KVM: s390: Fix virtual-real address confusions
cover KVM: s390: Fix virtual-real address confusions
This series fixes several virtual-real address confusions in the basic KVM infrastructure for s390.
Since physical addresses are currently equal to virtual addresses in kernel space, the address confusions do not represent issues at the moment.
IO-related fixes are going to follow in further series.
Nico Boehr (5):
s390/mm: gmap: sort out physical vs virtual pointers usage s390/entry: sort out physical vs virtual pointers usage in sie64a
KVM: s390: sort out physical vs virtual pointers usage
KVM: s390: sida: sort out physical vs virtual pointers usage
KVM: s390: pv: sort out physical vs virtual pointers usage
arch/s390/include/asm/kvm_host.h | 12 ++- arch/s390/include/asm/stacktrace.h | 1 + arch/s390/kernel/asm-offsets.c | 1 + arch/s390/kernel/entry.S | 26 ++--- arch/s390/kvm/intercept.c | 9 +- arch/s390/kvm/kvm-s390.c | 53 ++++++----- arch/s390/kvm/kvm-s390.h | 5 +- arch/s390/kvm/priv.c | 3 +- arch/s390/kvm/pv.c | 17 ++-- arch/s390/mm/gmap.c | 147 +++++++++++++++-------------- 10 files changed, 150 insertions(+), 124 deletions(-)
Signature · 1 line
2.37.3
PATCH 1/5 s390/mm: gmap: sort out physical vs virtual pointers usage
Fix virtual vs physical address confusion (which currently are the same).
--- arch/s390/mm/gmap.c | 147 +++++++++++++++++++++++--------------------- 1 file changed, 76 insertions(+), 71 deletions(-)
arch/s390/mm/gmap.c +76 -71
@@ -72,7 +72,7 @@ static struct gmap *gmap_alloc(unsigned long limit) goto out_free; page->index =0 ;list_add (&page->lru, &gmap->crst_list); - table = (unsigned long *)page_to_phys (page); + table =page_to_virt (page);crst_table_init (table, etype); gmap->table = table; gmap->asce = atype | _ASCE_TABLE_LENGTH |
@@ -311,12 +311,12 @@ static int gmap_alloc_table(struct gmap *gmap, unsigned long *table, page =alloc_pages (GFP_KERNEL_ACCOUNT ,CRST_ALLOC_ORDER );if (!page)return -ENOMEM ; - new = (unsigned long *)page_to_phys (page); + new =page_to_virt (page);crst_table_init (new, init);spin_lock (&gmap->guest_table_lock);if (*table & _REGION_ENTRY_INVALID) {list_add (&page->lru, &gmap->crst_list); - *table = (unsigned long ) new | _REGION_ENTRY_LENGTH | + *table =__pa (new) | _REGION_ENTRY_LENGTH | (*table & _REGION_ENTRY_TYPE_MASK); page->index = gaddr; page =NULL ;
@@ -557,7 +557,7 @@ int __gmap_link(struct gmap *gmap, unsigned long gaddr, unsigned long vmaddr)gmap_alloc_table (gmap, table, _REGION2_ENTRY_EMPTY, gaddr & _REGION1_MASK)) return -ENOMEM; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); + table =__va (*table & _REGION_ENTRY_ORIGIN); }if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION2) { table += (gaddr & _REGION2_INDEX) >> _REGION2_SHIFT;
