add per-domain IOMMU control
v13 cover add per-domain IOMMU control
These are the remaining uncommitted patches from my previous series:
https://lists.xenproject.org/archives/html/xen-devel/2019-09/msg01208.html
The only patch that has been revised is patch #4 (previously patch #6).
Ian Jackson (1):
tools/ocaml: abi check: Cope with consecutive relevant enums
Paul Durrant (3):
remove late (on-demand) construction of IOMMU page tables
iommu: tidy up iommu_use_hap_pt() and need_iommu_pt_sync() macros introduce a 'passthrough' configuration option to xl.cfg...
docs/man/xl.cfg.5.pod.in | 57 +++++++++ tools/libxl/libxl.h | 16 +++ tools/libxl/libxl_create.c | 30 ++++- tools/libxl/libxl_mem.c | 6 +- tools/libxl/libxl_types.idl | 9 ++ tools/libxl/libxl_utils.c | 15 +++ tools/libxl/libxl_utils.h | 1 + tools/ocaml/libs/xc/abi-check | 17 +-- tools/ocaml/libs/xc/xenctrl.ml | 4 + tools/ocaml/libs/xc/xenctrl.mli | 5 + tools/ocaml/libs/xc/xenctrl_stubs.c | 17 ++- tools/xl/xl_parse.c | 178 ++++++++++++++++++-------- xen/arch/arm/Kconfig | 1 + xen/arch/arm/domain.c | 10 +- xen/arch/arm/p2m.c | 2 +- xen/arch/x86/dom0_build.c | 2 +- xen/arch/x86/domain.c | 2 +- xen/arch/x86/hvm/mtrr.c | 5 +- xen/arch/x86/mm/mem_sharing.c | 2 +- xen/arch/x86/mm/p2m.c | 4 +- xen/arch/x86/mm/paging.c | 2 +- xen/arch/x86/x86_64/mm.c | 2 +- xen/common/domain.c | 7 + xen/common/domctl.c | 13 -- xen/common/memory.c | 4 +- xen/common/vm_event.c | 2 +- xen/drivers/passthrough/Kconfig | 3 + xen/drivers/passthrough/device_tree.c | 11 -- xen/drivers/passthrough/iommu.c | 147 ++++++--------------- xen/drivers/passthrough/pci.c | 12 -- xen/drivers/passthrough/vtd/iommu.c | 10 +- xen/drivers/passthrough/x86/iommu.c | 97 -------------- xen/include/asm-arm/iommu.h | 3 - xen/include/asm-x86/iommu.h | 4 - xen/include/public/domctl.h | 10 +- xen/include/xen/iommu.h | 40 +++--- xen/include/xen/sched.h | 8 -- 37 files changed, 388 insertions(+), 370 deletions(-)
---
Cc: Andrew Cooper <[email protected]>
Cc: Anthony PERARD <[email protected]>
Cc: David Scott <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Ian Jackson <[email protected]>
Cc: Konrad Rzeszutek Wilk <[email protected]>
Cc: Petre Pircalabu <[email protected]>
Cc: "Roger Pau Monné" <[email protected]>
Cc: Stefano Stabellini <[email protected]>
Cc: Tamas K Lengyel <[email protected]>
Cc: Tim Deegan <[email protected]>
Cc: Volodymyr Babchuk <[email protected]>
Cc: Wei Liu <[email protected]>
Signature · 7 lines
2.20.1.2.gb21ebb671 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH v13 1/4 remove late (on-demand) construction of IOMMU page tables
Now that there is a per-domain IOMMU-enable flag, which should be set if any device is going to be passed through, stop deferring page table construction until the assignment is done. Also don't tear down the tables again when the last device is de-assigned; defer that task until domain destruction.
This allows the has_iommu_pt() helper and iommu_status enumeration to be removed. Calls to has_iommu_pt() are simply replaced by calls to is_iommu_enabled(). Remaining open-coded tests of iommu_hap_pt_share can also be replaced by calls to iommu_use_hap_pt().
The arch_iommu_populate_page_table() and iommu_construct() functions become redundant, as does the 'strict mode' dom0 page_list mapping code in iommu_hwdom_init(), and iommu_teardown() can be made static is its only remaining caller, iommu_domain_destroy(), is within the same source module.
All in all, about 220 lines of code are removed from the hypervisor (at the expense of some additions in the toolstack).
NOTE: This patch will cause a small amount of extra resource to be used
to accommodate IOMMU page tables that may never be used, since the per-domain IOMMU-enable flag is currently set to the value of the global iommu_enable flag. A subsequent patch will add an option to the toolstack to allow it to be turned off if there is no intention to assign passthrough hardware to the domain.
To account for the extra resource, 'iommu_memkb' has been added to domain_build_info. This patch sets it to a value calculated based on the domain's maximum memory when the P2M sharing is either not supported or globally disabled, or zero otherwise. However, when the toolstack option mentioned above is added, it will also be zero if the per-domain IOMMU-enable flag is turned off.
---
Cc: Stefano Stabellini <[email protected]>
Cc: Volodymyr Babchuk <[email protected]>
Cc: Andrew Cooper <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Ian Jackson <[email protected]>
Cc: Konrad Rzeszutek Wilk <[email protected]>
Cc: Tim Deegan <[email protected]>
Cc: Wei Liu <[email protected]>
Cc: "Roger Pau Monné" <[email protected]>
Cc: Tamas K Lengyel <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Petre Pircalabu <[email protected]>
Previously part of series https://lists.xenproject.org/archives/html/xen-devel/2019-07/msg02267.html
v9:
- Avoid the iommu_memkb overhead if the IOMMU is disable or page tables are shared
v7:
- Add toolstack memory reservation for IOMMU page tables... Re-use of shadow calculation didn't seem appropriate so a new helper function is added
v5:
- Minor style fixes
--- tools/libxl/libxl.h | 7 ++ tools/libxl/libxl_mem.c | 6 +- tools/libxl/libxl_types.idl | 1 + tools/libxl/libxl_utils.c | 15 +++ tools/libxl/libxl_utils.h | 1 + tools/xl/xl_parse.c | 24 ++++- xen/arch/arm/p2m.c | 2 +- xen/arch/x86/dom0_build.c | 2 +- xen/arch/x86/hvm/mtrr.c | 5 +- xen/arch/x86/mm/mem_sharing.c | 2 +- xen/arch/x86/mm/p2m.c | 4 +- xen/arch/x86/mm/paging.c | 2 +- xen/arch/x86/x86_64/mm.c | 2 +- xen/common/memory.c | 4 +- xen/common/vm_event.c | 2 +- xen/drivers/passthrough/device_tree.c | 11 --- xen/drivers/passthrough/iommu.c | 134 ++++++-------------------- xen/drivers/passthrough/pci.c | 12 --- xen/drivers/passthrough/vtd/iommu.c | 10 +- xen/drivers/passthrough/x86/iommu.c | 97 ------------------- xen/include/asm-arm/iommu.h | 2 +- xen/include/asm-x86/iommu.h | 2 +- xen/include/xen/iommu.h | 16 --- xen/include/xen/sched.h | 2 - 24 files changed, 94 insertions(+), 271 deletions(-)
tools/libxl/libxl.h +7 -0
@@ -408,6 +408,13 @@*/ #define LIBXL_HAVE_PHYSINFO_CAP_IOMMU_HAP_PT_SHARE 1 +/* +* LIBXL_HAVE_BUILDINFO_IOMMU_MEMKB indicates thate libxl_domain_build_info +* has an iommu_memkb field which should be set with the amount of memory +* overhead needed by the domain for populating IOMMU page tables. +*/ +#define LIBXL_HAVE_BUILDINFO_IOMMU_MEMKB 1 +/* * libxl ABI compatibility *
tools/libxl/libxl_mem.c +4 -2
@@ -461,15 +461,17 @@ int libxl_domain_need_memory(libxl_ctx *ctx,if (rc) goto out; *need_memkb = b_info->target_memkb; + *need_memkb += b_info->shadow_memkb + b_info->iommu_memkb; +switch (b_info->type) {case LIBXL_DOMAIN_TYPE_PVH :case LIBXL_DOMAIN_TYPE_HVM : - *need_memkb += b_info->shadow_memkb +LIBXL_HVM_EXTRA_MEMORY ; + *need_memkb +=LIBXL_HVM_EXTRA_MEMORY ;if (libxl_defbool_val (b_info->device_model_stubdomain)) *need_memkb +=32 *1024 ;break ;case LIBXL_DOMAIN_TYPE_PV : - *need_memkb += b_info->shadow_memkb +LIBXL_PV_EXTRA_MEMORY ; + *need_memkb +=LIBXL_PV_EXTRA_MEMORY ;break ;default : rc =ERROR_INVAL ;
tools/libxl/libxl_types.idl +1 -0
@@ -486,6 +486,7 @@ libxl_domain_build_info = Struct("domain_build_info",[ ("target_memkb", MemKB), ("video_memkb", MemKB), ("shadow_memkb", MemKB), + ("iommu_memkb", MemKB), ("rtc_timeoffset", uint32), ("exec_ssidref", uint32), ("exec_ssid_label", string),
tools/libxl/libxl_utils.c +15 -0
@@ -48,6 +48,21 @@ unsigned long libxl_get_required_shadow_memory(unsigned long maxmem_kb, unsignedreturn 4 * (256 * smp_cpus +2 * (maxmem_kb /1024 )); } +unsigned long libxl_get_required_iommu_memory (unsigned long maxmem_kb) +{ +unsigned long iommu_pages =0 , mem_pages = maxmem_kb /4 ; +unsigned int level; + +/* Assume a 4 level page table with 512 entries per level */ +for (level =0 ; level <4 ; level++) + { + mem_pages =DIV_ROUNDUP (mem_pages,512 ); + iommu_pages += mem_pages; + } + +return iommu_pages *4 ; +} +char *libxl_domid_to_name (libxl_ctx *ctx,uint32_t domid) {unsigned int len;
tools/libxl/libxl_utils.h +1 -0
@@ -24,6 +24,7 @@ constchar *libxl_basename (const char *name);/* returns string from strdup */ unsigned long libxl_get_required_shadow_memory (unsigned long maxmem_kb,unsigned int smp_cpus); +unsigned long libxl_get_required_iommu_memory (unsigned long maxmem_kb);int libxl_name_to_domid (libxl_ctx *ctx,const char *name,uint32_t *domid);int libxl_domain_qualifier_to_domid (libxl_ctx *ctx,const char *name,uint32_t *domid);char *libxl_domid_to_name (libxl_ctx *ctx,uint32_t domid);
tools/xl/xl_parse.c +22 -2
@@ -1207,6 +1207,7 @@ void parse_config_data(const char *config_source,int config_len, libxl_domain_config *d_config) { + libxl_physinfo physinfo;const char *buf;long l, vcpus =0 ;XLU_Config *config;
@@ -1221,10 +1222,22 @@ void parse_config_data(const char *config_source,int pci_seize =0 ;int i, e;char *kernel_basename; +bool iommu_enabled, iommu_hap_pt_share;libxl_domain_create_info *c_info = &d_config->c_info;libxl_domain_build_info *b_info = &d_config->b_info; +libxl_physinfo_init (&physinfo); +if (libxl_get_physinfo (ctx, &physinfo) !=0 ) { +libxl_physinfo_dispose (&physinfo); +fprintf (stderr,"libxl_get_physinfo failed\n" ); +exit (EXIT_FAILURE ); + } + + iommu_enabled = physinfo.cap_hvm_directio; + iommu_hap_pt_share = physinfo.cap_iommu_hap_pt_share; +libxl_physinfo_dispose (&physinfo); + config=xlu_cfg_init (stderr, config_source);if (!config) {fprintf (stderr,"Failed to allocate for configuration\n" );
@@ -1448,14 +1461,21 @@ void parse_config_data(const char *config_source,exit (1 ); } -/* libxl_get_required_shadow_memory() must be called after final values +/* libxl_get_required_shadow_memory() and +* libxl_get_required_iommu_memory() must be called after final values * (default or specified) for vcpus and memory are set, because the -* calculation depends on those values. */ +* calculations depend on those values. */ b_info->shadow_memkb = !xlu_cfg_get_long (config,"shadow_memory" , &l,0 ) ? l *1024 :libxl_get_required_shadow_memory (b_info->max_memkb, b_info->max_vcpus); +/* No IOMMU reservation is needed if either the IOMMU is disabled or it +* can share the P2M. */ + b_info->iommu_memkb = (!iommu_enabled || iommu_hap_pt_share) + ?0 + :libxl_get_required_iommu_memory (b_info->max_memkb); +xlu_cfg_get_defbool (config,"nomigrate" , &b_info->disable_migrate,0 );if (!xlu_cfg_get_long (config,"tsc_mode" , &l,1 )) {
xen/arch/arm/p2m.c +1 -1
@@ -1056,7 +1056,7 @@ static int __p2m_set_entry(struct p2m_domain *p2m, !mfn_eq(lpae_get_mfn(*entry), lpae_get_mfn(orig_pte)) )p2m_free_entry (p2m, orig_pte, level); -if (has_iommu_pt (p2m->domain) && +if (is_iommu_enabled (p2m->domain) && (lpae_is_valid (orig_pte) ||lpae_is_valid (*entry)) ) {unsigned int flush_flags =0 ;
xen/arch/x86/dom0_build.c +1 -1
@@ -365,7 +365,7 @@ unsigned long __init dom0_compute_nr_pages( } need_paging =is_hvm_domain (d) && - (!iommu_hap_pt_share || !paging_mode_hap (d)); + (!iommu_use_hap_pt (d) || !paging_mode_hap (d));for ( ; ; need_paging = false ) { nr_pages =get_memsize (&dom0_size, avail);
xen/arch/x86/hvm/mtrr.c +3 -2
@@ -783,7 +783,8 @@ HVM_REGISTER_SAVE_RESTORE(MTRR, hvm_save_mtrr_msr, hvm_load_mtrr_msr, 1,void memory_type_changed (struct domain *d) { -if ( (has_iommu_pt (d) ||cache_flush_permitted (d)) && d->vcpu && d->vcpu[0 ] ) +if ( (is_iommu_enabled (d) ||cache_flush_permitted (d)) && + d->vcpu && d->vcpu[0 ] ) {p2m_memory_type_changed (d);flush_all (FLUSH_CACHE );
@@ -831,7 +832,7 @@ int epte_get_entry_emt(struct domain *d, unsigned long gfn, mfn_t mfn,return MTRR_TYPE_UNCACHABLE ; } -if ( !has_iommu_pt (d) && !cache_flush_permitted (d) ) +if ( !is_iommu_enabled (d) && !cache_flush_permitted (d) ) { *ipat =1 ;return MTRR_TYPE_WRBACK ;
xen/arch/x86/mm/mem_sharing.c +1 -1
@@ -1664,7 +1664,7 @@ int mem_sharing_domctl(struct domain *d, struct xen_domctl_mem_sharing_op *mec)case XEN_DOMCTL_MEM_SHARING_CONTROL : { rc =0 ; -if (unlikely (has_iommu_pt (d) && mec->u.enable) ) +if (unlikely (is_iommu_enabled (d) && mec->u.enable) ) rc = -EXDEV ;else d->arch.hvm.mem_sharing_enabled = mec->u.enable;
xen/arch/x86/mm/p2m.c +2 -2
@@ -1341,7 +1341,7 @@ int set_identity_p2m_entry(struct domain *d, unsigned long gfn_l,if ( !paging_mode_translate (p2m->domain) ) { -if ( !has_iommu_pt (d) ) +if ( !is_iommu_enabled (d) )return 0 ;return iommu_legacy_map (d,_dfn (gfn_l),_mfn (gfn_l),PAGE_ORDER_4K , IOMMUF_readable | IOMMUF_writable);
@@ -1432,7 +1432,7 @@ int clear_identity_p2m_entry(struct domain *d, unsigned long gfn_l)if ( !paging_mode_translate (d) ) { -if ( !has_iommu_pt (d) ) +if ( !is_iommu_enabled (d) )return 0 ;return iommu_legacy_unmap (d,_dfn (gfn_l),PAGE_ORDER_4K ); }
xen/arch/x86/mm/paging.c +1 -1
@@ -213,7 +213,7 @@ int paging_log_dirty_enable(struct domain *d, bool_t log_global) {int ret; -if (has_iommu_pt (d) && log_global ) +if (is_iommu_enabled (d) && log_global ) {/* * Refuse to turn on global log-dirty mode
xen/arch/x86/x86_64/mm.c +1 -1
@@ -1434,7 +1434,7 @@ int memory_add(unsigned long spfn, unsigned long epfn, unsigned int pxm)* shared or being kept in sync then newly added memory needs to be * mapped here. */ -if (has_iommu_pt (hardware_domain) && +if (is_iommu_enabled (hardware_domain) && !iommu_use_hap_pt (hardware_domain) && !need_iommu_pt_sync (hardware_domain) ) {
xen/common/memory.c +2 -2
@@ -823,7 +823,7 @@ int xenmem_add_to_physmap(struct domain *d, struct xen_add_to_physmap *xatp, xatp->gpfn += start; xatp->size -= start; -if (has_iommu_pt (d) ) +if (is_iommu_enabled (d) )this_cpu (iommu_dont_flush_iotlb) =1 ;while ( xatp->size > done )
@@ -844,7 +844,7 @@ int xenmem_add_to_physmap(struct domain *d, struct xen_add_to_physmap *xatp, } } -if (has_iommu_pt (d) ) +if (is_iommu_enabled (d) ) {int ret;
xen/common/vm_event.c +1 -1
@@ -630,7 +630,7 @@ int vm_event_domctl(struct domain *d, struct xen_domctl_vm_event_op *vec)/* No paging if iommu is used */ rc = -EMLINK ; -if (unlikely (has_iommu_pt (d)) ) +if (unlikely (is_iommu_enabled (d)) )break ; rc = -EXDEV ;
xen/drivers/passthrough/device_tree.c +0 -11
