vcpu *vcpu) vcpu->arch.local_int.pending_irqs; } +static inline int isc_to_irq_type(unsigned long isc) +{ + return IRQ_PEND_IO_ISC_0 + isc; +} + +static inline int irq_type_to_isc(unsigned long irq_type) +{ + return irq_type - IRQ_PEND_IO_ISC_0; +} + static unsigned long disable_iscs(struct kvm_vcpu *vcpu, unsigned long active_mask) { @@ -220,7 +230,7 @@ static unsigned long disable_iscs(struct kvm_vcpu *vcpu, for (i = 0; i <= MAX_ISC; i++) if (!(vcpu->arch.sie_block->gcr[6] & isc_to_isc_bits(i))) - active_mask &= ~(1UL << (IRQ_PEND_IO_ISC_0 + i)); + active_mask &= ~(1UL << (isc_to_irq_type(i))); return active_mask; } @@ -901,7 +911,7 @@ static int __must_check __deliver_io(struct kvm_vcpu *vcpu, fi = &vcpu->kvm->arch.float_int; spin_lock(&fi->lock); - isc_list = &fi->lists[irq_type - IRQ_PEND_IO_ISC_0]; + isc_list = &fi->lists[irq_type_to_isc(irq_type)]; inti = list_first_entry_or_null(isc_list, struct kvm_s390_interrupt_info, list); @@ -1074,6 +1084,12 @@ void kvm_s390_vcpu_wakeup(struct kvm_vcpu *vcpu) * in kvm_vcpu_block without having the waitqueue set (polling) */ vcpu->valid_wakeup = true; + /* + * This is mostly to document, that the read in swait_active could + * be moved before other stores, leading to subtle races. + * All current users do not store or use an atomic like update + */ + smp_mb__after_atomic(); if (swait_active(&vcpu->wq)) { /* * The vcpu gave up the cpu voluntarily, mark it as a good @@ -1395,7 +1411,7 @@ static struct kvm_s390_interrupt_info *get_io_int(struct kvm *kvm, list_del_init(&iter->list); fi->counters[FIRQ_CNTR_IO] -= 1; if (list_empty(isc_list)) - clear_bit(IRQ_PEND_IO_ISC_0 + isc, &fi->pending_irqs); + clear_bit(isc_to_irq_type(isc), &fi->pending_irqs); spin_unlock(&fi->lock); return iter; } @@ -1522,7 +1538,7 @@ static int __inject_io(struct kvm *kvm, struct kvm_s390_interrupt_info *inti) isc = int_word_to_isc(inti->io.io_int_word); list = &fi->lists[FIRQ_LIST_IO_ISC_0 + isc]; list_add_tail(&inti->list, list); - set_bit(IRQ_PEND_IO_ISC_0 + isc, &fi->pending_irqs); + set_bit(isc_to_irq_type(isc), &fi->pending_irqs); spin_unlock(&fi->lock); return 0; } @@ -2175,6 +2191,8 @@ static int clear_io_irq(struct kvm *kvm, struct kvm_device_attr *attr) return -EINVAL; if (copy_from_user(&schid, (void __user *) attr->addr, sizeof(schid))) return -EFAULT; + if (!schid) + return -EINVAL; kfree(kvm_s390_get_io_int(kvm, isc_mask, schid)); /* * If userspace is conforming to the architecture, we can have at most @@ -2483,11 +2501,11 @@ void kvm_s390_reinject_machine_check(struct kvm_vcpu *vcpu, mci.val = mcck_info->mcic; if (mci.sr) - cr14 |= MCCK_CR14_RECOVERY_SUB_MASK; + cr14 |= CR14_RECOVERY_SUBMASK; if (mci.dg) - cr14 |= MCCK_CR14_DEGRAD_SUB_MASK; + cr14 |= CR14_DEGRADATION_SUBMASK; if (mci.w) - cr14 |= MCCK_CR14_WARN_SUB_MASK; + cr14 |= CR14_WARNING_SUBMASK; mchk = mci.ck ? &inti.mchk : &irq.u.mchk; mchk->cr14 = cr14; diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c