@@ -565,7 +565,7 @@ int __gmap_link(struct gmap *gmap, unsigned long gaddr, unsigned long vmaddr)gmap_alloc_table (gmap, table, _REGION3_ENTRY_EMPTY, gaddr & _REGION2_MASK)) return -ENOMEM; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); + table =__va (*table & _REGION_ENTRY_ORIGIN); }if ((gmap->asce & _ASCE_TYPE_MASK) >= _ASCE_TYPE_REGION3) { table += (gaddr & _REGION3_INDEX) >> _REGION3_SHIFT;
@@ -573,7 +573,7 @@ int __gmap_link(struct gmap *gmap, unsigned long gaddr, unsigned long vmaddr)gmap_alloc_table (gmap, table, _SEGMENT_ENTRY_EMPTY, gaddr & _REGION3_MASK)) return -ENOMEM; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); + table =__va (*table & _REGION_ENTRY_ORIGIN); } table += (gaddr & _SEGMENT_INDEX) >> _SEGMENT_SHIFT;/* Walk the parent mm page table */
@@ -813,7 +813,7 @@ static inline unsigned long *gmap_table_walk(struct gmap *gmap,break ;if (*table & _REGION_ENTRY_INVALID)return NULL ; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); + table =__va (*table & _REGION_ENTRY_ORIGIN); fallthrough;case _ASCE_TYPE_REGION2: table += (gaddr & _REGION2_INDEX) >> _REGION2_SHIFT;
@@ -821,7 +821,7 @@ static inline unsigned long *gmap_table_walk(struct gmap *gmap,break ;if (*table & _REGION_ENTRY_INVALID)return NULL ; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); + table =__va (*table & _REGION_ENTRY_ORIGIN); fallthrough;case _ASCE_TYPE_REGION3: table += (gaddr & _REGION3_INDEX) >> _REGION3_SHIFT;
@@ -829,7 +829,7 @@ static inline unsigned long *gmap_table_walk(struct gmap *gmap,break ;if (*table & _REGION_ENTRY_INVALID)return NULL ; - table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN); + table =__va (*table & _REGION_ENTRY_ORIGIN); fallthrough;case _ASCE_TYPE_SEGMENT: table += (gaddr & _SEGMENT_INDEX) >> _SEGMENT_SHIFT;
@@ -837,7 +837,7 @@ static inline unsigned long *gmap_table_walk(struct gmap *gmap,break ;if (*table & _REGION_ENTRY_INVALID)return NULL ; - table = (unsigned long *)(*table & _SEGMENT_ENTRY_ORIGIN); + table =__va (*table & _SEGMENT_ENTRY_ORIGIN); table += (gaddr & _PAGE_INDEX) >> _PAGE_SHIFT; }return table;
@@ -1150,7 +1150,7 @@ int gmap_read_table(struct gmap *gmap, unsigned long gaddr, unsigned long *val)if (pte_present (pte) && (pte_val (pte) & _PAGE_READ)) { address =pte_val (pte) &PAGE_MASK ; address += gaddr & ~PAGE_MASK ; - *val = *(unsigned long *) address; + *val = *(unsigned long *)__va (address);set_pte (ptep,set_pte_bit (*ptep,__pgprot (_PAGE_YOUNG)));/* Do *NOT* clear the _PAGE_INVALID bit! */ rc =0 ;
@@ -1335,7 +1335,8 @@ static void __gmap_unshadow_pgt(struct gmap *sg, unsigned long raddr,*/ static void gmap_unshadow_pgt (struct gmap *sg,unsigned long raddr) { -unsigned long sto, *ste, *pgt; +unsigned long *ste; +phys_addr_t sto, pgt;struct page *page;BUG_ON (!gmap_is_shadow (sg));