@@ -40,17 +40,6 @@ int iommu_assign_dt_device(struct domain *d, struct dt_device_node *dev)if ( !list_empty (&dev->domain_list) ) goto fail; -/* -* The hwdom is forced to use IOMMU for protecting assigned -* device. Therefore the IOMMU data is already set up. -*/ -ASSERT (!is_hardware_domain (d) || - hd->status == IOMMU_STATUS_initialized); - - rc =iommu_construct (d); -if ( rc ) - goto fail; -/* The flag field doesn't matter to DT device. */ rc = hd->platform_ops->assign_device (d,0 ,dt_to_dev (dev),0 );
xen/drivers/passthrough/iommu.c +28 -106
182 lines not on this page · raw message
xen/drivers/passthrough/pci.c +0 -12
@@ -939,9 +939,6 @@ static int deassign_device(struct domain *d, uint16_t seg, uint8_t bus, pdev->fault.count =0 ; -if ( !has_arch_pdevs (d) &&has_iommu_pt (d) ) -iommu_teardown (d); -return ret; }
@@ -1490,13 +1487,6 @@ static int assign_device(struct domain *d, u16 seg, u8 bus, u8 devfn, u32 flag)if ( !pcidevs_trylock () )return -ERESTART ; - rc =iommu_construct (d); -if ( rc ) - { -pcidevs_unlock (); -return rc; - } - pdev =pci_get_pdev_by_domain (hardware_domain, seg, bus, devfn);if ( !pdev ) {
@@ -1525,8 +1515,6 @@ static int assign_device(struct domain *d, u16 seg, u8 bus, u8 devfn, u32 flag) } done: -if ( !has_arch_pdevs (d) &&has_iommu_pt (d) ) -iommu_teardown (d);pcidevs_unlock ();return rc;
xen/drivers/passthrough/vtd/iommu.c +1 -9
@@ -1721,15 +1721,7 @@ static void iommu_domain_teardown(struct domain *d)ASSERT (is_iommu_enabled (d)); -/* -* We can't use iommu_use_hap_pt here because either IOMMU state -* is already changed to IOMMU_STATUS_disabled at this point or -* has always been IOMMU_STATUS_disabled. -* -* We also need to test if HAP is enabled because PV guests can -* enter this path too. -*/ -if (hap_enabled (d) && iommu_hap_pt_share ) +if (iommu_use_hap_pt (d) )return ;spin_lock (&hd->arch.mapping_lock);
xen/drivers/passthrough/x86/iommu.c +0 -97
103 lines not on this page · raw message
xen/include/asm-arm/iommu.h +1 -1
@@ -21,7 +21,7 @@ struct arch_iommu };/* Always share P2M Table between the CPU and the IOMMU */ -#define iommu_use_hap_pt (d) (has_iommu_pt(d)) +#define iommu_use_hap_pt (d) is_iommu_enabled(d)const struct iommu_ops *iommu_get_ops (void );void iommu_set_ops (const struct iommu_ops *ops);
xen/include/asm-x86/iommu.h +1 -1
@@ -88,7 +88,7 @@ extern const struct iommu_init_ops *iommu_init_ops;/* Are we using the domain P2M table as its IOMMU pagetable? */ #define iommu_use_hap_pt (d) \ - (hap_enabled(d) && has_iommu_pt(d) && iommu_hap_pt_share) + (hap_enabled(d) && is_iommu_enabled(d) && iommu_hap_pt_share)void iommu_update_ire_from_apic (unsigned int apic,unsigned int reg,unsigned int value);unsigned int iommu_read_apic_from_ire (unsigned int apic,unsigned int reg);
xen/include/xen/iommu.h +0 -16
@@ -88,15 +88,9 @@ void iommu_domain_destroy(struct domain *d);void arch_iommu_domain_destroy (struct domain *d);int arch_iommu_domain_init (struct domain *d); -int arch_iommu_populate_page_table (struct domain *d);void arch_iommu_check_autotranslated_hwdom (struct domain *d);void arch_iommu_hwdom_init (struct domain *d); -int iommu_construct (struct domain *d); - -/* Function used internally, use iommu_domain_destroy */ -void iommu_teardown (struct domain *d); -/* * The following flags are passed to map operations and passed by lookup * operations.
@@ -263,13 +257,6 @@ struct iommu_ops {# define iommu_vcall iommu_call#endif -enum iommu_status -{ - IOMMU_STATUS_disabled, - IOMMU_STATUS_initializing, - IOMMU_STATUS_initialized -}; -struct domain_iommu {struct arch_iommu arch;
@@ -289,9 +276,6 @@ struct domain_iommu {/* Features supported by the IOMMU */ DECLARE_BITMAP (features, IOMMU_FEAT_count); -/* Status of guest IOMMU mappings */ -enum iommu_status status; -/* * Does the guest reqire mappings to be synchonized, to maintain * the default dfn == pfn map. (See comment on dfn at the top of
xen/include/xen/sched.h +0 -2
@@ -966,10 +966,8 @@ static inline bool is_hwdom_pinned_vcpu(const struct vcpu *v) }#ifdef CONFIG_HAS_PASSTHROUGH -#define has_iommu_pt (d) (dom_iommu(d)->status != IOMMU_STATUS_disabled)#define need_iommu_pt_sync (d) (dom_iommu(d)->need_sync)#else -#define has_iommu_pt (d) false#define need_iommu_pt_sync (d) false#endif
Signature · 7 lines
2.20.1.2.gb21ebb671 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [Xen-devel] [PATCH v13 1/4] remove late (on-demand) construction of IOMMU page tables
Libxl maintainers... ping?
Quoted 865 lines
-----Original Message-----
From: Paul Durrant <[email protected]>
Sent: 18 September 2019 11:41
To: [email protected]
Cc: Paul Durrant <[email protected]>; Alexandru Isaila <[email protected]>; Razvan Cojocaru <[email protected]>; Jan Beulich <[email protected]>; Julien Grall <[email protected]>; Stefano Stabellini <[email protected]>; Volodymyr Babchuk <[email protected]>; Andrew Cooper <[email protected]>; George Dunlap <[email protected]>; Ian Jackson <[email protected]>; Konrad Rzeszutek Wilk <[email protected]>; Tim (Xen.org) <[email protected]>; Wei Liu <[email protected]>; Roger Pau Monne <[email protected]>; Tamas K Lengyel <[email protected]>; George Dunlap <[email protected]>; Petre Pircalabu <[email protected]>
Subject: [PATCH v13 1/4] remove late (on-demand) construction of IOMMU page tablesNow that there is a per-domain IOMMU-enable flag, which should be set if any device is going to be passed through, stop deferring page table construction until the assignment is done. Also don't tear down the tables again when the last device is de-assigned; defer that task until domain destruction.
This allows the
has_iommu_pt()helper andiommu_statusenumeration to be removed. Calls tohas_iommu_pt()are simply replaced by calls tois_iommu_enabled(). Remaining open-coded tests ofiommu_hap_pt_sharecan also be replaced by calls toiommu_use_hap_pt().
Thearch_iommu_populate_page_table()andiommu_construct()functions become redundant, as does the 'strict mode' dom0page_listmapping code iniommu_hwdom_init(), andiommu_teardown()can be made static is its only remaining caller,iommu_domain_destroy(), is within the same source module.All in all, about 220 lines of code are removed from the hypervisor (at the expense of some additions in the toolstack).
NOTE: This patch will cause a small amount of extra resource to be used
to accommodate IOMMU page tables that may never be used, since the per-domain IOMMU-enable flag is currently set to the value of the globaliommu_enableflag. A subsequent patch will add an option to the toolstack to allow it to be turned off if there is no intention to assign passthrough hardware to the domain.
To account for the extra resource, 'iommu_memkb' has been added todomain_build_info. This patch sets it to a value calculated based on the domain's maximum memory when the P2M sharing is either not supported or globally disabled, or zero otherwise. However, when the toolstack option mentioned above is added, it will also be zero if the per-domain IOMMU-enable flag is turned off.Signed-off-by Paul Durrant <[email protected]>Reviewed-by Alexandru Isaila <[email protected]>Acked-by Razvan Cojocaru <[email protected]>Reviewed-by Jan Beulich <[email protected]>Acked-by Julien Grall <[email protected]>---
Cc: Stefano Stabellini <[email protected]>
Cc: Volodymyr Babchuk <[email protected]>
Cc: Andrew Cooper <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Ian Jackson <[email protected]>
Cc: Konrad Rzeszutek Wilk <[email protected]>
Cc: Tim Deegan <[email protected]>
Cc: Wei Liu <[email protected]>
Cc: "Roger Pau Monné" <[email protected]>
Cc: Tamas K Lengyel <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Petre Pircalabu <[email protected]>Previously part of series https://lists.xenproject.org/archives/html/xen-devel/2019-07/msg02267.html
v9:
- Avoid theiommu_memkboverhead if the IOMMU is disable or page tables are sharedv7:
- Add toolstack memory reservation for IOMMU page tables... Re-use of shadow calculation didn't seem appropriate so a new helper function is addedv5:
- Minor style fixes--- tools/libxl/libxl.h | 7 ++ tools/libxl/libxl_mem.c | 6 +- tools/libxl/libxl_types.idl | 1 + tools/libxl/libxl_utils.c | 15 +++ tools/libxl/libxl_utils.h | 1 + tools/xl/xl_parse.c | 24 ++++- xen/arch/arm/p2m.c | 2 +- xen/arch/x86/dom0_build.c | 2 +- xen/arch/x86/hvm/mtrr.c | 5 +- xen/arch/x86/mm/mem_sharing.c | 2 +- xen/arch/x86/mm/p2m.c | 4 +- xen/arch/x86/mm/paging.c | 2 +- xen/arch/x86/x86_64/mm.c | 2 +- xen/common/memory.c | 4 +- xen/common/vm_event.c | 2 +- xen/drivers/passthrough/device_tree.c | 11 --- xen/drivers/passthrough/iommu.c | 134 ++++++-------------------- xen/drivers/passthrough/pci.c | 12 --- xen/drivers/passthrough/vtd/iommu.c | 10 +- xen/drivers/passthrough/x86/iommu.c | 97 ------------------- xen/include/asm-arm/iommu.h | 2 +- xen/include/asm-x86/iommu.h | 2 +- xen/include/xen/iommu.h | 16 --- xen/include/xen/sched.h | 2 - 24 files changed, 94 insertions(+), 271 deletions(-)tools/libxl/libxl.h +7
-0@@ -408,6 +408,13 @@*/ #define LIBXL_HAVE_PHYSINFO_CAP_IOMMU_HAP_PT_SHARE 1 +/* +* LIBXL_HAVE_BUILDINFO_IOMMU_MEMKB indicates thate libxl_domain_build_info +* has an iommu_memkb field which should be set with the amount of memory +* overhead needed by the domain for populating IOMMU page tables. +*/ +#define LIBXL_HAVE_BUILDINFO_IOMMU_MEMKB 1 +/* * libxl ABI compatibility * tools/libxl/libxl_mem.c +4
-2@@ -461,15 +461,17 @@ int libxl_domain_need_memory(libxl_ctx *ctx,if (rc) goto out; *need_memkb = b_info->target_memkb; + *need_memkb += b_info->shadow_memkb + b_info->iommu_memkb; +switch (b_info->type) {case LIBXL_DOMAIN_TYPE_PVH :case LIBXL_DOMAIN_TYPE_HVM : - *need_memkb += b_info->shadow_memkb +LIBXL_HVM_EXTRA_MEMORY ; + *need_memkb +=LIBXL_HVM_EXTRA_MEMORY ;if (libxl_defbool_val (b_info->device_model_stubdomain)) *need_memkb +=32 *1024 ;break ;case LIBXL_DOMAIN_TYPE_PV : - *need_memkb += b_info->shadow_memkb +LIBXL_PV_EXTRA_MEMORY ; + *need_memkb +=LIBXL_PV_EXTRA_MEMORY ;break ;default : rc =ERROR_INVAL ;tools/libxl/libxl_types.idl +1
-0@@ -486,6 +486,7 @@ libxl_domain_build_info = Struct("domain_build_info",[ ("target_memkb", MemKB), ("video_memkb", MemKB), ("shadow_memkb", MemKB), + ("iommu_memkb", MemKB), ("rtc_timeoffset", uint32), ("exec_ssidref", uint32), ("exec_ssid_label", string),tools/libxl/libxl_utils.c +15
-0@@ -48,6 +48,21 @@ unsigned long libxl_get_required_shadow_memory(unsigned long maxmem_kb, unsignedreturn 4 * (256 * smp_cpus +2 * (maxmem_kb /1024 )); } +unsigned long libxl_get_required_iommu_memory (unsigned long maxmem_kb) +{ +unsigned long iommu_pages =0 , mem_pages = maxmem_kb /4 ; +unsigned int level; + +/* Assume a 4 level page table with 512 entries per level */ +for (level =0 ; level <4 ; level++) + { + mem_pages =DIV_ROUNDUP (mem_pages,512 ); + iommu_pages += mem_pages; + } + +return iommu_pages *4 ; +} +char *libxl_domid_to_name (libxl_ctx *ctx,uint32_t domid) {unsigned int len;tools/libxl/libxl_utils.h +1
-0@@ -24,6 +24,7 @@ constchar *libxl_basename (const char *name);/* returns string from strdup */ unsigned long libxl_get_required_shadow_memory (unsigned long maxmem_kb,unsigned int smp_cpus); +unsigned long libxl_get_required_iommu_memory (unsigned long maxmem_kb);int libxl_name_to_domid (libxl_ctx *ctx,const char *name,uint32_t *domid);int libxl_domain_qualifier_to_domid (libxl_ctx *ctx,const char *name,uint32_t *domid);char *libxl_domid_to_name (libxl_ctx *ctx,uint32_t domid);tools/xl/xl_parse.c +22
-2@@ -1207,6 +1207,7 @@ void parse_config_data(const char *config_source,int config_len, libxl_domain_config *d_config) { + libxl_physinfo physinfo;const char *buf;long l, vcpus =0 ;XLU_Config *config;@@ -1221,10 +1222,22 @@ void parse_config_data(const char *config_source,int pci_seize =0 ;int i, e;char *kernel_basename; +bool iommu_enabled, iommu_hap_pt_share;libxl_domain_create_info *c_info = &d_config->c_info;libxl_domain_build_info *b_info = &d_config->b_info; +libxl_physinfo_init (&physinfo); +if (libxl_get_physinfo (ctx, &physinfo) !=0 ) { +libxl_physinfo_dispose (&physinfo); +fprintf (stderr,"libxl_get_physinfo failed\n" ); +exit (EXIT_FAILURE ); + } + + iommu_enabled = physinfo.cap_hvm_directio; + iommu_hap_pt_share = physinfo.cap_iommu_hap_pt_share; +libxl_physinfo_dispose (&physinfo); + config=xlu_cfg_init (stderr, config_source);if (!config) {fprintf (stderr,"Failed to allocate for configuration\n" );@@ -1448,14 +1461,21 @@ void parse_config_data(const char *config_source,exit (1 ); } -/* libxl_get_required_shadow_memory() must be called after final values +/* libxl_get_required_shadow_memory() and +* libxl_get_required_iommu_memory() must be called after final values * (default or specified) for vcpus and memory are set, because the -* calculation depends on those values. */ +* calculations depend on those values. */ b_info->shadow_memkb = !xlu_cfg_get_long (config,"shadow_memory" , &l,0 ) ? l *1024 :libxl_get_required_shadow_memory (b_info->max_memkb, b_info->max_vcpus); +/* No IOMMU reservation is needed if either the IOMMU is disabled or it +* can share the P2M. */ + b_info->iommu_memkb = (!iommu_enabled || iommu_hap_pt_share) + ?0 + :libxl_get_required_iommu_memory (b_info->max_memkb); +xlu_cfg_get_defbool (config,"nomigrate" , &b_info->disable_migrate,0 );if (!xlu_cfg_get_long (config,"tsc_mode" , &l,1 )) {xen/arch/arm/p2m.c +1
-1@@ -1056,7 +1056,7 @@ static int __p2m_set_entry(struct p2m_domain *p2m, !mfn_eq(lpae_get_mfn(*entry), lpae_get_mfn(orig_pte)) )p2m_free_entry (p2m, orig_pte, level); -if (has_iommu_pt (p2m->domain) && +if (is_iommu_enabled (p2m->domain) && (lpae_is_valid (orig_pte) ||lpae_is_valid (*entry)) ) {unsigned int flush_flags =0 ;xen/arch/x86/dom0_build.c +1
-1@@ -365,7 +365,7 @@ unsigned long __init dom0_compute_nr_pages( } need_paging =is_hvm_domain (d) && - (!iommu_hap_pt_share || !paging_mode_hap (d)); + (!iommu_use_hap_pt (d) || !paging_mode_hap (d));for ( ; ; need_paging = false ) { nr_pages =get_memsize (&dom0_size, avail);xen/arch/x86/hvm/mtrr.c +3
-2@@ -783,7 +783,8 @@ HVM_REGISTER_SAVE_RESTORE(MTRR, hvm_save_mtrr_msr, hvm_load_mtrr_msr, 1,void memory_type_changed (struct domain *d) { -if ( (has_iommu_pt (d) ||cache_flush_permitted (d)) && d->vcpu && d->vcpu[0 ] ) +if ( (is_iommu_enabled (d) ||cache_flush_permitted (d)) && + d->vcpu && d->vcpu[0 ] ) {p2m_memory_type_changed (d);flush_all (FLUSH_CACHE );@@ -831,7 +832,7 @@ int epte_get_entry_emt(struct domain *d, unsigned long gfn, mfn_t mfn,return MTRR_TYPE_UNCACHABLE ; } -if ( !has_iommu_pt (d) && !cache_flush_permitted (d) ) +if ( !is_iommu_enabled (d) && !cache_flush_permitted (d) ) { *ipat =1 ;return MTRR_TYPE_WRBACK ;xen/arch/x86/mm/mem_sharing.c +1