index 40d0a1a97889..98ad8b9e0360 100644 --- a/arch/s390/kvm/kvm-s390.c +++ b/arch/s390/kvm/kvm-s390.c @@ -395,6 +395,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext) case KVM_CAP_S390_USER_INSTR0: case KVM_CAP_S390_CMMA_MIGRATION: case KVM_CAP_S390_AIS: + case KVM_CAP_S390_AIS_MIGRATION: r = 1; break; case KVM_CAP_S390_MEM_OP: @@ -1884,8 +1885,6 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) rc = -ENOMEM; - ratelimit_state_init(&kvm->arch.sthyi_limit, 5 * HZ, 500); - kvm->arch.use_esca = 0; /* start with basic SCA */ if (!sclp.has_64bscao) alloc_flags |= GFP_DMA; @@ -3283,7 +3282,7 @@ static void sync_regs(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run) */ if ((kvm_run->kvm_dirty_regs & KVM_SYNC_RICCB) && test_kvm_facility(vcpu->kvm, 64) && - riccb->valid && + riccb->v && !(vcpu->arch.sie_block->ecb3 & ECB3_RI)) { VCPU_EVENT(vcpu, 3, "%s", "ENABLE: RI (sync_regs)"); vcpu->arch.sie_block->ecb3 |= ECB3_RI; diff --git a/arch/s390/kvm/kvm-s390.h b/arch/s390/kvm/kvm-s390.h index 9f8fdd7b2311..10d65dfbc306 100644 --- a/arch/s390/kvm/kvm-s390.h +++ b/arch/s390/kvm/kvm-s390.h @@ -242,6 +242,8 @@ static inline void kvm_s390_retry_instr(struct kvm_vcpu *vcpu) kvm_s390_rewind_psw(vcpu, kvm_s390_get_ilen(vcpu)); } +int handle_sthyi(struct kvm_vcpu *vcpu); + /* implemented in priv.c */ int is_valid_psw(psw_t *psw); int kvm_s390_handle_aa(struct kvm_vcpu *vcpu); @@ -268,9 +270,6 @@ void kvm_s390_vsie_destroy(struct kvm *kvm); int kvm_s390_handle_sigp(struct kvm_vcpu *vcpu); int kvm_s390_handle_sigp_pei(struct kvm_vcpu *vcpu); -/* implemented in sthyi.c */ -int handle_sthyi(struct kvm_vcpu *vcpu); - /* implemented in kvm-s390.c */ void kvm_s390_set_tod_clock_ext(struct kvm *kvm, const struct kvm_s390_vm_tod_clock *gtod); diff --git a/arch/s390/kvm/sthyi.c b/arch/s390/kvm/sthyi.c deleted file mode 100644 index 395926b8c1ed..000000000000 --- a/arch/s390/kvm/sthyi.c +++ /dev/null @@ -1,469 +0,0 @@ -/* - * store hypervisor information instruction emulation functions. - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License (version 2 only) - * as published by the Free Software Foundation. - * - * Copyright IBM Corp. 2016 - * Author(s): Janosch Frank <frankja@linux.vnet.ibm.com> - */ -#include <linux/kvm_host.h> -#include <linux/errno.h> -#include <linux/pagemap.h> -#include <linux/vmalloc.h> -#include <linux/ratelimit.h> - -#include <asm/kvm_host.h> -#include <asm/asm-offsets.h> -#include <asm/sclp.h> -#include <asm/diag.h> -#include <asm/sysinfo.h> -#include <asm/ebcdic.h> - -#include "kvm-s390.h" -#include "gaccess.h" -#include "trace.h" - -#define DED_WEIGHT 0xffff -/* - * CP and IFL as EBCDIC strings, SP/0x40 determines the end of string - * as they are justified with spaces. - */ -#define CP 0xc3d7404040404040UL -#define IFL 0xc9c6d34040404040UL - -enum hdr_flags { - HDR_NOT_LPAR = 0x10, - HDR_STACK_INCM = 0x20, - HDR_STSI_UNAV = 0x40, - HDR_PERF_UNAV = 0x80, -}; - -enum mac_validity { - MAC_NAME_VLD = 0x20, - MAC_ID_VLD = 0x40, - MAC_CNT_VLD = 0x80, -}; - -enum par_flag { - PAR_MT_EN = 0x80, -}; - -enum par_validity { - PAR_GRP_VLD = 0x08, - PAR_ID_VLD = 0x10, - PAR_ABS_VLD = 0x20, - PAR_WGHT_VLD = 0x40, - PAR_PCNT_VLD = 0x80, -}; - -struct hdr_sctn { - u8 infhflg1; - u8 infhflg2; /* reserved */ - u8 infhval1; /* reserved */ - u8 infhval2; /* reserved */ - u8 reserved[3]; - u8 infhygct; - u16 infhtotl; - u16 infhdln; - u16 infmoff; - u16 infmlen; - u16 infpoff; - u16 infplen; - u16 infhoff1; - u16 infhlen1; - u16 infgoff1; - u16 infglen1; - u16 infhoff2; - u16 infhlen2; - u16 infgoff2; - u16 infglen2; - u16 infhoff3; - u16 infhlen3; - u16 infgoff3; - u16 infglen3; - u8 reserved2[4]; -} __packed; - -struct mac_sctn { - u8 infmflg1; /* reserved */ - u8 infmflg2; /* reserved */ - u8 infmval1; - u8 infmval2; /* reserved */ - u16 infmscps; - u16 infmdcps; - u16 infmsifl; - u16 infmdifl; - char infmname[8]; - char infmtype[4]; - char infmmanu[16]; - char infmseq[16]; - char infmpman[4]; - u8 reserved[4]; -} __packed; - -struct par_sctn { - u8 infpflg1; - u8 infpflg2; /* reserved */ - u8 infpval1; - u8 infpval2; /* reserved */ - u16 infppnum; - u16 infpscps; - u16 infpdcps; - u16 infpsifl; - u16 infpdifl; - u16 reserved; - char infppnam[8]; - u32 infpwbcp; - u32 infpabcp; - u32 infpwbif; - u32 infpabif; - char infplgnm[8]; - u32 infplgcp; - u32 infplgif; -} __packed; - -struct sthyi_sctns { - struct hdr_sctn hdr; - struct mac_sctn mac; - struct par_sctn par; -} __packed; - -struct cpu_inf { - u64 lpar_cap; - u64 lpar_grp_cap; - u64 lpar_weight; - u64 all_weight; - int cpu_num_ded; - int cpu_num_shd; -}; - -struct lpar_cpu_inf { - struct cpu_inf cp; - struct cpu_inf ifl; -}; - -static inline u64 cpu_id(u8 ctidx, void *diag224_buf) -{ - return *((u64 *)(diag224_buf + (ctidx + 1) * DIAG204_CPU_NAME_LEN)); -} - -/* - * Scales the cpu capping from the lpar range to the one expected in - * sthyi data. - * - * diag204 reports a cap in hundredths of processor units. - * z/VM's range for one core is 0 - 0x10000. - */ -static u32 scale_cap(u32 in) -{ - return (0x10000 * in) / 100; -} - -static void fill_hdr(struct sthyi_sctns *sctns) -{ - sctns->hdr.infhdln = sizeof(sctns->hdr); - sctns->hdr.infmoff = sizeof(sctns->hdr); - sctns->hdr.infmlen = sizeof(sctns->mac); - sctns->hdr.infplen = sizeof(sctns->par); - sctns->hdr.infpoff = sctns->hdr.infhdln + sctns->hdr.infmlen; - sctns->hdr.infhtotl = sctns->hdr.infpoff + sctns->hdr.infplen; -} - -static void fill_stsi_mac(struct sthyi_sctns *sctns, - struct sysinfo_1_1_1 *sysinfo) -{ - if (stsi(sysinfo, 1, 1, 1)) - return; - - sclp_ocf_cpc_name_copy(sctns->mac.infmname); - - memcpy(sctns->mac.infmtype, sysinfo->type, sizeof(sctns->mac.infmtype)); - memcpy(sctns->mac.infmmanu, sysinfo->manufacturer, sizeof(sctns->mac.infmmanu)); - memcpy(sctns->mac.infmpman, sysinfo->plant, sizeof(sctns->mac.infmpman)); - memcpy(sctns->mac.infmseq, sysinfo->sequence, sizeof(sctns->mac.infmseq)); - - sctns->mac.infmval1 |= MAC_ID_VLD | MAC_NAME_VLD; -} - -static void fill_stsi_par(struct sthyi_sctns *sctns, - struct sysinfo_2_2_2 *sysinfo) -{ - if (stsi(sysinfo, 2, 2, 2)) - return; - - sctns->par.infppnum = sysinfo->lpar_number; - memcpy(sctns->par.infppnam, sysinfo->name, sizeof(sctns->par.infppnam)); - - sctns->par.infpval1 |= PAR_ID_VLD; -} - -static void fill_stsi(struct sthyi_sctns *sctns) -{ - void *sysinfo; - - /* Errors are handled through the validity bits in the response. */ - sysinfo = (void *)__get_free_page(GFP_KERNEL); - if (!sysinfo) - return; - - fill_stsi_mac(sctns, sysinfo); - fill_stsi_par(sctns, sysinfo); - - free_pages((unsigned long)sysinfo, 0); -} - -static void fill_diag_mac(struct sthyi_sctns *sctns, - struct diag204_x_phys_block *block, - void *diag224_buf) -{ - int i; - - for (i = 0; i < block->hdr.cpus; i++) { - switch (cpu_id(block->cpus[i].ctidx, diag224_buf)) { - case CP: - if (block->cpus[i].weight == DED_WEIGHT) - sctns->mac.infmdcps++; - else - sctns->mac.infmscps++; - break; - case IFL: - if (block->cpus[i].weight == DED_WEIGHT) - sctns->mac.infmdifl++; - else - sctns->mac.infmsifl++; - break; - } - } - sctns->mac.infmval1 |= MAC_CNT_VLD; -} - -/* Returns a pointer to the the next partition block. */ -static struct diag204_x_part_block *lpar_cpu_inf(struct lpar_cpu_inf *part_inf, - bool this_lpar, - void *diag224_buf, - struct diag204_x_part_block *block) -{ - int i, capped = 0, weight_cp = 0, weight_ifl = 0; - struct cpu_inf *cpu_inf; - - for (i = 0; i < block->hdr.rcpus; i++) { - if (!(block->cpus[i].cflag & DIAG204_CPU_ONLINE)) - continue; - - switch (cpu_id(block->cpus[i].ctidx, diag224_buf)) { - case CP: - cpu_inf = &part_inf->cp; - if (block->cpus[i].cur_weight < DED_WEIGHT) - weight_cp |= block->cpus[i].cur_weight; - break; - case IFL: - cpu_inf = &part_inf->ifl; - if (block->cpus[i].cur_weight < DED_WEIGHT) - weight_ifl |= block->cpus[i].cur_weight; - break; - default: - continue; - } - - if (!this_lpar) - continue; - - capped |= block->cpus[i].cflag & DIAG204_CPU_CAPPED; - cpu_inf->lpar_cap |= block->cpus[i].cpu_type_cap; - cpu_inf->lpar_grp_cap |= block->cpus[i].group_cpu_type_cap; - - if (block->cpus[i].weight == DED_WEIGHT) - cpu_inf->cpu_num_ded += 1; - else - cpu_inf->cpu_num_shd += 1; - } - - if (this_lpar && capped) { - part_inf->cp.lpar_weight = weight_cp; - part_inf->ifl.lpar_weight = weight_ifl; - } - part_inf->cp.all_weight += weight_cp; - part_inf->ifl.all_weight += weight_ifl; - return (struct diag204_x_part_block *)&block->cpus[i]; -} - -static void fill_diag(struct sthyi_sctns *sctns) -{ - int i, r, pages; - bool this_lpar; - void *diag204_buf; - void *diag224_buf = NULL; - struct diag204_x_info_blk_hdr *ti_hdr; - struct diag204_x_part_block *part_block; - struct diag204_x_phys_block *phys_block; - struct