@@ -1343,13 +1344,13 @@ static void gmap_unshadow_pgt(struct gmap *sg, unsigned long raddr)if (!ste || !(*ste & _SEGMENT_ENTRY_ORIGIN))return ;gmap_call_notifier (sg, raddr, raddr + _SEGMENT_SIZE -1 ); - sto = (unsigned long ) (ste - ((raddr & _SEGMENT_INDEX) >> _SEGMENT_SHIFT)); + sto =__pa (ste - ((raddr & _SEGMENT_INDEX) >> _SEGMENT_SHIFT));gmap_idte_one (sto | _ASCE_TYPE_SEGMENT, raddr); - pgt = (unsigned long *)(*ste & _SEGMENT_ENTRY_ORIGIN); + pgt = *ste & _SEGMENT_ENTRY_ORIGIN; *ste = _SEGMENT_ENTRY_EMPTY; -__gmap_unshadow_pgt (sg, raddr, pgt); +__gmap_unshadow_pgt (sg, raddr,__va (pgt));/* Free page table */ - page =pfn_to_page (__pa (pgt) >>PAGE_SHIFT ); + page =phys_to_page (pgt);list_del (&page->lru);page_table_free_pgste (page); }
@@ -1365,19 +1366,19 @@ static void gmap_unshadow_pgt(struct gmap *sg, unsigned long raddr)static void __gmap_unshadow_sgt (struct gmap *sg,unsigned long raddr,unsigned long *sgt) { -unsigned long *pgt;struct page *page; +phys_addr_t pgt;int i;BUG_ON (!gmap_is_shadow (sg));for (i =0 ; i < _CRST_ENTRIES; i++, raddr += _SEGMENT_SIZE) {if (!(sgt[i] & _SEGMENT_ENTRY_ORIGIN))continue ; - pgt = (unsigned long *)(sgt[i] & _REGION_ENTRY_ORIGIN); + pgt = sgt[i] & _REGION_ENTRY_ORIGIN; sgt[i] = _SEGMENT_ENTRY_EMPTY; -__gmap_unshadow_pgt (sg, raddr, pgt); +__gmap_unshadow_pgt (sg, raddr,__va (pgt));/* Free page table */ - page =pfn_to_page (__pa (pgt) >>PAGE_SHIFT ); + page =phys_to_page (pgt);list_del (&page->lru);page_table_free_pgste (page); }
@@ -1392,7 +1393,8 @@ static void __gmap_unshadow_sgt(struct gmap *sg, unsigned long raddr,*/ static void gmap_unshadow_sgt (struct gmap *sg,unsigned long raddr) { -unsigned long r3o, *r3e, *sgt; +unsigned long r3o, *r3e; +phys_addr_t sgt;struct page *page;BUG_ON (!gmap_is_shadow (sg));
@@ -1401,12 +1403,12 @@ static void gmap_unshadow_sgt(struct gmap *sg, unsigned long raddr)return ;gmap_call_notifier (sg, raddr, raddr + _REGION3_SIZE -1 ); r3o = (unsigned long ) (r3e - ((raddr & _REGION3_INDEX) >> _REGION3_SHIFT)); -gmap_idte_one (r3o | _ASCE_TYPE_REGION3, raddr); - sgt = (unsigned long *)(*r3e & _REGION_ENTRY_ORIGIN); +gmap_idte_one (__pa (r3o) | _ASCE_TYPE_REGION3, raddr); + sgt = *r3e & _REGION_ENTRY_ORIGIN; *r3e = _REGION3_ENTRY_EMPTY; -__gmap_unshadow_sgt (sg, raddr, sgt); +__gmap_unshadow_sgt (sg, raddr,__va (sgt));/* Free segment table */ - page =pfn_to_page (__pa (sgt) >>PAGE_SHIFT ); + page =phys_to_page (sgt);list_del (&page->lru);__free_pages (page,CRST_ALLOC_ORDER ); }