-1@@ -1664,7 +1664,7 @@ int mem_sharing_domctl(struct domain *d, struct xen_domctl_mem_sharing_op *mec)case XEN_DOMCTL_MEM_SHARING_CONTROL : { rc =0 ; -if (unlikely (has_iommu_pt (d) && mec->u.enable) ) +if (unlikely (is_iommu_enabled (d) && mec->u.enable) ) rc = -EXDEV ;else d->arch.hvm.mem_sharing_enabled = mec->u.enable;xen/arch/x86/mm/p2m.c +2
-2@@ -1341,7 +1341,7 @@ int set_identity_p2m_entry(struct domain *d, unsigned long gfn_l,if ( !paging_mode_translate (p2m->domain) ) { -if ( !has_iommu_pt (d) ) +if ( !is_iommu_enabled (d) )return 0 ;return iommu_legacy_map (d,_dfn (gfn_l),_mfn (gfn_l),PAGE_ORDER_4K , IOMMUF_readable | IOMMUF_writable);@@ -1432,7 +1432,7 @@ int clear_identity_p2m_entry(struct domain *d, unsigned long gfn_l)if ( !paging_mode_translate (d) ) { -if ( !has_iommu_pt (d) ) +if ( !is_iommu_enabled (d) )return 0 ;return iommu_legacy_unmap (d,_dfn (gfn_l),PAGE_ORDER_4K ); }xen/arch/x86/mm/paging.c +1
-1@@ -213,7 +213,7 @@ int paging_log_dirty_enable(struct domain *d, bool_t log_global) {int ret; -if (has_iommu_pt (d) && log_global ) +if (is_iommu_enabled (d) && log_global ) {/* * Refuse to turn on global log-dirty mode xen/arch/x86/x86_64/mm.c +1
-1@@ -1434,7 +1434,7 @@ int memory_add(unsigned long spfn, unsigned long epfn, unsigned int pxm)* shared or being kept in sync then newly added memory needs to be * mapped here. */ -if (has_iommu_pt (hardware_domain) && +if (is_iommu_enabled (hardware_domain) && !iommu_use_hap_pt (hardware_domain) && !need_iommu_pt_sync (hardware_domain) ) {xen/common/memory.c +2
-2@@ -823,7 +823,7 @@ int xenmem_add_to_physmap(struct domain *d, struct xen_add_to_physmap *xatp, xatp->gpfn += start; xatp->size -= start; -if (has_iommu_pt (d) ) +if (is_iommu_enabled (d) )this_cpu (iommu_dont_flush_iotlb) =1 ;while ( xatp->size > done )@@ -844,7 +844,7 @@ int xenmem_add_to_physmap(struct domain *d, struct xen_add_to_physmap *xatp, } } -if (has_iommu_pt (d) ) +if (is_iommu_enabled (d) ) {int ret;xen/common/vm_event.c +1
-1@@ -630,7 +630,7 @@ int vm_event_domctl(struct domain *d, struct xen_domctl_vm_event_op *vec)/* No paging if iommu is used */ rc = -EMLINK ; -if (unlikely (has_iommu_pt (d)) ) +if (unlikely (is_iommu_enabled (d)) )break ; rc = -EXDEV ;xen/drivers/passthrough/device_tree.c +0
-11@@ -40,17 +40,6 @@ int iommu_assign_dt_device(struct domain *d, struct dt_device_node *dev)if ( !list_empty (&dev->domain_list) ) goto fail; -/* -* The hwdom is forced to use IOMMU for protecting assigned -* device. Therefore the IOMMU data is already set up. -*/ -ASSERT (!is_hardware_domain (d) || - hd->status == IOMMU_STATUS_initialized); - - rc =iommu_construct (d); -if ( rc ) - goto fail; -/* The flag field doesn't matter to DT device. */ rc = hd->platform_ops->assign_device (d,0 ,dt_to_dev (dev),0 );xen/drivers/passthrough/iommu.c +28
-106@@ -152,6 +152,17 @@ static int __init parse_dom0_iommu_param(const char *s) }custom_param ("dom0-iommu" , parse_dom0_iommu_param); +static void __hwdom_init check_hwdom_reqs(struct domain *d) +{ +if ( iommu_hwdom_none || !paging_mode_translate (d) ) +return ; + +arch_iommu_check_autotranslated_hwdom (d); + + iommu_hwdom_passthrough = false; + iommu_hwdom_strict = true; +} +int iommu_domain_init (struct domain *d) {struct domain_iommu *hd =dom_iommu (d);@@ -169,129 +180,44 @@ int iommu_domain_init(struct domain *d)return ret; hd->platform_ops =iommu_get_ops (); -return hd->platform_ops->init (d); -} + ret = hd->platform_ops->init (d); +if ( ret ) +return ret; -static void __hwdom_init check_hwdom_reqs(struct domain *d) -{ -if ( iommu_hwdom_none || !paging_mode_translate (d) ) -return ; +if (is_hardware_domain (d) ) +check_hwdom_reqs (d);/* may modify iommu_hwdom_strict */ -arch_iommu_check_autotranslated_hwdom (d); +/* +* NB: 'relaxed' h/w domains don't need the IOMMU mappings to be kept +* in-sync with their assigned pages because all host RAM will be +* mapped during hwdom_init(). +*/ +if ( !is_hardware_domain (d) || iommu_hwdom_strict ) + hd->need_sync = !iommu_use_hap_pt (d); - iommu_hwdom_passthrough = false; - iommu_hwdom_strict = true; +return 0 ; } void __hwdom_init iommu_hwdom_init(struct domain *d) {struct domain_iommu *hd =dom_iommu (d); -check_hwdom_reqs (d); -if ( !is_iommu_enabled (d) )return ;register_keyhandler ('o' , &iommu_dump_p2m_table,"dump iommu p2m table" ,0 ); - hd->status = IOMMU_STATUS_initializing; -/* -* NB: relaxed hw domains don't need sync because all ram is already -* mapped in the iommu page tables. -*/ - hd->need_sync = iommu_hwdom_strict && !iommu_use_hap_pt (d); -if (need_iommu_pt_sync (d) ) - { -struct page_info *page; -unsigned int i =0 , flush_flags =0 ; -int rc =0 ; - - page_list_for_each ( page, &d->page_list ) - { -unsigned long mfn =mfn_x (page_to_mfn (page)); -unsigned long dfn =mfn_to_gmfn (d, mfn); -unsigned int mapping = IOMMUF_readable; -int ret; - -if ( ((page->u.inuse.type_info & PGT_count_mask) ==0 ) || - ((page->u.inuse.type_info & PGT_type_mask) - == PGT_writable_page) ) - mapping |= IOMMUF_writable; - - ret =iommu_map (d,_dfn (dfn),_mfn (mfn),0 , mapping, - &flush_flags); - -if ( !rc ) - rc = ret; - -if ( !(i++ &0xfffff ) ) -process_pending_softirqs (); - } - -/* Use while-break to avoid compiler warning */ -while (iommu_iotlb_flush_all (d, flush_flags) ) -break ; - -if ( rc ) -printk (XENLOG_WARNING "d%d: IOMMU mapping failed: %d\n" , - d->domain_id, rc); - } - hd->platform_ops->hwdom_init (d); - - hd->status = IOMMU_STATUS_initialized; } -void iommu_teardown (struct domain *d) +static void iommu_teardown (struct domain *d) {struct domain_iommu *hd =dom_iommu (d); - hd->status = IOMMU_STATUS_disabled; hd->platform_ops->teardown (d);tasklet_schedule (&iommu_pt_cleanup_tasklet); } -int iommu_construct (struct domain *d) -{ -struct domain_iommu *hd =dom_iommu (d); - -if ( hd->status == IOMMU_STATUS_initialized ) -return 0 ; - - hd->status = IOMMU_STATUS_initializing; - -if ( !iommu_use_hap_pt (d) ) - { -int rc; - - hd->need_sync = true; - - rc =arch_iommu_populate_page_table (d); -if ( rc ) - { -if ( rc != -ERESTART ) - { - hd->need_sync = false; - hd->status = IOMMU_STATUS_disabled; - } - -return rc; - } - } - - hd->status = IOMMU_STATUS_initialized; - -/* -* There may be dirty cache lines when a device is assigned -* and before has_iommu_pt(d) becoming true, this will cause -* memory_type_changed lose effect if memory type changes. -* Call memory_type_changed here to amend this. -*/ -memory_type_changed (d); - -return 0 ; -} -void iommu_domain_destroy (struct domain *d) {if ( !is_iommu_enabled (d) )@@ -587,11 +513,8 @@ int iommu_do_domctl(void iommu_share_p2m_table (struct domain * d) {ASSERT (hap_enabled (d)); -/* -* iommu_use_hap_pt(d) cannot be used here because during domain -* construction has_iommu_pt(d) will always return false here. -*/ -if (is_iommu_enabled (d) && iommu_hap_pt_share ) + +if (iommu_use_hap_pt (d) )iommu_get_ops ()->share_p2m (d); }@@ -638,8 +561,7 @@ static void iommu_dump_p2m_table(unsigned char key) ops =iommu_get_ops ();for_each_domain (d) { -if (is_hardware_domain (d) || -dom_iommu (d)->status < IOMMU_STATUS_initialized ) +if (is_hardware_domain (d) || !is_iommu_enabled (d) )continue ;if (iommu_use_hap_pt (d) )xen/drivers/passthrough/pci.c +0
-12@@ -939,9 +939,6 @@ static int deassign_device(struct domain *d, uint16_t seg, uint8_t bus, pdev->fault.count =0 ; -if ( !has_arch_pdevs (d) &&has_iommu_pt (d) ) -iommu_teardown (d); -return ret; }@@ -1490,13 +1487,6 @@ static int assign_device(struct domain *d, u16 seg, u8 bus, u8 devfn, u32 flag)if ( !pcidevs_trylock () )return -ERESTART ; - rc =iommu_construct (d); -if ( rc ) - { -pcidevs_unlock (); -return rc; - } - pdev =pci_get_pdev_by_domain (hardware_domain, seg, bus, devfn);if ( !pdev ) {@@ -1525,8 +1515,6 @@ static int assign_device(struct domain *d, u16 seg, u8 bus, u8 devfn, u32 flag) } done: -if ( !has_arch_pdevs (d) &&has_iommu_pt (d) ) -iommu_teardown (d);pcidevs_unlock ();return rc;xen/drivers/passthrough/vtd/iommu.c +1
-9@@ -1721,15 +1721,7 @@ static void iommu_domain_teardown(struct domain *d)ASSERT (is_iommu_enabled (d)); -/* -* We can't use iommu_use_hap_pt here because either IOMMU state -* is already changed to IOMMU_STATUS_disabled at this point or -* has always been IOMMU_STATUS_disabled. -* -* We also need to test if HAP is enabled because PV guests can -* enter this path too. -*/ -if (hap_enabled (d) && iommu_hap_pt_share ) +if (iommu_use_hap_pt (d) )return ;spin_lock (&hd->arch.mapping_lock);xen/drivers/passthrough/x86/iommu.c +0
-97@@ -81,103 +81,6 @@ int __init iommu_setup_hpet_msi(struct msi_desc *msi)return ops->setup_hpet_msi ? ops->setup_hpet_msi (msi) : -ENODEV ; } -int arch_iommu_populate_page_table (struct domain *d) -{ -struct page_info *page; -int rc =0 , n =0 ; - -spin_lock (&d->page_alloc_lock); - -if (unlikely (d->is_dying) ) - rc = -ESRCH ; - -while ( !rc && (page =page_list_remove_head (&d->page_list)) ) - { -if (is_hvm_domain (d) || - (page->u.inuse.type_info & PGT_type_mask) == PGT_writable_page ) - { -mfn_t mfn =page_to_mfn (page); -gfn_t gfn =mfn_to_gfn (d, mfn); -unsigned int flush_flags =0 ; - -if ( !gfn_eq (gfn,INVALID_GFN ) ) - { -dfn_t dfn =_dfn (gfn_x (gfn)); - -ASSERT (!(gfn_x (gfn) >>DEFAULT_DOMAIN_ADDRESS_WIDTH )); -BUG_ON (SHARED_M2P (gfn_x (gfn))); - rc =iommu_map (d, dfn, mfn,PAGE_ORDER_4K , - IOMMUF_readable | IOMMUF_writable, - &flush_flags); - -/* -* We may be working behind the back of a running guest, which -* may change the type of a page at any time. We can't prevent -* this (for instance, by bumping the type count while mapping -* the page) without causing legitimate guest type-change -* operations to fail. So after adding the page to the IOMMU, -* check again to make sure this is still valid. NB that the -* writable entry in the iommu is harmless until later, when -* the actual device gets assigned. -*/ -if ( !rc && !is_hvm_domain (d) && - ((page->u.inuse.type_info & PGT_type_mask) != - PGT_writable_page) ) - { - rc =iommu_unmap (d, dfn,PAGE_ORDER_4K , &flush_flags); -/* If the type changed yet again, simply force a retry. */ -if ( !rc && ((page->u.inuse.type_info & PGT_type_mask) == - PGT_writable_page) ) - rc = -ERESTART ; - } - } -if ( rc ) - { -page_list_add (page, &d->page_list); -break ; - } - } -page_list_add_tail (page, &d->arch.relmem_list); -if ( !(++n &0xff ) && !page_list_empty (&d->page_list) && -hypercall_preempt_check () ) - rc = -ERESTART ; - } - -if ( !rc ) - { -/* -* The expectation here is that generally there are many normal pages -* on relmem_list (the ones we put there) and only few being in an -* offline/broken state. The latter ones are always at the head of the -* list. Hence we first move the whole list, and then move back the -* first few entries. -*/ -page_list_move (&d->page_list, &d->arch.relmem_list); -while ( !page_list_empty (&d->page_list) && - (page =page_list_first (&d->page_list), - (page->count_info & (PGC_state|PGC_broken))) ) - { -page_list_del (page, &d->page_list); -page_list_add_tail (page, &d->arch.relmem_list); - } - } - -spin_unlock (&d->page_alloc_lock); - -if ( !rc ) -/* -* flush_flags are not tracked across hypercall pre-emption so -* assume a full flush is necessary. -*/ - rc =iommu_iotlb_flush_all ( - d, IOMMU_FLUSHF_added | IOMMU_FLUSHF_modified); - -if ( rc && rc != -ERESTART ) -iommu_teardown (d); - -return rc; -} - void __hwdom_init arch_iommu_check_autotranslated_hwdom(struct domain *d) {if ( !is_iommu_enabled (d) )xen/include/asm-arm/iommu.h +1
-1@@ -21,7 +21,7 @@ struct arch_iommu };/* Always share P2M Table between the CPU and the IOMMU */ -#define iommu_use_hap_pt (d) (has_iommu_pt(d)) +#define iommu_use_hap_pt (d) is_iommu_enabled(d)const struct iommu_ops *iommu_get_ops (void );void iommu_set_ops (const struct iommu_ops *ops);xen/include/asm-x86/iommu.h +1
-1@@ -88,7 +88,7 @@ extern const struct iommu_init_ops *iommu_init_ops;/* Are we using the domain P2M table as its IOMMU pagetable? */ #define iommu_use_hap_pt (d) \ - (hap_enabled(d) && has_iommu_pt(d) && iommu_hap_pt_share) + (hap_enabled(d) && is_iommu_enabled(d) && iommu_hap_pt_share)void iommu_update_ire_from_apic (unsigned int apic,unsigned int reg,unsigned int value);unsigned int iommu_read_apic_from_ire (unsigned int apic,unsigned int reg);xen/include/xen/iommu.h +0
-16@@ -88,15 +88,9 @@ void iommu_domain_destroy(struct domain *d);void arch_iommu_domain_destroy (struct domain *d);int arch_iommu_domain_init (struct domain *d); -int arch_iommu_populate_page_table (struct domain *d);void arch_iommu_check_autotranslated_hwdom (struct domain *d);void arch_iommu_hwdom_init (struct domain *d); -int iommu_construct (struct domain *d); - -/* Function used internally, use iommu_domain_destroy */ -void iommu_teardown (struct domain *d); -/* * The following flags are passed to map operations and passed by lookup * operations. @@ -263,13 +257,6 @@ struct iommu_ops {# define iommu_vcall iommu_call#endif -enum iommu_status -{ - IOMMU_STATUS_disabled, - IOMMU_STATUS_initializing, - IOMMU_STATUS_initialized -}; -struct domain_iommu {struct arch_iommu arch;@@ -289,9 +276,6 @@ struct domain_iommu {/* Features supported by the IOMMU */ DECLARE_BITMAP (features, IOMMU_FEAT_count); -/* Status of guest IOMMU mappings */ -enum iommu_status status; -/* * Does the guest reqire mappings to be synchonized, to maintain * the default dfn == pfn map. (See comment on dfn at the top of xen/include/xen/sched.h +0
-3@@ -966,10 +966,8 @@ static inline bool is_hwdom_pinned_vcpu(const struct vcpu *v) }#ifdef CONFIG_HAS_PASSTHROUGH -#define has_iommu_pt (d) (dom_iommu(d)->status != IOMMU_STATUS_disabled)#define need_iommu_pt_sync (d) (dom_iommu(d)->need_sync)#else -#define has_iommu_pt (d) false#define need_iommu_pt_sync (d) false#endif --
Quoted 1 line
2.20.1.2.gb21ebb671
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [Xen-devel] [PATCH v13 1/4] remove late (on-demand) construction of IOMMU page tables
In reply to Paul DurrantQuoted 32 lines · Paul Durrant
Now that there is a per-domain IOMMU-enable flag, which should be set if any device is going to be passed through, stop deferring page table construction until the assignment is done. Also don't tear down the tables again when the last device is de-assigned; defer that task until domain destruction.