lpar_cpu_inf lpar_inf = {}; - - /* Errors are handled through the validity bits in the response. */ - pages = diag204((unsigned long)DIAG204_SUBC_RSI | - (unsigned long)DIAG204_INFO_EXT, 0, NULL); - if (pages <= 0) - return; - - diag204_buf = vmalloc(PAGE_SIZE * pages); - if (!diag204_buf) - return; - - r = diag204((unsigned long)DIAG204_SUBC_STIB7 | - (unsigned long)DIAG204_INFO_EXT, pages, diag204_buf); - if (r < 0) - goto out; - - diag224_buf = (void *)__get_free_page(GFP_KERNEL | GFP_DMA); - if (!diag224_buf || diag224(diag224_buf)) - goto out; - - ti_hdr = diag204_buf; - part_block = diag204_buf + sizeof(*ti_hdr); - - for (i = 0; i < ti_hdr->npar; i++) { - /* - * For the calling lpar we also need to get the cpu - * caps and weights. The time information block header - * specifies the offset to the partition block of the - * caller lpar, so we know when we process its data. - */ - this_lpar = (void *)part_block - diag204_buf == ti_hdr->this_part; - part_block = lpar_cpu_inf(&lpar_inf, this_lpar, diag224_buf, - part_block); - } - - phys_block = (struct diag204_x_phys_block *)part_block; - part_block = diag204_buf + ti_hdr->this_part; - if (part_block->hdr.mtid) - sctns->par.infpflg1 = PAR_MT_EN; - - sctns->par.infpval1 |= PAR_GRP_VLD; - sctns->par.infplgcp = scale_cap(lpar_inf.cp.lpar_grp_cap); - sctns->par.infplgif = scale_cap(lpar_inf.ifl.lpar_grp_cap); - memcpy(sctns->par.infplgnm, part_block->hdr.hardware_group_name, - sizeof(sctns->par.infplgnm)); - - sctns->par.infpscps = lpar_inf.cp.cpu_num_shd; - sctns->par.infpdcps = lpar_inf.cp.cpu_num_ded; - sctns->par.infpsifl = lpar_inf.ifl.cpu_num_shd; - sctns->par.infpdifl = lpar_inf.ifl.cpu_num_ded; - sctns->par.infpval1 |= PAR_PCNT_VLD; - - sctns->par.infpabcp = scale_cap(lpar_inf.cp.lpar_cap); - sctns->par.infpabif = scale_cap(lpar_inf.ifl.lpar_cap); - sctns->par.infpval1 |= PAR_ABS_VLD; - - /* - * Everything below needs global performance data to be - * meaningful. - */ - if (!(ti_hdr->flags & DIAG204_LPAR_PHYS_FLG)) { - sctns->hdr.infhflg1 |= HDR_PERF_UNAV; - goto out; - } - - fill_diag_mac(sctns, phys_block, diag224_buf); - - if (lpar_inf.cp.lpar_weight) { - sctns->par.infpwbcp = sctns->mac.infmscps * 0x10000 * - lpar_inf.cp.lpar_weight / lpar_inf.cp.all_weight; - } - - if (lpar_inf.ifl.lpar_weight) { - sctns->par.infpwbif = sctns->mac.infmsifl * 0x10000 * - lpar_inf.ifl.lpar_weight / lpar_inf.ifl.all_weight; - } - sctns->par.infpval1 |= PAR_WGHT_VLD; - -out: - free_page((unsigned long)diag224_buf); - vfree(diag204_buf); -} - -static int sthyi(u64 vaddr) -{ - register u64 code asm("0") = 0; - register u64 addr asm("2") = vaddr; - int cc; - - asm volatile( - ".insn rre,0xB2560000,%[code],%[addr]\n" - "ipm %[cc]\n" - "srl %[cc],28\n" - : [cc] "=d" (cc) - : [code] "d" (code), [addr] "a" (addr) - : "3", "memory", "cc"); - return cc; -} - -int handle_sthyi(struct kvm_vcpu *vcpu) -{ - int reg1, reg2, r = 0; - u64 code, addr, cc = 0; - struct sthyi_sctns *sctns = NULL; - - if (!test_kvm_facility(vcpu->kvm, 74)) - return kvm_s390_inject_program_int(vcpu, PGM_OPERATION); - - /* - * STHYI requires extensive locking in the higher hypervisors - * and is very computational/memory expensive. Therefore we - * ratelimit the executions per VM. - */ - if (!