@@ -1422,19 +1424,19 @@ static void gmap_unshadow_sgt(struct gmap *sg, unsigned long raddr)static void __gmap_unshadow_r3t (struct gmap *sg,unsigned long raddr,unsigned long *r3t) { -unsigned long *sgt;struct page *page; +phys_addr_t sgt;int i;BUG_ON (!gmap_is_shadow (sg));for (i =0 ; i < _CRST_ENTRIES; i++, raddr += _REGION3_SIZE) {if (!(r3t[i] & _REGION_ENTRY_ORIGIN))continue ; - sgt = (unsigned long *)(r3t[i] & _REGION_ENTRY_ORIGIN); + sgt = r3t[i] & _REGION_ENTRY_ORIGIN; r3t[i] = _REGION3_ENTRY_EMPTY; -__gmap_unshadow_sgt (sg, raddr, sgt); +__gmap_unshadow_sgt (sg, raddr,__va (sgt));/* Free segment table */ - page =pfn_to_page (__pa (sgt) >>PAGE_SHIFT ); + page =phys_to_page (sgt);list_del (&page->lru);__free_pages (page,CRST_ALLOC_ORDER ); }
@@ -1449,7 +1451,8 @@ static void __gmap_unshadow_r3t(struct gmap *sg, unsigned long raddr,*/ static void gmap_unshadow_r3t (struct gmap *sg,unsigned long raddr) { -unsigned long r2o, *r2e, *r3t; +unsigned long r2o, *r2e; +phys_addr_t r3t;struct page *page;BUG_ON (!gmap_is_shadow (sg));
@@ -1458,12 +1461,12 @@ static void gmap_unshadow_r3t(struct gmap *sg, unsigned long raddr)return ;gmap_call_notifier (sg, raddr, raddr + _REGION2_SIZE -1 ); r2o = (unsigned long ) (r2e - ((raddr & _REGION2_INDEX) >> _REGION2_SHIFT)); -gmap_idte_one (r2o | _ASCE_TYPE_REGION2, raddr); - r3t = (unsigned long *)(*r2e & _REGION_ENTRY_ORIGIN); +gmap_idte_one (__pa (r2o) | _ASCE_TYPE_REGION2, raddr); + r3t = *r2e & _REGION_ENTRY_ORIGIN; *r2e = _REGION2_ENTRY_EMPTY; -__gmap_unshadow_r3t (sg, raddr, r3t); +__gmap_unshadow_r3t (sg, raddr,__va (r3t));/* Free region 3 table */ - page =pfn_to_page (__pa (r3t) >>PAGE_SHIFT ); + page =phys_to_page (r3t);list_del (&page->lru);__free_pages (page,CRST_ALLOC_ORDER ); }
@@ -1479,7 +1482,7 @@ static void gmap_unshadow_r3t(struct gmap *sg, unsigned long raddr)static void __gmap_unshadow_r2t (struct gmap *sg,unsigned long raddr,unsigned long *r2t) { -unsigned long *r3t; +phys_addr_t r3t;struct page *page;int i;
@@ -1487,11 +1490,11 @@ static void __gmap_unshadow_r2t(struct gmap *sg, unsigned long raddr,for (i =0 ; i < _CRST_ENTRIES; i++, raddr += _REGION2_SIZE) {if (!(r2t[i] & _REGION_ENTRY_ORIGIN))continue ; - r3t = (unsigned long *)(r2t[i] & _REGION_ENTRY_ORIGIN); + r3t = r2t[i] & _REGION_ENTRY_ORIGIN; r2t[i] = _REGION2_ENTRY_EMPTY; -__gmap_unshadow_r3t (sg, raddr, r3t); +__gmap_unshadow_r3t (sg, raddr,__va (r3t));/* Free region 3 table */ - page =pfn_to_page (__pa (r3t) >>PAGE_SHIFT ); + page =phys_to_page (r3t);list_del (&page->lru);__free_pages (page,CRST_ALLOC_ORDER ); }
@@ -1506,8 +1509,9 @@ static void __gmap_unshadow_r2t(struct gmap *sg, unsigned long raddr,*/ static void gmap_unshadow_r2t (struct gmap *sg,unsigned long raddr) { -unsigned long r1o, *r1e, *r2t; +unsigned long r1o, *r1e;struct page *page; +phys_addr_t r2t;BUG_ON (!gmap_is_shadow (sg)); r1e =gmap_table_walk (sg, raddr,4 );/* get region-1 pointer */