This allows the
has_iommu_pt()helper andiommu_statusenumeration to be removed. Calls tohas_iommu_pt()are simply replaced by calls tois_iommu_enabled(). Remaining open-coded tests ofiommu_hap_pt_sharecan also be replaced by calls toiommu_use_hap_pt().
Thearch_iommu_populate_page_table()andiommu_construct()functions become redundant, as does the 'strict mode' dom0page_listmapping code iniommu_hwdom_init(), andiommu_teardown()can be made static is its only remaining caller,iommu_domain_destroy(), is within the same source module.All in all, about 220 lines of code are removed from the hypervisor (at the expense of some additions in the toolstack).
NOTE: This patch will cause a small amount of extra resource to be used
to accommodate IOMMU page tables that may never be used, since the per-domain IOMMU-enable flag is currently set to the value of the globaliommu_enableflag. A subsequent patch will add an option to the toolstack to allow it to be turned off if there is no intention to assign passthrough hardware to the domain.
To account for the extra resource, 'iommu_memkb' has been added todomain_build_info. This patch sets it to a value calculated based on the domain's maximum memory when the P2M sharing is either not supported or globally disabled, or zero otherwise. However, when the toolstack option mentioned above is added, it will also be zero if the per-domain IOMMU-enable flag is turned off.
Quoted 5 lines
Signed-off-by Paul Durrant <[email protected]>Reviewed-by Alexandru Isaila <[email protected]>Acked-by Razvan Cojocaru <[email protected]>Reviewed-by Jan Beulich <[email protected]>Acked-by Julien Grall <[email protected]>
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH v13 2/4 iommu: tidy up iommu_use_hap_pt() and need_iommu_pt_sync() macros
In reply to Paul DurrantThes macros really ought to live in the common xen/iommu.h header rather then being distributed amongst architecture specific iommu headers and xen/sched.h. This patch moves them there.
NOTE: Disabling 'sharept' in the command line iommu options should really
be hard error on ARM (as opposed to just being ignored), so define 'iommu_hap_pt_share' to be true for ARM (via ARM-selected CONFIG_IOMMU_FORCE_PT_SHARE).
---
Cc: Andrew Cooper <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Ian Jackson <[email protected]>
Cc: Konrad Rzeszutek Wilk <[email protected]>
Cc: Stefano Stabellini <[email protected]>
Cc: Tim Deegan <[email protected]>
Cc: Wei Liu <[email protected]>
Cc: Volodymyr Babchuk <[email protected]>
Cc: "Roger Pau Monné" <[email protected]>
Previously part of https://lists.xenproject.org/archives/html/xen-devel/2019-07/msg02267.html
v11:
- Fix #if/#elif ordering issue
v9:
- Add new Kconfig option to cause 'iommu_hap_pt_share' to be defined to true, rather than using CONFIG_ARM, as requested by Julien
- Assuming Jan's R-b stands since this is a mainly a cosmetic change directly requested by another maintainer
v7:
- Re-work the ARM handling of 'sharept' as suggested by Jan
- Make sure that need_iommu_pt_sync() always evaluates its argument
--- xen/arch/arm/Kconfig | 1 + xen/drivers/passthrough/Kconfig | 3 +++ xen/include/asm-arm/iommu.h | 3 --- xen/include/asm-x86/iommu.h | 4 ---- xen/include/xen/iommu.h | 15 ++++++++++++++- xen/include/xen/sched.h | 6 ------ 6 files changed, 18 insertions(+), 14 deletions(-)
xen/arch/arm/Kconfig +1 -0
@@ -20,6 +20,7 @@ config ARM select HAS_DEVICE_TREE select HAS_PASSTHROUGH select HAS_PDX + select IOMMU_FORCE_PT_SHARE config ARCH_DEFCONFIG string
xen/drivers/passthrough/Kconfig +3 -0
@@ -13,3 +13,6 @@ config ARM_SMMU Say Y here if your SoC includes an IOMMU device implementing the ARM SMMU architecture. endif + +config IOMMU_FORCE_PT_SHARE + bool
xen/include/asm-arm/iommu.h +0 -3
@@ -20,9 +20,6 @@ struct arch_iommuvoid *priv; }; -/* Always share P2M Table between the CPU and the IOMMU */ -#define iommu_use_hap_pt (d) is_iommu_enabled(d) -const struct iommu_ops *iommu_get_ops (void );void iommu_set_ops (const struct iommu_ops *ops);
xen/include/asm-x86/iommu.h +0 -4
@@ -86,10 +86,6 @@ struct iommu_init_ops {extern const struct iommu_init_ops *iommu_init_ops; -/* Are we using the domain P2M table as its IOMMU pagetable? */ -#define iommu_use_hap_pt (d) \ - (hap_enabled(d) && is_iommu_enabled(d) && iommu_hap_pt_share) -void iommu_update_ire_from_apic (unsigned int apic,unsigned int reg,unsigned int value);unsigned int iommu_read_apic_from_ire (unsigned int apic,unsigned int reg);int iommu_setup_hpet_msi (struct msi_desc *);
xen/include/xen/iommu.h +14 -1
@@ -56,7 +56,9 @@ extern bool_t iommu_enable, iommu_enabled;extern bool_t force_iommu, iommu_verbose, iommu_igfx;extern bool_t iommu_snoop, iommu_qinval, iommu_intremap, iommu_intpost; -#ifdef CONFIG_HVM +#if defined(CONFIG_IOMMU_FORCE_PT_SHARE ) +#define iommu_hap_pt_share true +#elif defined(CONFIG_HVM )extern bool iommu_hap_pt_share;#else #define iommu_hap_pt_share false
@@ -288,6 +290,17 @@ struct domain_iommu {#define iommu_set_feature (d, f) set_bit(f, dom_iommu(d)->features)#define iommu_clear_feature (d, f) clear_bit(f, dom_iommu(d)->features) +/* Are we using the domain P2M table as its IOMMU pagetable? */ +#define iommu_use_hap_pt (d) \ + (hap_enabled(d) && is_iommu_enabled(d) && iommu_hap_pt_share) + +/* Does the IOMMU pagetable need to be kept synchronized with the P2M */ +#ifdef CONFIG_HAS_PASSTHROUGH +#define need_iommu_pt_sync (d) (dom_iommu(d)->need_sync) +#else +#define need_iommu_pt_sync (d) ({ (void)(d); false; }) +#endif +int __must_check iommu_suspend (void );void iommu_resume (void );void iommu_crash_shutdown (void );
xen/include/xen/sched.h +0 -6
@@ -965,12 +965,6 @@ static inline bool is_hwdom_pinned_vcpu(const struct vcpu *v) cpumask_weight(v->cpu_hard_affinity) == 1); } -#ifdef CONFIG_HAS_PASSTHROUGH -#define need_iommu_pt_sync (d) (dom_iommu(d)->need_sync) -#else -#define need_iommu_pt_sync (d) false -#endif -static inline bool is_vcpu_online (const struct vcpu *v) {return !test_bit (_VPF_down, &v->pause_flags);
Signature · 7 lines
2.20.1.2.gb21ebb671 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH v13 3/4 tools/ocaml: abi check: Cope with consecutive relevant enums
In reply to Paul DurrantFrom: Ian Jackson <[email protected]>
If the end of one enum is the `type' line for the next enum, we would not notice it.
Fix this by reordering the code, and getting rid of the else: now if the "we are within an enum" branch decides that it's the end of the enum, it unsets $ei and we then immediately process the line as a "not within an enum" line - ie as the start of the next one.
v13:
- New in this version
--- tools/ocaml/libs/xc/abi-check | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-)
tools/ocaml/libs/xc/abi-check +9 -8
@@ -70,14 +70,7 @@ my $cval; $ei = undef; my $bitnum = 0; while (<OCAML_FILE>) { - if (!$ei) { - if (m{^type \s+ (\w+) \s* \= \s* $}x && $enums{$1}) { - print "// found ocaml type $1 at $o:$.\n" or die $!; - $ei = $enums{$1}; - $cval = ''; - $bitnum = 0; - } - } else { + if ($ei) { if (m{^\s+ \| \s* $ei->{OPrefix} (\w+) \s*$}x) { $cval = $1; if ($ei->{Mangle} eq 'lc') {
@@ -104,6 +97,14 @@ while (<OCAML_FILE>) { die "$_ ?"; } } + if (!$ei) { + if (m{^type \s+ (\w+) \s* \= \s* $}x && $enums{$1}) { + print "// found ocaml type $1 at $o:$.\n" or die $!; + $ei = $enums{$1}; + $cval = ''; + $bitnum = 0; + } + } } foreach $ei (values %enums) {
Signature · 7 lines
2.20.1.2.gb21ebb671 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
PATCH v13 4/4 introduce a 'passthrough' configuration option to xl.cfg...
In reply to Paul Durrant...and hence the ability to disable IOMMU mappings, and control EPT sharing.
This patch introduces a new 'libxl_passthrough' enumeration into libxl_domain_create_info. The value will be set by xl either when it parses a new 'passthrough' option in xl.cfg, or implicitly if there is passthrough hardware specified for the domain.
If the value of the passthrough configuration option is 'disabled' then the XEN_DOMCTL_CDF_iommu flag will be clear in the xen_domctl_createdomain flags, thus allowing the toolstack to control whether the domain gets IOMMU mappings or not (where previously they were globally set).
If the value of the passthrough configuration option is 'sync_pt' then a new 'iommu_opts' field in xen_domctl_createdomain will be set with the value XEN_DOMCTL_IOMMU_no_sharept. This will override the global default set in iommu_hap_pt_share, thus allowing the toolstack to control whether EPT sharing is used for the domain.
If the value of passthrough is 'enabled' then xl will choose an appropriate default according to the type of domain and hardware support.
NOTE: The 'iommu_memkb' overhead in libxl_domain_build_info will now only
be set if passthrough is 'sync_pt' (or xl has chosen this mode as a default).