__ratelimit(&vcpu->kvm->arch.sthyi_limit)) { - kvm_s390_retry_instr(vcpu); - return 0; - } - - kvm_s390_get_regs_rre(vcpu, ®1, ®2); - code = vcpu->run->s.regs.gprs[reg1]; - addr = vcpu->run->s.regs.gprs[reg2]; - - vcpu->stat.instruction_sthyi++; - VCPU_EVENT(vcpu, 3, "STHYI: fc: %llu addr: 0x%016llx", code, addr); - trace_kvm_s390_handle_sthyi(vcpu, code, addr); - - if (reg1 == reg2 || reg1 & 1 || reg2 & 1) - return kvm_s390_inject_program_int(vcpu, PGM_SPECIFICATION); - - if (code & 0xffff) { - cc = 3; - goto out; - } - - if (addr & ~PAGE_MASK) - return kvm_s390_inject_program_int(vcpu, PGM_SPECIFICATION); - - sctns = (void *)get_zeroed_page(GFP_KERNEL); - if (!sctns) - return -ENOMEM; - - /* - * If we are a guest, we don't want to emulate an emulated - * instruction. We ask the hypervisor to provide the data. - */ - if (test_facility(74)) { - cc = sthyi((u64)sctns); - goto out; - } - - fill_hdr(sctns); - fill_stsi(sctns); - fill_diag(sctns); - -out: - if (!cc) { - r = write_guest(vcpu, addr, reg2, sctns, PAGE_SIZE); - if (r) { - free_page((unsigned long)sctns); - return kvm_s390_inject_prog_cond(vcpu, r); - } - } - - free_page((unsigned long)sctns); - vcpu->run->s.regs.gprs[reg2 + 1] = cc ? 4 : 0; - kvm_s390_set_psw_cc(vcpu, cc); - return r; -} diff --git a/arch/s390/kvm/trace-s390.h b/arch/s390/kvm/trace-s390.h index 78b7e847984a..6f0209d45164 100644 --- a/arch/s390/kvm/trace-s390.h +++ b/arch/s390/kvm/trace-s390.h @@ -1,3 +1,4 @@ +/* SPDX-License-Identifier: GPL-2.0 */ #if !defined(_TRACE_KVMS390_H) || defined(TRACE_HEADER_MULTI_READ) #define _TRACE_KVMS390_H diff --git a/arch/s390/kvm/trace.h b/arch/s390/kvm/trace.h index 4fc9d4e5be89..aa419eb6a0c8 100644 --- a/arch/s390/kvm/trace.h +++ b/arch/s390/kvm/trace.h @@ -1,3 +1,4 @@ +/* SPDX-License-Identifier: GPL-2.0 */ #if !defined(_TRACE_KVM_H) || defined(TRACE_HEADER_MULTI_READ) #define _TRACE_KVM_H diff --git a/arch/s390/kvm/vsie.c b/arch/s390/kvm/vsie.c index b18b5652e5c5..a311938b63b3 100644 --- a/arch/s390/kvm/vsie.c +++ b/arch/s390/kvm/vsie.c @@ -443,22 +443,14 @@ static int map_prefix(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page) * * Returns: - 0 on success * - -EINVAL if the gpa is not valid guest storage - * - -ENOMEM if out of memory */ static int pin_guest_page(struct kvm *kvm, gpa_t gpa, hpa_t *hpa) { struct page *page; - hva_t hva; - int rc; - hva = gfn_to_hva(kvm, gpa_to_gfn(gpa)); - if (kvm_is_error_hva(hva)) + page = gfn_to_page(kvm, gpa_to_gfn(gpa)); + if (is_error_page(page)) return -EINVAL; - rc = get_user_pages_fast(hva, 1, 1, &page); - if (rc < 0) - return