@@ -1515,12 +1519,12 @@ static void gmap_unshadow_r2t(struct gmap *sg, unsigned long raddr)return ;gmap_call_notifier (sg, raddr, raddr + _REGION1_SIZE -1 ); r1o = (unsigned long ) (r1e - ((raddr & _REGION1_INDEX) >> _REGION1_SHIFT)); -gmap_idte_one (r1o | _ASCE_TYPE_REGION1, raddr); - r2t = (unsigned long *)(*r1e & _REGION_ENTRY_ORIGIN); +gmap_idte_one (__pa (r1o) | _ASCE_TYPE_REGION1, raddr); + r2t = *r1e & _REGION_ENTRY_ORIGIN; *r1e = _REGION1_ENTRY_EMPTY; -__gmap_unshadow_r2t (sg, raddr, r2t); +__gmap_unshadow_r2t (sg, raddr,__va (r2t));/* Free region 2 table */ - page =pfn_to_page (__pa (r2t) >>PAGE_SHIFT ); + page =phys_to_page (r2t);list_del (&page->lru);__free_pages (page,CRST_ALLOC_ORDER ); }
@@ -1536,22 +1540,23 @@ static void gmap_unshadow_r2t(struct gmap *sg, unsigned long raddr)static void __gmap_unshadow_r1t (struct gmap *sg,unsigned long raddr,unsigned long *r1t) { -unsigned long asce, *r2t; +unsigned long asce;struct page *page; +phys_addr_t r2t;int i;BUG_ON (!gmap_is_shadow (sg)); - asce = (unsigned long ) r1t | _ASCE_TYPE_REGION1; + asce =__pa (r1t) | _ASCE_TYPE_REGION1;for (i =0 ; i < _CRST_ENTRIES; i++, raddr += _REGION1_SIZE) {if (!(r1t[i] & _REGION_ENTRY_ORIGIN))continue ; - r2t = (unsigned long *)(r1t[i] & _REGION_ENTRY_ORIGIN); -__gmap_unshadow_r2t (sg, raddr, r2t); + r2t = r1t[i] & _REGION_ENTRY_ORIGIN; +__gmap_unshadow_r2t (sg, raddr,__va (r2t));/* Clear entry and flush translation r1t -> r2t */ gmap_idte_one (asce, raddr); r1t[i] = _REGION1_ENTRY_EMPTY;/* Free region 2 table */ - page =pfn_to_page (__pa (r2t) >>PAGE_SHIFT ); + page =phys_to_page (r2t);list_del (&page->lru);__free_pages (page,CRST_ALLOC_ORDER ); }
@@ -1573,7 +1578,7 @@ static void gmap_unshadow(struct gmap *sg) sg->removed =1 ;gmap_call_notifier (sg,0 ,-1UL );gmap_flush_tlb (sg); - table = (unsigned long *)(sg->asce & _ASCE_ORIGIN); + table =__va (sg->asce & _ASCE_ORIGIN);switch (sg->asce & _ASCE_TYPE_MASK) {case _ASCE_TYPE_REGION1:__gmap_unshadow_r1t (sg,0 , table);
@@ -1748,7 +1753,8 @@ int gmap_shadow_r2t(struct gmap *sg, unsigned long saddr, unsigned long r2t, int fake) {unsigned long raddr, origin, offset, len; -unsigned long *s_r2t, *table; +unsigned long *table; +phys_addr_t s_r2t;struct page *page;int rc;
@@ -1760,7 +1766,7 @@ int gmap_shadow_r2t(struct gmap *sg, unsigned long saddr, unsigned long r2t, page->index = r2t & _REGION_ENTRY_ORIGIN;if (fake) page->index |=GMAP_SHADOW_FAKE_TABLE ; - s_r2t = (unsigned long *)page_to_phys (page); + s_r2t =page_to_phys (page);/* Install shadow region second table */ spin_lock (&sg->guest_table_lock); table =gmap_table_walk (sg, saddr,4 );/* get region-1 pointer */