---
Cc: Ian Jackson <[email protected]>
Cc: Wei Liu <[email protected]>
Cc: Andrew Cooper <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Konrad Rzeszutek Wilk <[email protected]>
Cc: Stefano Stabellini <[email protected]>
Cc: Tim Deegan <[email protected]>
Cc: Anthony PERARD <[email protected]>
Cc: David Scott <[email protected]>
Cc: Volodymyr Babchuk <[email protected]>
Cc: "Roger Pau Monné" <[email protected]>
Previously part of series https://lists.xenproject.org/archives/html/xen-devel/2019-07/msg02267.html
v13:
- Small documentation adjustments requested by George and Julien
- Fix default passthrough option
v12:
- Set passthrough default in libxl
v11:
- Fixed abi-check runes
v10:
- Added abi-check runes
v9:
- Added the passthrough='enabled' option to xl
- One cosmetic change in xen
- Assume Jan's R-b stands since non-cosmetic changes are only in the toolstack
v7:
- Added missing breaks
- Added missing ocaml binding changes
v6:
- Remove the libxl_physinfo() call since it's usefulness is limited to x86
v5:
- Expand xen_domctl_createdomain flags field and hence bump interface version
- Fix spelling mistakes in context line
--- docs/man/xl.cfg.5.pod.in | 57 ++++++++++ tools/libxl/libxl.h | 9 ++ tools/libxl/libxl_create.c | 30 ++++- tools/libxl/libxl_types.idl | 8 ++ tools/ocaml/libs/xc/xenctrl.ml | 4 + tools/ocaml/libs/xc/xenctrl.mli | 5 + tools/ocaml/libs/xc/xenctrl_stubs.c | 17 ++- tools/xl/xl_parse.c | 164 +++++++++++++++++++--------- xen/arch/arm/domain.c | 10 +- xen/arch/x86/domain.c | 2 +- xen/common/domain.c | 7 ++ xen/common/domctl.c | 13 --- xen/drivers/passthrough/iommu.c | 13 ++- xen/include/public/domctl.h | 10 +- xen/include/xen/iommu.h | 15 ++- 15 files changed, 277 insertions(+), 87 deletions(-)
docs/man/xl.cfg.5.pod.in +57 -0
@@ -605,6 +605,63 @@ option should only be used with a trusted device tree. Note that the partial device tree should avoid using the phandle 65000 which is reserved by the toolstack. +=item B<passthrough="STRING"> + +Specify whether IOMMU mappings are enabled for the domain and hence whether +it will be enabled for passthrough hardware. Valid values for this option +are: + +=over 4 + +=item B<disabled> + +IOMMU mappings are disabled for the domain and so hardware may not be +passed through. + +This option is the default if no passthrough hardware is specified in the +domain's configuration. + +=item B<enabled> + +This option enables IOMMU mappings and selects an appropriate default +operating mode (see below for details of the operating modes). For HVM/PVH +domains running on platforms where the option is available, this is +equivalent to B<share_pt>. Otherwise, and also for PV domains, this +option is equivalent to B<sync_pt>. + +This option is the default if passthrough hardware is specified in the +domain's configuration. + +=item B<sync_pt> + +This option means that IOMMU mappings will be synchronized with the +domain's P2M table as follows: + +For a PV domain, all writable pages assigned to the domain are identity +mapped by MFN in the IOMMU page table. Thus a device driver running in the +domain may program passthrough hardware for DMA using MFN values +(i.e. host/machine frame numbers) looked up in its P2M. + +For an HVM/PVH domain, all non-foreign RAM pages present in its P2M will be +mapped by GFN in the IOMMU page table. Thus a device driver running in the +domain may program passthrough hardware using GFN values (i.e. guest +physical frame numbers) without any further translation. + +This option is not currently available on Arm. + +=item B<share_pt> + +This option is unavailable for a PV domain. For an HVM/PVH domain, this +option means that the IOMMU will be programmed to directly reference the +domain's P2M table as its page table. From the point of view of a device +driver running in the domain this is functionally equivalent to B<sync_pt> +but places less load on the hypervisor and so should generally be selected +in preference. However, the availability of this option is hardware +specific. If B<xl info> reports B<virt_caps> containing +B<iommu_hap_pt_share> then this option may be used. + +=back + =back =head2 Devices
tools/libxl/libxl.h +9 -0
@@ -415,6 +415,15 @@*/ #define LIBXL_HAVE_BUILDINFO_IOMMU_MEMKB 1 +/* +* LIBXL_HAVE_CREATEINFO_PASSTHROUGH indicates that +* libxl_domain_create_info has a passthrough field (which is a +* libxl_passthrough enumeration) that indicates whether device pass- +* through is enabled for the domain and, if so, whether the IOMMU and +* HAP page tables may be shared or not. +*/ +#define LIBXL_HAVE_CREATEINFO_PASSTHROUGH 1 +/* * libxl ABI compatibility *
tools/libxl/libxl_create.c +24 -6
@@ -30,6 +30,13 @@int libxl__domain_create_info_setdefault (libxl__gc *gc,libxl_domain_create_info *c_info) { +libxl_physinfo info; +int rc; + + rc =libxl_get_physinfo (CTX , &info); +if (rc) +return rc; +if (!c_info->type) {LOG (ERROR ,"domain type unspecified" );return ERROR_INVAL ;
@@ -38,12 +45,6 @@ int libxl__domain_create_info_setdefault(libxl__gc *gc,libxl__arch_domain_create_info_setdefault (gc, c_info);if (c_info->type !=LIBXL_DOMAIN_TYPE_PV ) { -libxl_physinfo info; -int rc =libxl_get_physinfo (CTX , &info); - -if (rc) -return rc; -if (info.cap_hap)libxl_defbool_setdefault (&c_info->hap, true);else if (info.cap_shadow)
@@ -62,6 +63,13 @@ int libxl__domain_create_info_setdefault(libxl__gc *gc,if (!c_info->ssidref) c_info->ssidref =SECINITSID_DOMU ; +if (info.cap_hvm_directio && + (c_info->passthrough ==LIBXL_PASSTHROUGH_ENABLED )) { + c_info->passthrough = ((c_info->type ==LIBXL_DOMAIN_TYPE_PV ) || + !info.cap_iommu_hap_pt_share) ? +LIBXL_PASSTHROUGH_SYNC_PT :LIBXL_PASSTHROUGH_SHARE_PT ; + } +return 0 ; }
@@ -578,6 +586,16 @@ int libxl__domain_make(libxl__gc *gc, libxl_domain_config *d_config,libxl_defbool_val (info->oos) ?0 : XEN_DOMCTL_CDF_oos_off; } +assert (info->passthrough !=LIBXL_PASSTHROUGH_ENABLED ); +LOG (DETAIL ,"passthrough: %s" , +libxl_passthrough_to_string (info->passthrough)); + +if (info->passthrough !=LIBXL_PASSTHROUGH_DISABLED ) + create.flags |= XEN_DOMCTL_CDF_iommu; + +if (info->passthrough ==LIBXL_PASSTHROUGH_SYNC_PT ) + create.iommu_opts |= XEN_DOMCTL_IOMMU_no_sharept; +/* Ultimately, handle is an array of 16 uint8_t, same as uuid */ libxl_uuid_copy (ctx, (libxl_uuid *)&create.handle, &info->uuid);
tools/libxl/libxl_types.idl +8 -0
@@ -263,6 +263,13 @@ libxl_vkb_backend = Enumeration("vkb_backend", [ (2, "LINUX") ]) +libxl_passthrough = Enumeration("passthrough", [ + (0, "enabled"), + (1, "disabled"), + (2, "sync_pt"), + (3, "share_pt"), + ]) + # # Complex libxl types #
@@ -408,6 +415,7 @@ libxl_domain_create_info = Struct("domain_create_info",[ ("pool_name", string), ("run_hotplug_scripts",libxl_defbool), ("driver_domain",libxl_defbool), + ("passthrough", libxl_passthrough), ], dir=DIR_IN) libxl_domain_restore_params = Struct("domain_restore_params", [
tools/ocaml/libs/xc/xenctrl.ml +4 -0
@@ -65,11 +65,15 @@ type domain_create_flag = | CDF_XS_DOMAIN | CDF_IOMMU +type domain_create_iommu_opts = + | IOMMU_NO_SHAREPT + type domctl_create_config = { ssidref: int32; handle: string; flags: domain_create_flag list; + iommu_opts: domain_create_iommu_opts list; max_vcpus: int; max_evtchn_port: int; max_grant_frames: int;
tools/ocaml/libs/xc/xenctrl.mli +5 -0
@@ -57,10 +57,15 @@ type domain_create_flag = | CDF_OOS_OFF | CDF_XS_DOMAIN | CDF_IOMMU + +type domain_create_iommu_opts = + | IOMMU_NO_SHAREPT + type domctl_create_config = { ssidref: int32; handle: string; flags: domain_create_flag list; + iommu_opts: domain_create_iommu_opts list; max_vcpus: int; max_evtchn_port: int; max_grant_frames: int;
tools/ocaml/libs/xc/xenctrl_stubs.c +12 -5
@@ -190,11 +190,12 @@ CAMLprim value stub_xc_domain_create(value xch, value config)#define VAL_SSIDREF Field(config, 0)#define VAL_HANDLE Field(config, 1)#define VAL_FLAGS Field(config, 2) -#define VAL_MAX_VCPUS Field(config, 3) -#define VAL_MAX_EVTCHN_PORT Field(config, 4) -#define VAL_MAX_GRANT_FRAMES Field(config, 5) -#define VAL_MAX_MAPTRACK_FRAMES Field(config, 6) -#define VAL_ARCH Field(config, 7) +#define VAL_IOMMU_OPTS Field(config, 3) +#define VAL_MAX_VCPUS Field(config, 4) +#define VAL_MAX_EVTCHN_PORT Field(config, 5) +#define VAL_MAX_GRANT_FRAMES Field(config, 6) +#define VAL_MAX_MAPTRACK_FRAMES Field(config, 7) +#define VAL_ARCH Field(config, 8)uint32_t domid =0 ;int result;
@@ -213,6 +214,11 @@ CAMLprim value stub_xc_domain_create(value xch, value config)/* ! XEN_DOMCTL_CDF_ XEN_DOMCTL_CDF_MAX max */ (VAL_FLAGS ); + cfg.iommu_opts =ocaml_list_to_c_bitmap +/* ! domain_create_iommu_opts IOMMU_ lc */ +/* ! XEN_DOMCTL_IOMMU_ XEN_DOMCTL_IOMMU_MAX max */ + (VAL_IOMMU_OPTS ); + arch_domconfig =Field (VAL_ARCH ,0 );switch (Tag_val (VAL_ARCH ) ) {
@@ -247,6 +253,7 @@ CAMLprim value stub_xc_domain_create(value xch, value config)#undef VAL_MAX_GRANT_FRAMES#undef VAL_MAX_EVTCHN_PORT#undef VAL_MAX_VCPUS +#undef VAL_IOMMU_OPTS#undef VAL_FLAGS#undef VAL_HANDLE#undef VAL_SSIDREF
tools/xl/xl_parse.c +111 -53
182 lines not on this page · raw message
xen/arch/arm/domain.c +9 -1
@@ -617,6 +617,14 @@ int arch_sanitise_domain_config(struct xen_domctl_createdomain *config)return -EINVAL ; } +/* The P2M table must always be shared between the CPU and the IOMMU */ +if ( config->iommu_opts & XEN_DOMCTL_IOMMU_no_sharept ) + { +dprintk (XENLOG_INFO , +"Unsupported iommu option: XEN_DOMCTL_IOMMU_no_sharept\n" ); +return -EINVAL ; + } +/* Fill in the native GIC version, passed back to the toolstack. */ if ( config->arch.gic_version ==XEN_DOMCTL_CONFIG_GIC_NATIVE ) {
@@ -677,7 +685,7 @@ int arch_domain_create(struct domain *d,ASSERT (config !=NULL );/* p2m_init relies on some value initialized by the IOMMU subsystem */ -if ( (rc =iommu_domain_init (d)) !=0 ) +if ( (rc =iommu_domain_init (d, config->iommu_opts)) !=0 ) goto fail;if ( (rc =p2m_init (d)) !=0 )
xen/arch/x86/domain.c +1 -1
@@ -611,7 +611,7 @@ int arch_domain_create(struct domain *d,if ( (rc =init_domain_irq_mapping (d)) !=0 ) goto fail; -if ( (rc =iommu_domain_init (d)) !=0 ) +if ( (rc =iommu_domain_init (d, config->iommu_opts)) !=0 ) goto fail;psr_domain_init (d);
xen/common/domain.c +7 -0
@@ -308,6 +308,13 @@ static int sanitise_domain_config(struct xen_domctl_createdomain *config)return -EINVAL ; } +if ( !(config->flags & XEN_DOMCTL_CDF_iommu) && config->iommu_opts ) + { +dprintk (XENLOG_INFO , +"IOMMU options specified but IOMMU not enabled\n" ); +return -EINVAL ; + } +if ( config->max_vcpus <1 ) {dprintk (XENLOG_INFO ,"No vCPUS\n" );
xen/common/domctl.c +0 -13
@@ -515,19 +515,6 @@ long do_domctl(XEN_GUEST_HANDLE_PARAM(xen_domctl_t) u_domctl) rover = dom; } -/* -* For now, make sure the createdomain IOMMU flag is set if the -* IOMMU is enabled. When the flag comes under toolstack control -* this can go away. -*/ -if ( op->u.createdomain.flags & XEN_DOMCTL_CDF_iommu ) - { -ASSERT_UNREACHABLE (); -return -EINVAL ; - } -if ( iommu_enabled ) - op->u.createdomain.flags |= XEN_DOMCTL_CDF_iommu; - d =domain_create (dom, &op->u.createdomain, false);if (IS_ERR (d) ) {
xen/drivers/passthrough/iommu.c +12 -1
@@ -163,7 +163,7 @@ static void __hwdom_init check_hwdom_reqs(struct domain *d) iommu_hwdom_strict = true; } -int iommu_domain_init (struct domain *d) +int iommu_domain_init (struct domain *d,unsigned int opts) {struct domain_iommu *hd =dom_iommu (d);int ret =0 ;
@@ -187,6 +187,15 @@ int iommu_domain_init(struct domain *d)if (is_hardware_domain (d) )check_hwdom_reqs (d);/* may modify iommu_hwdom_strict */ +/* +* Use shared page tables for HAP and IOMMU if the global option +* is enabled (from which we can infer the h/w is capable) and +* the domain options do not disallow it. HAP must, of course, also +* be enabled. +*/ + hd->hap_pt_share =hap_enabled (d) && iommu_hap_pt_share && + !(opts & XEN_DOMCTL_IOMMU_no_sharept); +/* * NB: 'relaxed' h/w domains don't need the IOMMU mappings to be kept * in-sync with their assigned pages because all host RAM will be
@@ -195,6 +204,8 @@ int iommu_domain_init(struct domain *d)if ( !is_hardware_domain (d) || iommu_hwdom_strict ) hd->need_sync = !iommu_use_hap_pt (d); +ASSERT (!(hd->need_sync && hd->hap_pt_share)); +return 0 ; }
xen/include/public/domctl.h +9 -1
@@ -38,7 +38,7 @@#include "hvm/save.h" #include "memory.h" -#define XEN_DOMCTL_INTERFACE_VERSION 0x00000011 +#define XEN_DOMCTL_INTERFACE_VERSION 0x00000012/* * NB. xen_domctl.domain is an IN/OUT parameter for this operation.
@@ -73,6 +73,14 @@ struct xen_domctl_createdomain {uint32_t flags; +#define _XEN_DOMCTL_IOMMU_no_sharept 0 +#define XEN_DOMCTL_IOMMU_no_sharept (1U << _XEN_DOMCTL_IOMMU_no_sharept) + +/* Max XEN_DOMCTL_IOMMU_* constant. Used for ABI checking. */ +#define XEN_DOMCTL_IOMMU_MAX XEN_DOMCTL_IOMMU_no_sharept + +uint32_t iommu_opts; +/* * Various domain limits, which impact the quantity of resources (global * mapping space, xenheap, etc) a guest may consume.
xen/include/xen/iommu.h +9 -6
@@ -84,7 +84,7 @@ extern unsigned int iommu_dev_iotlb_timeout;int iommu_setup (void );int iommu_hardware_setup (void ); -int iommu_domain_init (struct domain *d); +int iommu_domain_init (struct domain *d,unsigned int opts);void iommu_hwdom_init (struct domain *d);void iommu_domain_destroy (struct domain *d);
@@ -278,10 +278,14 @@ struct domain_iommu {/* Features supported by the IOMMU */ DECLARE_BITMAP (features, IOMMU_FEAT_count); +/* Does the guest share HAP mapping with the IOMMU? */ +bool hap_pt_share; +/* -* Does the guest reqire mappings to be synchonized, to maintain -* the default dfn == pfn map. (See comment on dfn at the top of -* include/xen/mm.h). +* Does the guest require mappings to be synchronized, to maintain +* the default dfn == pfn map? (See comment on dfn at the top of +* include/xen/mm.h). Note that hap_pt_share == false does not +* necessarily imply this is true. */ bool need_sync; };
@@ -291,8 +295,7 @@ struct domain_iommu {#define iommu_clear_feature (d, f) clear_bit(f, dom_iommu(d)->features)/* Are we using the domain P2M table as its IOMMU pagetable? */ -#define iommu_use_hap_pt (d) \ - (hap_enabled(d) && is_iommu_enabled(d) && iommu_hap_pt_share) +#define iommu_use_hap_pt (d) (dom_iommu(d)->hap_pt_share)/* Does the IOMMU pagetable need to be kept synchronized with the P2M */ #ifdef CONFIG_HAS_PASSTHROUGH
Signature · 7 lines
2.20.1.2.gb21ebb671 _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Re: [Xen-devel] [PATCH v13 4/4] introduce a 'passthrough' configuration option to xl.cfg...
Quoted 22 lines · Paul Durrant
...and hence the ability to disable IOMMU mappings, and control EPT sharing.