rc; - else if (rc != 1) - return -ENOMEM; *hpa = (hpa_t) page_to_virt(page) + (gpa & ~PAGE_MASK); return 0; } @@ -466,11 +458,7 @@ static int pin_guest_page(struct kvm *kvm, gpa_t gpa, hpa_t *hpa) /* Unpins a page previously pinned via pin_guest_page, marking it as dirty. */ static void unpin_guest_page(struct kvm *kvm, gpa_t gpa, hpa_t hpa) { - struct page *page; - - page = virt_to_page(hpa); - set_page_dirty_lock(page); - put_page(page); + kvm_release_pfn_dirty(hpa >> PAGE_SHIFT); /* mark the page always as dirty for migration */ mark_page_dirty(kvm, gpa_to_gfn(gpa)); } @@ -557,7 +545,7 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page) rc = set_validity_icpt(scb_s, 0x003bU); if (!rc) { rc = pin_guest_page(vcpu->kvm, gpa, &hpa); - if (rc == -EINVAL) + if (rc) rc = set_validity_icpt(scb_s, 0x0034U); } if (rc) @@ -574,10 +562,10 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page) } /* 256 bytes cannot cross page boundaries */ rc = pin_guest_page(vcpu->kvm, gpa, &hpa); - if (rc == -EINVAL) + if (rc) { rc = set_validity_icpt(scb_s, 0x0080U); - if (rc) goto unpin; + } scb_s->itdba = hpa; } @@ -592,10 +580,10 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page) * if this block gets bigger, we have to shadow it. */ rc = pin_guest_page(vcpu->kvm, gpa, &hpa); - if (rc == -EINVAL) + if (rc) { rc = set_validity_icpt(scb_s, 0x1310U); - if (rc) goto unpin; + } scb_s->gvrd = hpa; } @@ -607,11 +595,11 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page) } /* 64 bytes cannot cross page boundaries */ rc = pin_guest_page(vcpu->kvm, gpa, &hpa); - if (rc == -EINVAL) + if (rc) { rc = set_validity_icpt(scb_s, 0x0043U); - /* Validity 0x0044 will be checked by SIE */ - if (rc) goto unpin; + } + /* Validity 0x0044 will be checked by SIE */ scb_s->riccbd = hpa; } if ((scb_s->ecb & ECB_GS) && !(scb_s->ecd & ECD_HOSTREGMGMT)) { @@ -635,10 +623,10 @@ static int pin_blocks(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page) * cross page boundaries */ rc = pin_guest_page(vcpu->kvm, gpa, &hpa); - if (rc == -EINVAL) + if (rc) { rc = set_validity_icpt(scb_s, 0x10b0U); - if (rc) goto unpin; + } scb_s->sdnxo = hpa | sdnxc; } return 0; @@ -663,7 +651,6 @@ static void unpin_scb(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page, * * Returns: - 0 if the scb was pinned. * - > 0 if control has to be given to guest 2 - * - -ENOMEM if out of memory */ static int pin_scb(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page, gpa_t gpa) @@ -672,14 +659,13 @@ static int pin_scb(struct kvm_vcpu *vcpu, struct vsie_page *vsie_page, int rc; rc = pin_guest_page(vcpu->kvm, gpa, &hpa); - if (rc == -EINVAL) { + if (rc) { rc = kvm_s390_inject_program_int(vcpu, PGM_ADDRESSING); - if (!rc) - rc = 1; + WARN_ON_ONCE(rc); + return 1; } - if (!rc) - vsie_page->scb_o = (struct kvm_s390_sie_block *) hpa; - return rc; + vsie_page->scb_o = (struct kvm_s390_sie_block *) hpa; + return 0; } /* |