@@ -1775,9 +1781,9 @@ int gmap_shadow_r2t(struct gmap *sg, unsigned long saddr, unsigned long r2t, rc = -EAGAIN ;/* Race with shadow */ goto out_free; } -crst_table_init (s_r2t, _REGION2_ENTRY_EMPTY); +crst_table_init (__va (s_r2t), _REGION2_ENTRY_EMPTY);/* mark as invalid as long as the parent table is not protected */ - *table = (unsigned long ) s_r2t | _REGION_ENTRY_LENGTH | + *table = s_r2t | _REGION_ENTRY_LENGTH | _REGION_ENTRY_TYPE_R1 | _REGION_ENTRY_INVALID;if (sg->edat_level >=1 ) *table |= (r2t & _REGION_ENTRY_PROTECT);
@@ -1798,8 +1804,7 @@ int gmap_shadow_r2t(struct gmap *sg, unsigned long saddr, unsigned long r2t,spin_lock (&sg->guest_table_lock);if (!rc) { table =gmap_table_walk (sg, saddr,4 ); -if (!table || (*table & _REGION_ENTRY_ORIGIN) != - (unsigned long ) s_r2t) +if (!table || (*table & _REGION_ENTRY_ORIGIN) != s_r2t) rc = -EAGAIN ;/* Race with unshadow */ else *table &= ~_REGION_ENTRY_INVALID;
@@ -1832,7 +1837,8 @@ int gmap_shadow_r3t(struct gmap *sg, unsigned long saddr, unsigned long r3t, int fake) {unsigned long raddr, origin, offset, len; -unsigned long *s_r3t, *table; +unsigned long *table; +phys_addr_t s_r3t;struct page *page;int rc;
@@ -1844,7 +1850,7 @@ int gmap_shadow_r3t(struct gmap *sg, unsigned long saddr, unsigned long r3t, page->index = r3t & _REGION_ENTRY_ORIGIN;if (fake) page->index |=GMAP_SHADOW_FAKE_TABLE ; - s_r3t = (unsigned long *)page_to_phys (page); + s_r3t =page_to_phys (page);/* Install shadow region second table */ spin_lock (&sg->guest_table_lock); table =gmap_table_walk (sg, saddr,3 );/* get region-2 pointer */
@@ -1859,9 +1865,9 @@ int gmap_shadow_r3t(struct gmap *sg, unsigned long saddr, unsigned long r3t, rc = -EAGAIN ;/* Race with shadow */ goto out_free; } -crst_table_init (s_r3t, _REGION3_ENTRY_EMPTY); +crst_table_init (__va (s_r3t), _REGION3_ENTRY_EMPTY);/* mark as invalid as long as the parent table is not protected */ - *table = (unsigned long ) s_r3t | _REGION_ENTRY_LENGTH | + *table = s_r3t | _REGION_ENTRY_LENGTH | _REGION_ENTRY_TYPE_R2 | _REGION_ENTRY_INVALID;if (sg->edat_level >=1 ) *table |= (r3t & _REGION_ENTRY_PROTECT);
@@ -1882,8 +1888,7 @@ int gmap_shadow_r3t(struct gmap *sg, unsigned long saddr, unsigned long r3t,spin_lock (&sg->guest_table_lock);if (!rc) { table =gmap_table_walk (sg, saddr,3 ); -if (!table || (*table & _REGION_ENTRY_ORIGIN) != - (unsigned long ) s_r3t) +if (!table || (*table & _REGION_ENTRY_ORIGIN) != s_r3t) rc = -EAGAIN ;/* Race with unshadow */ else *table &= ~_REGION_ENTRY_INVALID;
@@ -1916,7 +1921,8 @@ int gmap_shadow_sgt(struct gmap *sg, unsigned long saddr, unsigned long sgt, int fake) {unsigned long raddr, origin, offset, len; -unsigned long *s_sgt, *table; +unsigned long *table; +phys_addr_t s_sgt;struct page *page;int rc;