This patch introduces a new '
libxl_passthrough' enumeration intolibxl_domain_create_info. The value will be set by xl either when it parses a new 'passthrough' option in xl.cfg, or implicitly if there is passthrough hardware specified for the domain.If the value of the passthrough configuration option is 'disabled' then the
XEN_DOMCTL_CDF_iommuflag will be clear in thexen_domctl_createdomainflags, thus allowing the toolstack to control whether the domain gets IOMMU mappings or not (where previously they were globally set).If the value of the passthrough configuration option is '
sync_pt' then a new 'iommu_opts' field inxen_domctl_createdomainwill be set with the valueXEN_DOMCTL_IOMMU_no_sharept. This will override the global default set iniommu_hap_pt_share, thus allowing the toolstack to control whether EPT sharing is used for the domain.If the value of passthrough is 'enabled' then xl will choose an appropriate default according to the type of domain and hardware support.
Quoted 8 lines
NOTE: The '
iommu_memkb' overhead inlibxl_domain_build_infowill now only
be set if passthrough is 'sync_pt' (or xl has chosen this mode as a default).Signed-off-by Paul Durrant <[email protected]>Reviewed-by Jan Beulich <[email protected]>Acked-by Christian Lindig <[email protected]>Acked-by Julien Grall <[email protected]>
Thanks,
Signature · 6 lines
Anthony PERARD _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Ping? I don't think this series is awaiting any more acks.
Paul
Quoted 64 lines
-----Original Message-----
From: Paul Durrant <[email protected]>
Sent: 18 September 2019 11:41
To: [email protected]
Cc: Paul Durrant <[email protected]>; Andrew Cooper <[email protected]>; Anthony Perard <[email protected]>; David Scott <[email protected]>; George Dunlap <[email protected]>; Ian Jackson <[email protected]>; Konrad Rzeszutek Wilk <[email protected]>; Petre Pircalabu <[email protected]>; Roger Pau Monne <[email protected]>; Stefano Stabellini <[email protected]>; Tamas K Lengyel <[email protected]>; Tim (Xen.org) <[email protected]>; Volodymyr Babchuk <[email protected]>; Wei Liu <[email protected]>
Subject: [PATCH v13 0/4] add per-domain IOMMU controlThese are the remaining uncommitted patches from my previous series:
https://lists.xenproject.org/archives/html/xen-devel/2019-09/msg01208.html
The only patch that has been revised is patch #4 (previously patch #6).
Ian Jackson (1):
tools/ocaml: abi check: Cope with consecutive relevant enumsPaul Durrant (3):
remove late (on-demand) construction of IOMMU page tables
iommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macros introduce a 'passthrough' configuration option to xl.cfg...docs/man/xl.cfg.5.pod.in | 57 +++++++++ tools/libxl/libxl.h | 16 +++ tools/libxl/libxl_create.c | 30 ++++- tools/libxl/libxl_mem.c | 6 +- tools/libxl/libxl_types.idl | 9 ++ tools/libxl/libxl_utils.c | 15 +++ tools/libxl/libxl_utils.h | 1 + tools/ocaml/libs/xc/abi-check | 17 +-- tools/ocaml/libs/xc/xenctrl.ml | 4 + tools/ocaml/libs/xc/xenctrl.mli | 5 + tools/ocaml/libs/xc/xenctrl_stubs.c | 17 ++- tools/xl/xl_parse.c | 178 ++++++++++++++++++-------- xen/arch/arm/Kconfig | 1 + xen/arch/arm/domain.c | 10 +- xen/arch/arm/p2m.c | 2 +- xen/arch/x86/dom0_build.c | 2 +- xen/arch/x86/domain.c | 2 +- xen/arch/x86/hvm/mtrr.c | 5 +- xen/arch/x86/mm/mem_sharing.c | 2 +- xen/arch/x86/mm/p2m.c | 4 +- xen/arch/x86/mm/paging.c | 2 +- xen/arch/x86/x86_64/mm.c | 2 +- xen/common/domain.c | 7 + xen/common/domctl.c | 13 -- xen/common/memory.c | 4 +- xen/common/vm_event.c | 2 +- xen/drivers/passthrough/Kconfig | 3 + xen/drivers/passthrough/device_tree.c | 11 -- xen/drivers/passthrough/iommu.c | 147 ++++++--------------- xen/drivers/passthrough/pci.c | 12 -- xen/drivers/passthrough/vtd/iommu.c | 10 +- xen/drivers/passthrough/x86/iommu.c | 97 -------------- xen/include/asm-arm/iommu.h | 3 - xen/include/asm-x86/iommu.h | 4 - xen/include/public/domctl.h | 10 +- xen/include/xen/iommu.h | 40 +++--- xen/include/xen/sched.h | 8 -- 37 files changed, 388 insertions(+), 370 deletions(-)
Quoted 16 lines
---
Cc: Andrew Cooper <[email protected]>
Cc: Anthony PERARD <[email protected]>
Cc: David Scott <[email protected]>
Cc: George Dunlap <[email protected]>
Cc: Ian Jackson <[email protected]>
Cc: Konrad Rzeszutek Wilk <[email protected]>
Cc: Petre Pircalabu <[email protected]>
Cc: "Roger Pau Monné" <[email protected]>
Cc: Stefano Stabellini <[email protected]>
Cc: Tamas K Lengyel <[email protected]>
Cc: Tim Deegan <[email protected]>
Cc: Volodymyr Babchuk <[email protected]>
Cc: Wei Liu <[email protected]>
--
2.20.1.2.gb21ebb671
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 1 line · Paul Durrant
Ping? I don't think this series is awaiting any more acks.
I did enumerate the other day on irc what is still missing according to my records. Would you mind pointing me at a libxl maintainer ack for patch 1? I think I've managed to spot the missing one for patch 3, which hung off a "REPOST" thread. And I realize I was wrongly looking for a tool stack maintainer ack for patch 4, when a libxl one is sufficient and already there.
Jan
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 13 lines
-----Original Message-----
From: Jan Beulich <[email protected]>
Sent: 25 September 2019 09:51
To: Paul Durrant <[email protected]>
Cc: [email protected]; Petre Pircalabu <[email protected]>; Andrew Cooper <[email protected]>; Anthony Perard <[email protected]>; George Dunlap <[email protected]>; Ian Jackson <[email protected]>; Roger Pau Monne <[email protected]>; Volodymyr Babchuk <[email protected]>; Stefano Stabellini <[email protected]>; KonradRzeszutek Wilk <[email protected]>; David Scott <[email protected]>; Tamas K Lengyel <[email protected]>; Tim (Xen.org) <[email protected]>; Wei Liu <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control
Quoted 5 lines
Quoted 1 line · Paul Durrant
Ping? I don't think this series is awaiting any more acks.
I did enumerate the other day on irc what is still missing according to my records.
Oh, I must have dropped off an missed that. Connectivity can be somewhat flakey here in the office :-(
Quoted 5 lines
Would you mind pointing me at a libxl maintainer ack
for patch 1? I think I've managed to spot the missing one for patch 3, which hung off a "REPOST" thread. And I realize I was wrongly looking for a tool stack maintainer ack for patch 4, when a libxl one is sufficient and already there.
I thought I'd had an toolstack ack on patch #1 but now that I look again, you're right... it's still missing. I'll chase that one.
Thanks,
Paul
Quoted 2 lines
Jan
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ce iommu: tidy up iommu_use_hap_pt() and need_iommu_pt_sync() macros
triggers ASSERT_UNREACHABLE on Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: # CONFIG_IPMMU_VMSA is not set) .
So, iommu_setup() calls clear_iommu_hap_pt_share() with iommu_hap_pt_share being set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.
...
(XEN) Assertion 'unreachable' failed at ...ild-workspace/build/xen/xen/include/xen/iommu.h:72 (XEN) ----[ Xen-4.13-unstable arm64 debug=y Not tainted ]----
(XEN) CPU: 0
(XEN) PC: 00000000002b3ae0 iommu_setup+0xa0/0x18c
(XEN) LR: 00000000002b3a8c
(XEN) SP: 00000000002f7dc0
(XEN) CPSR: a0000249 MODE:64-bit EL2h (Hypervisor, handler)
(XEN) X0: 00000000002a7000 X1: 0000000000000000 X2: 2c736173656e6572 (XEN) X3: 0000000000000002 X4: 0000000000000001 X5: 0000000000000000 (XEN) X6: 0000000000000080 X7: 2b726072646d6471 X8: 7f7f7f7f7f7f7f7f (XEN) X9: ff65685e6c6f7275 X10: 7f7f7f7f7f7f7f7f X11: 0101010101010101 (XEN) X12: 0000000000000038 X13: 0000000000280910 X14: 0000000000000020 (XEN) X15: 0000000000000000 X16: 00000000002a7000 X17: 00000000002a7000 (XEN) X18: 00000000002a7000 X19: 0000000000000000 X20: 00000000ffffffed (XEN) X21: 00000000002a6380 X22: 0000000000335430 X23: 0000000000000002 (XEN) X24: 000000000029b1f0 X25: 00000000002d83d0 X26: 0000000048000000 (XEN) X27: 00000000c0000000 X28: 0000000000000001 FP: 00000000002f7dc0 (XEN)
(XEN) VTCR_EL2: 80000000
(XEN) VTTBR_EL2: 0000000000000000
(XEN)
(XEN) SCTLR_EL2: 30cd183d
(XEN) HCR_EL2: 0000000000000038
(XEN) TTBR0_EL2: 00000000781b4000
(XEN)
(XEN) ESR_EL2: f2000001
(XEN) HPFAR_EL2: 0000000000000000
(XEN) FAR_EL2: 0000000000000000
(XEN)
(XEN) Xen stack trace from sp=00000000002f7dc0:
(XEN) 00000000002f7de0 00000000002bdd94 0000000000000002 0000000000000002 (XEN) 00000000bfe0b660 00000000002001b4 0000000078080000 0000000077e80000 (XEN) 0000000048000000 0000000000000000 0000000000400000 0000000000000003 (XEN) 0000000000000001 0000000000000000 0000000078080000 0000000048080040 (XEN) 0000000000000000 000000000000f080 0000000048000000 0000000078000000 (XEN) 00000000002d83c0 00000000002aa440 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000300000000 0000000000000000 00000040ffffffff (XEN) 0000000000000400 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) Xen call trace:
(XEN) [<00000000002b3ae0>] iommu_setup+0xa0/0x18c (PC)
(XEN) [<00000000002b3a8c>] iommu_setup+0x4c/0x18c (LR)
(XEN) [<00000000002bdd94>] start_xen+0xaa0/0xc7c
(XEN) [<00000000002001b4>] arm64/head.o#primary_switched+0xc/0x2c
(XEN)
(XEN)
(XEN) ****************************************
(XEN) Panic on CPU 0:
(XEN) Assertion 'unreachable' failed at ...ild-workspace/build/xen/xen/include/xen/iommu.h:72 (XEN) ****************************************
(XEN)
Signature · 9 lines
Regards, Oleksandr Tyshchenko _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 19 lines
-----Original Message-----
From: Oleksandr <[email protected]>
Sent: 25 September 2019 16:50
To: Paul Durrant <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; [email protected]; Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien Grall <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control[CC Julien]
Hi Paul
Quoted 7 lines
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macrostriggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .
Oh, I'm sure I was told this was not a possibility for ARM, which is why iommu_hap_pt_share ended up being #defined. Seems I was misled, in which case ARM ought to be have a more dynamic config. as with x86.
Paul
Quoted 72 lines
So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT....
(XEN) Assertion 'unreachable' failed at
...ild-workspace/build/xen/xen/include/xen/iommu.h:72
(XEN) ----[ Xen-4.13-unstable arm64 debug=y Not tainted ]---- (XEN) CPU: 0
(XEN) PC:00000000002b3ae0iommu_setup+0xa0/0x18c
(XEN) LR:00000000002b3a8c
(XEN) SP:00000000002f7dc0
(XEN) CPSR:a0000249MODE:64-bit EL2h (Hypervisor, handler) (XEN) X0: 00000000002a7000 X1: 0000000000000000 X2:2c736173656e6572(XEN) X3: 0000000000000002 X4: 0000000000000001 X5: 0000000000000000 (XEN) X6: 0000000000000080 X7: 2b726072646d6471 X8:7f7f7f7f7f7f7f7f(XEN) X9:ff65685e6c6f7275X10:7f7f7f7f7f7f7f7fX11: 0101010101010101 (XEN) X12: 0000000000000038 X13: 0000000000280910 X14: 0000000000000020 (XEN) X15: 0000000000000000 X16:00000000002a7000X17:00000000002a7000(XEN) X18:00000000002a7000X19: 0000000000000000 X20:00000000ffffffed(XEN) X21:00000000002a6380X22: 0000000000335430 X23: 0000000000000002 (XEN) X24:000000000029b1f0X25:00000000002d83d0X26: 0000000048000000 (XEN) X27:00000000c0000000X28: 0000000000000001 FP:00000000002f7dc0(XEN)
(XEN)VTCR_EL2: 80000000
(XEN)VTTBR_EL2: 0000000000000000
(XEN)
(XEN)SCTLR_EL2:30cd183d
(XEN)HCR_EL2: 0000000000000038
(XEN)TTBR0_EL2:00000000781b4000
(XEN)
(XEN)ESR_EL2:f2000001
(XEN)HPFAR_EL2: 0000000000000000
(XEN)FAR_EL2: 0000000000000000
(XEN)
(XEN) Xen stack trace from sp=00000000002f7dc0:
(XEN)00000000002f7de000000000002bdd940000000000000002 0000000000000002 (XEN)00000000bfe0b66000000000002001b400000000780800000000000077e80000(XEN) 0000000048000000 0000000000000000 0000000000400000 0000000000000003 (XEN) 0000000000000001 0000000000000000 0000000078080000 0000000048080040 (XEN) 0000000000000000000000000000f0800000000048000000 0000000078000000 (XEN)00000000002d83c000000000002aa4400000000000000000 0000000000000000 (XEN) 0000000000000000 0000000300000000 000000000000000000000040ffffffff(XEN) 0000000000000400 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) Xen call trace:
(XEN) [<00000000002b3ae0>]iommu_setup+0xa0/0x18c(PC)
(XEN) [<00000000002b3a8c>]iommu_setup+0x4c/0x18c(LR)
(XEN) [<00000000002bdd94>]start_xen+0xaa0/0xc7c
(XEN) [<00000000002001b4>] arm64/head.o#primary_switched+0xc/0x2c (XEN)
(XEN)
(XEN) ****************************************
(XEN) Panic on CPU 0:
(XEN) Assertion 'unreachable' failed at...ild-workspace/build/xen/xen/include/xen/iommu.h:72
(XEN) ****************************************
(XEN)
Quoted 4 lines
--
Regards,Oleksandr Tyshchenko
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 24 lines
-----Original Message-----
From: Oleksandr <[email protected]>
Sent: 25 September 2019 16:50
To: Paul Durrant <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; [email protected]; Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien Grall <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macros
Quoted 6 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.
BTW, I assume disabling the iommu on the xen command like will work round the issue?
Paul
Quoted 68 lines
...
(XEN) Assertion 'unreachable' failed at
...ild-workspace/build/xen/xen/include/xen/iommu.h:72
(XEN) ----[ Xen-4.13-unstable arm64 debug=y Not tainted ]---- (XEN) CPU: 0
(XEN) PC:00000000002b3ae0iommu_setup+0xa0/0x18c
(XEN) LR:00000000002b3a8c
(XEN) SP:00000000002f7dc0
(XEN) CPSR:a0000249MODE:64-bit EL2h (Hypervisor, handler) (XEN) X0: 00000000002a7000 X1: 0000000000000000 X2:2c736173656e6572(XEN) X3: 0000000000000002 X4: 0000000000000001 X5: 0000000000000000 (XEN) X6: 0000000000000080 X7: 2b726072646d6471 X8:7f7f7f7f7f7f7f7f(XEN) X9:ff65685e6c6f7275X10:7f7f7f7f7f7f7f7fX11: 0101010101010101 (XEN) X12: 0000000000000038 X13: 0000000000280910 X14: 0000000000000020 (XEN) X15: 0000000000000000 X16:00000000002a7000X17:00000000002a7000(XEN) X18:00000000002a7000X19: 0000000000000000 X20:00000000ffffffed(XEN) X21:00000000002a6380X22: 0000000000335430 X23: 0000000000000002 (XEN) X24:000000000029b1f0X25:00000000002d83d0X26: 0000000048000000 (XEN) X27:00000000c0000000X28: 0000000000000001 FP:00000000002f7dc0(XEN)
(XEN)VTCR_EL2: 80000000
(XEN)VTTBR_EL2: 0000000000000000
(XEN)
(XEN)SCTLR_EL2:30cd183d
(XEN)HCR_EL2: 0000000000000038
(XEN)TTBR0_EL2:00000000781b4000
(XEN)
(XEN)ESR_EL2:f2000001
(XEN)HPFAR_EL2: 0000000000000000
(XEN)FAR_EL2: 0000000000000000
(XEN)
(XEN) Xen stack trace from sp=00000000002f7dc0:
(XEN)00000000002f7de000000000002bdd940000000000000002 0000000000000002 (XEN)00000000bfe0b66000000000002001b400000000780800000000000077e80000(XEN) 0000000048000000 0000000000000000 0000000000400000 0000000000000003 (XEN) 0000000000000001 0000000000000000 0000000078080000 0000000048080040 (XEN) 0000000000000000000000000000f0800000000048000000 0000000078000000 (XEN)00000000002d83c000000000002aa4400000000000000000 0000000000000000 (XEN) 0000000000000000 0000000300000000 000000000000000000000040ffffffff(XEN) 0000000000000400 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) Xen call trace:
(XEN) [<00000000002b3ae0>]iommu_setup+0xa0/0x18c(PC)
(XEN) [<00000000002b3a8c>]iommu_setup+0x4c/0x18c(LR)
(XEN) [<00000000002bdd94>]start_xen+0xaa0/0xc7c
(XEN) [<00000000002001b4>] arm64/head.o#primary_switched+0xc/0x2c (XEN)
(XEN)
(XEN) ****************************************
(XEN) Panic on CPU 0:
(XEN) Assertion 'unreachable' failed at...ild-workspace/build/xen/xen/include/xen/iommu.h:72
(XEN) ****************************************
(XEN)
Quoted 4 lines
--
Regards,Oleksandr Tyshchenko
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 40 lines
-----Original Message-----
From: Paul Durrant
Sent: 25 September 2019 17:04
To: 'Oleksandr' <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; [email protected]; Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien Grall <[email protected]>
Subject: RE: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU controlQuoted 5 lines
-----Original Message-----
From: Oleksandr <[email protected]>
Sent: 25 September 2019 16:50
To: Paul Durrant <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; WeiLiu
Quoted 2 lines
<[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George
Dunlap
Quoted 2 lines
<[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; xen-
Quoted 15 lines
Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien Grall <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macros
Quoted 8 lines
Quoted 6 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.BTW, I assume disabling the iommu on the xen command like will work round the issue?
Actually, no, that would be no good either.
Paul
Quoted 75 lines
Paul
Quoted 72 lines
...
(XEN) Assertion 'unreachable' failed at
...ild-workspace/build/xen/xen/include/xen/iommu.h:72
(XEN) ----[ Xen-4.13-unstable arm64 debug=y Not tainted ]---- (XEN) CPU: 0
(XEN) PC:00000000002b3ae0iommu_setup+0xa0/0x18c
(XEN) LR:00000000002b3a8c
(XEN) SP:00000000002f7dc0
(XEN) CPSR:a0000249MODE:64-bit EL2h (Hypervisor, handler) (XEN) X0: 00000000002a7000 X1: 0000000000000000 X2:2c736173656e6572(XEN) X3: 0000000000000002 X4: 0000000000000001 X5: 0000000000000000 (XEN) X6: 0000000000000080 X7: 2b726072646d6471 X8:7f7f7f7f7f7f7f7f(XEN) X9:ff65685e6c6f7275X10:7f7f7f7f7f7f7f7fX11: 0101010101010101 (XEN) X12: 0000000000000038 X13: 0000000000280910 X14: 0000000000000020 (XEN) X15: 0000000000000000 X16:00000000002a7000X17:00000000002a7000(XEN) X18:00000000002a7000X19: 0000000000000000 X20:00000000ffffffed(XEN) X21:00000000002a6380X22: 0000000000335430 X23: 0000000000000002 (XEN) X24:000000000029b1f0X25:00000000002d83d0X26: 0000000048000000 (XEN) X27:00000000c0000000X28: 0000000000000001 FP:00000000002f7dc0(XEN)
(XEN)VTCR_EL2: 80000000
(XEN)VTTBR_EL2: 0000000000000000
(XEN)
(XEN)SCTLR_EL2:30cd183d
(XEN)HCR_EL2: 0000000000000038
(XEN)TTBR0_EL2:00000000781b4000
(XEN)
(XEN)ESR_EL2:f2000001
(XEN)HPFAR_EL2: 0000000000000000
(XEN)FAR_EL2: 0000000000000000
(XEN)
(XEN) Xen stack trace from sp=00000000002f7dc0:
(XEN)00000000002f7de000000000002bdd940000000000000002 0000000000000002 (XEN)00000000bfe0b66000000000002001b400000000780800000000000077e80000(XEN) 0000000048000000 0000000000000000 0000000000400000 0000000000000003 (XEN) 0000000000000001 0000000000000000 0000000078080000 0000000048080040 (XEN) 0000000000000000000000000000f0800000000048000000 0000000078000000 (XEN)00000000002d83c000000000002aa4400000000000000000 0000000000000000 (XEN) 0000000000000000 0000000300000000 000000000000000000000040ffffffff(XEN) 0000000000000400 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) Xen call trace:
(XEN) [<00000000002b3ae0>]iommu_setup+0xa0/0x18c(PC)
(XEN) [<00000000002b3a8c>]iommu_setup+0x4c/0x18c(LR)
(XEN) [<00000000002bdd94>]start_xen+0xaa0/0xc7c
(XEN) [<00000000002001b4>] arm64/head.o#primary_switched+0xc/0x2c (XEN)
(XEN)
(XEN) ****************************************
(XEN) Panic on CPU 0:
(XEN) Assertion 'unreachable' failed at...ild-workspace/build/xen/xen/include/xen/iommu.h:72
(XEN) ****************************************
(XEN)--
Regards,Oleksandr Tyshchenko
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
On 25.09.19 19:03, Paul Durrant wrote:
Hi, Paul
Quoted 13 lines
Quoted 13 lines
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control
[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macros
Quoted 7 lines
Quoted 6 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.BTW, I assume disabling the iommu on the xen command like will work round the issue?
No. Disabling the iommu will lead to ASSERT in all cases.
Signature · 9 lines
Regards, Oleksandr Tyshchenko _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 24 lines
-----Original Message-----
From: Oleksandr <[email protected]>
Sent: 25 September 2019 16:50
To: Paul Durrant <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; [email protected]; Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien Grall <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macros
Quoted 7 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.
Here a minimal patch, leaving 'force pt share' in place. Does this avoid the problem?
---8<---
xen/common/sysctl.c +4 -2
@@ -268,9 +268,11 @@ long do_sysctl(XEN_GUEST_HANDLE_PARAM(xen_sysctl_t) u_sysctl) pi->max_mfn =get_upper_mfn_bound ();arch_do_physinfo (pi);if ( iommu_enabled ) + { pi->capabilities |= XEN_SYSCTL_PHYSCAP_directio; -if ( iommu_hap_pt_share ) - pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; +if ( iommu_hap_pt_share ) + pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + }if (copy_to_guest (u_sysctl, op,1 ) ) ret = -EFAULT ;
xen/include/xen/iommu.h +0 -3
@@ -68,8 +68,6 @@ static inline void clear_iommu_hap_pt_share(void) {#ifndef iommu_hap_pt_share iommu_hap_pt_share = false; -#elif iommu_hap_pt_share -ASSERT_UNREACHABLE ();#endif } ---8<---
Paul
Quoted 67 lines
...
(XEN) Assertion 'unreachable' failed at
...ild-workspace/build/xen/xen/include/xen/iommu.h:72
(XEN) ----[ Xen-4.13-unstable arm64 debug=y Not tainted ]---- (XEN) CPU: 0
(XEN) PC:00000000002b3ae0iommu_setup+0xa0/0x18c
(XEN) LR:00000000002b3a8c
(XEN) SP:00000000002f7dc0
(XEN) CPSR:a0000249MODE:64-bit EL2h (Hypervisor, handler) (XEN) X0: 00000000002a7000 X1: 0000000000000000 X2:2c736173656e6572(XEN) X3: 0000000000000002 X4: 0000000000000001 X5: 0000000000000000 (XEN) X6: 0000000000000080 X7: 2b726072646d6471 X8:7f7f7f7f7f7f7f7f(XEN) X9:ff65685e6c6f7275X10:7f7f7f7f7f7f7f7fX11: 0101010101010101 (XEN) X12: 0000000000000038 X13: 0000000000280910 X14: 0000000000000020 (XEN) X15: 0000000000000000 X16:00000000002a7000X17:00000000002a7000(XEN) X18:00000000002a7000X19: 0000000000000000 X20:00000000ffffffed(XEN) X21:00000000002a6380X22: 0000000000335430 X23: 0000000000000002 (XEN) X24:000000000029b1f0X25:00000000002d83d0X26: 0000000048000000 (XEN) X27:00000000c0000000X28: 0000000000000001 FP:00000000002f7dc0(XEN)
(XEN)VTCR_EL2: 80000000
(XEN)VTTBR_EL2: 0000000000000000
(XEN)
(XEN)SCTLR_EL2:30cd183d
(XEN)HCR_EL2: 0000000000000038
(XEN)TTBR0_EL2:00000000781b4000
(XEN)
(XEN)ESR_EL2:f2000001
(XEN)HPFAR_EL2: 0000000000000000
(XEN)FAR_EL2: 0000000000000000
(XEN)
(XEN) Xen stack trace from sp=00000000002f7dc0:
(XEN)00000000002f7de000000000002bdd940000000000000002 0000000000000002 (XEN)00000000bfe0b66000000000002001b400000000780800000000000077e80000(XEN) 0000000048000000 0000000000000000 0000000000400000 0000000000000003 (XEN) 0000000000000001 0000000000000000 0000000078080000 0000000048080040 (XEN) 0000000000000000000000000000f0800000000048000000 0000000078000000 (XEN)00000000002d83c000000000002aa4400000000000000000 0000000000000000 (XEN) 0000000000000000 0000000300000000 000000000000000000000040ffffffff(XEN) 0000000000000400 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) 0000000000000000 0000000000000000 0000000000000000 0000000000000000 (XEN) Xen call trace:
(XEN) [<00000000002b3ae0>]iommu_setup+0xa0/0x18c(PC)
(XEN) [<00000000002b3a8c>]iommu_setup+0x4c/0x18c(LR)
(XEN) [<00000000002bdd94>]start_xen+0xaa0/0xc7c
(XEN) [<00000000002001b4>] arm64/head.o#primary_switched+0xc/0x2c (XEN)
(XEN)
(XEN) ****************************************
(XEN) Panic on CPU 0:
(XEN) Assertion 'unreachable' failed at...ild-workspace/build/xen/xen/include/xen/iommu.h:72
(XEN) ****************************************
(XEN)
Quoted 4 lines
--
Regards,Oleksandr Tyshchenko
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
On 25.09.19 19:10, Paul Durrant wrote:
Hi Paul
Quoted 57 lines
Quoted 24 lines
-----Original Message-----
From: Oleksandr <[email protected]>
Sent: 25 September 2019 16:50
To: Paul Durrant <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; [email protected]; Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien Grall <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macrosQuoted 7 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.Here a minimal patch, leaving 'force pt share' in place. Does this avoid the problem?
---8<---
xen/common/sysctl.c +4
-2@@ -268,9 +268,11 @@ long do_sysctl(XEN_GUEST_HANDLE_PARAM(xen_sysctl_t) u_sysctl) pi->max_mfn =get_upper_mfn_bound ();arch_do_physinfo (pi);if ( iommu_enabled ) + { pi->capabilities |= XEN_SYSCTL_PHYSCAP_directio; -if ( iommu_hap_pt_share ) - pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; +if ( iommu_hap_pt_share ) + pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + }if (copy_to_guest (u_sysctl, op,1 ) ) ret = -EFAULT ;xen/include/xen/iommu.h +0
-0
xen/include/xen/iommu.h static inline void clear_iommu_hap_pt_share(void)
Quoted 10 lines
xen/include/xen/iommu.h +0
-3@@ -68,8 +68,6 @@ static inline void clear_iommu_hap_pt_share(void) {#ifndef iommu_hap_pt_share iommu_hap_pt_share = false; -#elif iommu_hap_pt_share -ASSERT_UNREACHABLE ();#endif } ---8 <--- Paul
Thank you for the patch, but I need some time to check it (now I have some troubles with starting guest VM). I will inform you once I check.
Signature · 9 lines
Regards, Oleksandr Tyshchenko _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Hi Paul
Quoted 50 lines
Quoted 50 lines
Quoted 6 lines
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macrosQuoted 7 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.Here a minimal patch, leaving 'force pt share' in place. Does this avoid the problem?
---8<---
xen/common/sysctl.c +4
-2@@ -268,9 +268,11 @@ longdo_sysctl (XEN_GUEST_HANDLE_PARAM (xen_sysctl_t ) u_sysctl) pi->max_mfn = get_upper_mfn_bound(); arch_do_physinfo(pi); if ( iommu_enabled ) + { pi->capabilities |= XEN_SYSCTL_PHYSCAP_directio; - if ( iommu_hap_pt_share ) - pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + if ( iommu_hap_pt_share ) + pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + }if (
copy_to_guest(u_sysctl, op, 1))
ret =-EFAULT;xen/include/xen/iommu.h +0
-3@@ -68,8 +68,6 @@ static inline void clear_iommu_hap_pt_share(void) { #ifndef iommu_hap_pt_share iommu_hap_pt_share = false; -#elif iommu_hap_pt_share - ASSERT_UNREACHABLE(); #endif } ---8<---
Quoted 3 lines
Quoted 1 line
Paul
Thank you for the patch, but I need some time to check it (now I have some troubles with starting guest VM). I will inform you once I check.
With the patch applied, the issue I have faced (during Xen boot) is gone away. But, I haven't analyzed your "per-domain IOMMU control series" yet to have opinion regarding your patch itself.
I noticed the following:
When iommu_enabled = false (my case, when IOMMU driver is disable), I can't create guest VM if "dtdev" property is present and contains DMA masters in the domain config:
Parsing config from /xt/dom.cfg/domd.cfg
ERROR: passthrough not supported on this platform
Even if I add passthrough = "disabled", it still denies:
Parsing config from /xt/dom.cfg/domd.cfg
ERROR: passthrough disabled but devices are specified
Looks like, correct behavior...
Signature · 9 lines
Regards, Oleksandr Tyshchenko _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 89 lines
-----Original Message-----
From: Oleksandr <[email protected]>
Sent: 25 September 2019 19:07
To: Paul Durrant <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; [email protected]; Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien Grall <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU controlHi Paul
Quoted 52 lines
Quoted 52 lines
Quoted 6 lines
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macrosQuoted 8 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.Here a minimal patch, leaving 'force pt share' in place. Does this avoid the problem?
---8<---
xen/common/sysctl.c +4
-2@@ -268,9 +268,11 @@ longdo_sysctl (XEN_GUEST_HANDLE_PARAM (xen_sysctl_t ) u_sysctl) pi->max_mfn = get_upper_mfn_bound(); arch_do_physinfo(pi); if ( iommu_enabled ) + { pi->capabilities |= XEN_SYSCTL_PHYSCAP_directio; - if ( iommu_hap_pt_share ) - pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + if ( iommu_hap_pt_share ) + pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + }if (
copy_to_guest(u_sysctl, op, 1))
ret =-EFAULT;xen/include/xen/iommu.h +0
-3@@ -68,8 +68,6 @@ static inline void clear_iommu_hap_pt_share(void) { #ifndef iommu_hap_pt_share iommu_hap_pt_share = false; -#elif iommu_hap_pt_share - ASSERT_UNREACHABLE(); #endif } ---8<---Quoted 4 lines
Quoted 1 line
Paul
Thank you for the patch, but I need some time to check it (now I have some troubles with starting guest VM). I will inform you once I check.
With the patch applied, the issue I have faced (during Xen boot) is gone away. But, I haven't analyzed your "per-domain IOMMU control series" yet to have opinion regarding your patch itself.
I noticed the following:
When
iommu_enabled= false (my case, when IOMMU driver is disable), I can't create guest VM if "dtdev" property is present and contains DMA masters in the domain config:Parsing config from /xt/dom.cfg/domd.cfg
ERROR: passthrough not supported on this platform
Quoted 7 lines
Even if I add passthrough = "disabled", it still denies:
Parsing config from /xt/dom.cfg/domd.cfg
ERROR: passthrough disabled but devices are specifiedLooks like, correct behavior...
Yes, that all seems correct. Passthrough should not be done without an IOMMU.
Paul
Quoted 5 lines
--
Regards,Oleksandr Tyshchenko
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Hi,
Quoted 58 lines · Paul Durrant
Quoted 24 lines
-----Original Message-----
From: Oleksandr <[email protected]>
Sent: 25 September 2019 16:50
To: Paul Durrant <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; [email protected]; Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien Grall <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macrosQuoted 7 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.Here a minimal patch, leaving 'force pt share' in place. Does this avoid the problem?
---8<---
xen/common/sysctl.c +4
-2@@ -268,9 +268,11 @@ long do_sysctl(XEN_GUEST_HANDLE_PARAM(xen_sysctl_t) u_sysctl) pi->max_mfn =get_upper_mfn_bound ();arch_do_physinfo (pi);if ( iommu_enabled ) + { pi->capabilities |= XEN_SYSCTL_PHYSCAP_directio; -if ( iommu_hap_pt_share ) - pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; +if ( iommu_hap_pt_share ) + pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + }if (copy_to_guest (u_sysctl, op,1 ) ) ret = -EFAULT ;xen/include/xen/iommu.h +0
-0
xen/include/xen/iommu.h static inline void clear_iommu_hap_pt_share(void) · line 71
Quoted 6 lines
xen/include/xen/iommu.h +0
-2@@ -68,8 +68,6 @@ static inline void clear_iommu_hap_pt_share(void) {#ifndef iommu_hap_pt_share iommu_hap_pt_share = false; -#elif iommu_hap_pt_share -ASSERT_UNREACHABLE ();#endif
IHMO, calling this function is a mistake on platform only supporting shared page-table so the ASSERT() should be kept here.
This raises the question why the function is actually called from common code. iommu_hap_enabled() should technically not be used in any code if the IOMMU is not enabled/present. So what are you trying to prevent here?
Cheers,
Signature · 5 lines
Julien Grall _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 88 lines
-----Original Message-----
From: Julien Grall <[email protected]>
Sent: 25 September 2019 22:34
To: Paul Durrant <[email protected]>; 'Oleksandr' <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: nd <[email protected]>; Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; [email protected]; Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU controlHi,
Quoted 5 lines · Paul Durrant
Quoted 5 lines
-----Original Message-----
From: Oleksandr <[email protected]>
Sent: 25 September 2019 16:50
To: Paul Durrant <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; WeiLiu
Quoted 2 lines
Quoted 2 lines
<[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George
Dunlap
Quoted 2 lines
Quoted 2 lines
<[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; xen-
Quoted 1 line
Quoted 1 line
Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien
Grall
Quoted 48 lines
Quoted 14 lines
<[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macrosQuoted 7 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.Here a minimal patch, leaving 'force pt share' in place. Does this avoid the problem?
---8<---
xen/common/sysctl.c +4
-2@@ -268,9 +268,11 @@ long do_sysctl(XEN_GUEST_HANDLE_PARAM(xen_sysctl_t) u_sysctl) pi->max_mfn =get_upper_mfn_bound ();arch_do_physinfo (pi);if ( iommu_enabled ) + { pi->capabilities |= XEN_SYSCTL_PHYSCAP_directio; -if ( iommu_hap_pt_share ) - pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; +if ( iommu_hap_pt_share ) + pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + }if (copy_to_guest (u_sysctl, op,1 ) ) ret = -EFAULT ;xen/include/xen/iommu.h +0
-0xen/include/xen/iommu.h static inline void clear_iommu_hap_pt_share(void) · line 71
Quoted 6 lines
xen/include/xen/iommu.h +0
-2@@ -68,8 +68,6 @@ static inline void clear_iommu_hap_pt_share(void) {#ifndef iommu_hap_pt_share iommu_hap_pt_share = false; -#elif iommu_hap_pt_share -ASSERT_UNREACHABLE ();#endif IHMO, calling this function is a mistake on platform only supporting shared page-table so the
ASSERT()should be kept here.
Quoted 3 lines
This raises the question why the function is actually called from common code.
iommu_hap_enabled()should technically not be used in any code if the IOMMU is not enabled/present. So what are you trying to prevent here?
What I'm trying to prevent, on x86, is a situation where the iommu_enabled == false but iommu_hap_pt_share == true. I had, mistakenly, believed that iommu_enabled would never be false for ARM but it seems this is not the case so that situation has to be tolerated. I guess, given the other hunk of my patch, I can actually leave the ASSERT in place and avoid making the call from common code, in which case the function ought to move into an x86 header as well.
Paul
Quoted 5 lines
Cheers,
--
Julien Grall
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Hi,
Quoted 88 lines · Paul Durrant
Quoted 88 lines
-----Original Message-----
From: Julien Grall <[email protected]>
Sent: 25 September 2019 22:34
To: Paul Durrant <[email protected]>; 'Oleksandr' <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: nd <[email protected]>; Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; [email protected]; Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU controlHi,
Quoted 5 lines · Paul Durrant
Quoted 5 lines
-----Original Message-----
From: Oleksandr <[email protected]>
Sent: 25 September 2019 16:50
To: Paul Durrant <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; WeiLiu
Quoted 2 lines
Quoted 2 lines
<[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George
Dunlap
Quoted 2 lines
Quoted 2 lines
<[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; xen-
Quoted 1 line
Quoted 1 line
Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien
Grall
Quoted 48 lines
Quoted 14 lines
<[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macrosQuoted 7 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.Here a minimal patch, leaving 'force pt share' in place. Does this avoid the problem?
---8<---
xen/common/sysctl.c +4
-2@@ -268,9 +268,11 @@ long do_sysctl(XEN_GUEST_HANDLE_PARAM(xen_sysctl_t) u_sysctl) pi->max_mfn =get_upper_mfn_bound ();arch_do_physinfo (pi);if ( iommu_enabled ) + { pi->capabilities |= XEN_SYSCTL_PHYSCAP_directio; -if ( iommu_hap_pt_share ) - pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; +if ( iommu_hap_pt_share ) + pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + }if (copy_to_guest (u_sysctl, op,1 ) ) ret = -EFAULT ;xen/include/xen/iommu.h +0
-0xen/include/xen/iommu.h static inline void clear_iommu_hap_pt_share(void) · line 71
Quoted 6 lines
xen/include/xen/iommu.h +0
-2@@ -68,8 +68,6 @@ static inline void clear_iommu_hap_pt_share(void) {#ifndef iommu_hap_pt_share iommu_hap_pt_share = false; -#elif iommu_hap_pt_share -ASSERT_UNREACHABLE ();#endif IHMO, calling this function is a mistake on platform only supporting shared page-table so the
ASSERT()should be kept here.
Quoted 5 lines
Quoted 3 lines
This raises the question why the function is actually called from common code.
iommu_hap_enabled()should technically not be used in any code if the IOMMU is not enabled/present. So what are you trying to prevent here?What I'm trying to prevent, on x86, is a situation where the
iommu_enabled== false butiommu_hap_pt_share== true.
This is not entirely uncommon to have other variables gated by others.
So what could happen if you have iommu_enabled == false and iommu_hap_pt_share == true on x86?
Quoted 1 line
I had, mistakenly, believed that
iommu_enabledwould never be false for ARM but it seems this is not the case so that situation has to be tolerated. I guess, given the other hunk of my patch, I can actually leave the ASSERT in place and avoid making the call from common code, in which case the function ought to move into an x86 header as well.
By "the function", do you mean clear_iommu_hap_pt_share? If so, I think it should stay were it is. This is a generic function that might be re-used for other architecture in the future.
Cheers,
Signature · 6 lines
Julien Grall _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
Quoted 112 lines
-----Original Message-----
From: Julien Grall <[email protected]>
Sent: 26 September 2019 10:13
To: Paul Durrant <[email protected]>; 'Oleksandr' <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: nd <[email protected]>; Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; [email protected]; Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU controlHi,
Quoted 7 lines · Paul Durrant
Quoted 7 lines
-----Original Message-----
From: Julien Grall <[email protected]>
Sent: 25 September 2019 22:34
To: Paul Durrant <[email protected]>; 'Oleksandr' <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: nd <[email protected]>; Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>; Wei Liu <[email protected]>; KonradRzeszutek Wilk <[email protected]>;Andrew
Quoted 1 line
Quoted 1 line
Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>;
George
Quoted 2 lines
Quoted 2 lines
Dunlap <[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; xen-
Quoted 11 lines
Quoted 11 lines
Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU controlHi,
Quoted 5 lines · Paul Durrant
Quoted 5 lines
-----Original Message-----
From: Oleksandr <[email protected]>
Sent: 25 September 2019 16:50
To: Paul Durrant <[email protected]>; 'Jan Beulich' <[email protected]>
Cc: Petre Pircalabu <[email protected]>; Stefano Stabellini <[email protected]>;Wei
Quoted 71 lines
Quoted 70 lines
Liu
Quoted 2 lines
Quoted 2 lines
<[email protected]>; KonradRzeszutek Wilk <[email protected]>; Andrew Cooper <[email protected]>; David Scott <[email protected]>; Tim (Xen.org) <[email protected]>; George
Dunlap
Quoted 2 lines
Quoted 2 lines
<[email protected]>; Tamas K Lengyel <[email protected]>; Ian Jackson <[email protected]>; Anthony Perard <[email protected]>; xen-
Quoted 1 line
Quoted 1 line
Volodymyr Babchuk <[email protected]>; Roger Pau Monne <[email protected]>; Julien
Grall
Quoted 48 lines
Quoted 14 lines
<[email protected]>
Subject: Re: [Xen-devel] [PATCH v13 0/4] add per-domain IOMMU control[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macrosQuoted 7 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.Here a minimal patch, leaving 'force pt share' in place. Does this avoid the problem?
---8<---
xen/common/sysctl.c +4
-2@@ -268,9 +268,11 @@ long do_sysctl(XEN_GUEST_HANDLE_PARAM(xen_sysctl_t) u_sysctl) pi->max_mfn =get_upper_mfn_bound ();arch_do_physinfo (pi);if ( iommu_enabled ) + { pi->capabilities |= XEN_SYSCTL_PHYSCAP_directio; -if ( iommu_hap_pt_share ) - pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; +if ( iommu_hap_pt_share ) + pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + }if (copy_to_guest (u_sysctl, op,1 ) ) ret = -EFAULT ;xen/include/xen/iommu.h +0
-0xen/include/xen/iommu.h static inline void clear_iommu_hap_pt_share(void) · line 71
Quoted 6 lines
xen/include/xen/iommu.h +0
-2@@ -68,8 +68,6 @@ static inline void clear_iommu_hap_pt_share(void) {#ifndef iommu_hap_pt_share iommu_hap_pt_share = false; -#elif iommu_hap_pt_share -ASSERT_UNREACHABLE ();#endif IHMO, calling this function is a mistake on platform only supporting shared page-table so the
ASSERT()should be kept here.This raises the question why the function is actually called from common code.
iommu_hap_enabled()should technically not be used in any code if the IOMMU is not enabled/present. So what are you trying to prevent here?
Quoted 7 lines
Quoted 1 line
What I'm trying to prevent, on x86, is a situation where the
iommu_enabled== false but
iommu_hap_pt_share== true.This is not entirely uncommon to have other variables gated by others. So what could happen if you have
iommu_enabled== false andiommu_hap_pt_share== true on x86?
Well, I was hoping to avoid the nested if in sysctl.c.
Quoted 6 lines
Quoted 1 line
I had, mistakenly, believed that
iommu_enabledwould never be false for ARM but it seems this is notthe case so that situation has to be tolerated. I guess, given the other hunk of my patch, I can actually leave the ASSERT in place and avoid making the call from common code, in which case the function ought to move into an x86 header as well.
By "the function", do you mean
clear_iommu_hap_pt_share?
Yes.
Quoted 4 lines
If so, I think
it should stay were it is. This is a generic function that might be re-used for other architecture in the future.
That seems like a bit of a long shot. If I remove the call from iommu_setup() then the only remaining callers are in x86 code, but I suppose it can stay where it is to avoid the churn. I'll spin a new test patch.
Paul
_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
Hi Paul,
Quoted 6 lines · Paul Durrant
Quoted 4 lines
If so, I think
it should stay were it is. This is a generic function that might be re-used for other architecture in the future.That seems like a bit of a long shot. If I remove the call from
iommu_setup()then the only remaining callers are in x86 code, but I suppose it can stay where it is to avoid the churn. I'll spin a new test patch.
Not really, I know that we will likely need it on Arm when MSI doorbell will be exposed to the guest because some mappings cannot be accessed by the processor. Although, I can't tell when this will happen.
Anyway, I will have a look at your next patch :).
Cheers,
Signature · 6 lines
Julien Grall _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel
On 26.09.19 11:39, Paul Durrant wrote:
Hi Paul
Quoted 53 lines
Quoted 53 lines
Quoted 44 lines
Quoted 11 lines
[CC Julien]
Hi Paul
I may mistake, but looks like
80ff3d338dc93260b41ffeeebb0f852c2edef9ceiommu: tidy upiommu_use_hap_pt()andneed_iommu_pt_sync()macrosQuoted 7 lines
triggers
ASSERT_UNREACHABLEon Arm if no IOMMU has been found (I built with my platform's IOMMU driver disabled: #CONFIG_IPMMU_VMSAis not set) .So,
iommu_setup()callsclear_iommu_hap_pt_share()withiommu_hap_pt_sharebeing set (CONFIG_IOMMU_FORCE_PT_SHARE=y) which, actually, triggers ASSERT.Here a minimal patch, leaving 'force pt share' in place. Does this avoid the problem?
---8<---
xen/common/sysctl.c +4
-2@@ -268,9 +268,11 @@ long do_sysctl(XEN_GUEST_HANDLE_PARAM(xen_sysctl_t) u_sysctl) pi->max_mfn =get_upper_mfn_bound ();arch_do_physinfo (pi);if ( iommu_enabled ) + { pi->capabilities |= XEN_SYSCTL_PHYSCAP_directio; -if ( iommu_hap_pt_share ) - pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; +if ( iommu_hap_pt_share ) + pi->capabilities |= XEN_SYSCTL_PHYSCAP_iommu_hap_pt_share; + }if (copy_to_guest (u_sysctl, op,1 ) ) ret = -EFAULT ;xen/include/xen/iommu.h +0
-0xen/include/xen/iommu.h static inline void clear_iommu_hap_pt_share(void) · line 71
Quoted 6 lines
xen/include/xen/iommu.h +0
-2@@ -68,8 +68,6 @@ static inline void clear_iommu_hap_pt_share(void) {#ifndef iommu_hap_pt_share iommu_hap_pt_share = false; -#elif iommu_hap_pt_share -ASSERT_UNREACHABLE ();#endif IHMO, calling this function is a mistake on platform only supporting shared page-table so the
ASSERT()should be kept here.
Quoted 4 lines
Quoted 3 lines
This raises the question why the function is actually called from common code.
iommu_hap_enabled()should technically not be used in any code if the IOMMU is not enabled/present. So what are you trying to prevent here?What I'm trying to prevent, on x86, is a situation where the
iommu_enabled== false butiommu_hap_pt_share== true. I had, mistakenly, believed thatiommu_enabledwould never be false for ARM but it seems this is not the case so that situation has to be tolerated. I guess, given the other hunk of my patch, I can actually leave the ASSERT in place and avoid making the call from common code, in which case the function ought to move into an x86 header as well.
Not all Arm based SoCs (which supported by Xen) contains SMMU (the only one supported driver at the moment) or IPMMU-VMSA (on review now, but, it will be under CONFIG_EXPERT when merged, so disabled by default). So, "iommu_enabled" can be false.
Signature · 9 lines
Regards, Oleksandr Tyshchenko _______________________________________________ Xen-devel mailing list [email protected] https://lists.xenproject.org/mailman/listinfo/xen-devel