@@ -1928,7 +1934,7 @@ int gmap_shadow_sgt(struct gmap *sg, unsigned long saddr, unsigned long sgt, page->index = sgt & _REGION_ENTRY_ORIGIN;if (fake) page->index |=GMAP_SHADOW_FAKE_TABLE ; - s_sgt = (unsigned long *)page_to_phys (page); + s_sgt =page_to_phys (page);/* Install shadow region second table */ spin_lock (&sg->guest_table_lock); table =gmap_table_walk (sg, saddr,2 );/* get region-3 pointer */
@@ -1943,9 +1949,9 @@ int gmap_shadow_sgt(struct gmap *sg, unsigned long saddr, unsigned long sgt, rc = -EAGAIN ;/* Race with shadow */ goto out_free; } -crst_table_init (s_sgt, _SEGMENT_ENTRY_EMPTY); +crst_table_init (__va (s_sgt), _SEGMENT_ENTRY_EMPTY);/* mark as invalid as long as the parent table is not protected */ - *table = (unsigned long ) s_sgt | _REGION_ENTRY_LENGTH | + *table = s_sgt | _REGION_ENTRY_LENGTH | _REGION_ENTRY_TYPE_R3 | _REGION_ENTRY_INVALID;if (sg->edat_level >=1 ) *table |= sgt & _REGION_ENTRY_PROTECT;
@@ -1966,8 +1972,7 @@ int gmap_shadow_sgt(struct gmap *sg, unsigned long saddr, unsigned long sgt,spin_lock (&sg->guest_table_lock);if (!rc) { table =gmap_table_walk (sg, saddr,2 ); -if (!table || (*table & _REGION_ENTRY_ORIGIN) != - (unsigned long ) s_sgt) +if (!table || (*table & _REGION_ENTRY_ORIGIN) != s_sgt) rc = -EAGAIN ;/* Race with unshadow */ else *table &= ~_REGION_ENTRY_INVALID;
@@ -2040,8 +2045,9 @@ int gmap_shadow_pgt(struct gmap *sg, unsigned long saddr, unsigned long pgt, int fake) {unsigned long raddr, origin; -unsigned long *s_pgt, *table; +unsigned long *table;struct page *page; +phys_addr_t s_pgt;int rc;BUG_ON (!gmap_is_shadow (sg) || (pgt & _SEGMENT_ENTRY_LARGE));
@@ -2052,7 +2058,7 @@ int gmap_shadow_pgt(struct gmap *sg, unsigned long saddr, unsigned long pgt, page->index = pgt & _SEGMENT_ENTRY_ORIGIN;if (fake) page->index |=GMAP_SHADOW_FAKE_TABLE ; - s_pgt = (unsigned long *)page_to_phys (page); + s_pgt =page_to_phys (page);/* Install shadow page table */ spin_lock (&sg->guest_table_lock); table =gmap_table_walk (sg, saddr,1 );/* get segment pointer */
@@ -2085,8 +2091,7 @@ int gmap_shadow_pgt(struct gmap *sg, unsigned long saddr, unsigned long pgt,spin_lock (&sg->guest_table_lock);if (!rc) { table =gmap_table_walk (sg, saddr,1 ); -if (!table || (*table & _SEGMENT_ENTRY_ORIGIN) != - (unsigned long ) s_pgt) +if (!table || (*table & _SEGMENT_ENTRY_ORIGIN) != s_pgt) rc = -EAGAIN ;/* Race with unshadow */ else *table &= ~_SEGMENT_ENTRY_INVALID;
Signature · 1 line
2.37.3
PATCH 2/5 s390/entry: sort out physical vs virtual pointers usage in sie64a
In reply to Nico BoehrFix virtual vs physical address confusion (which currently are the same).
sie_block is accessed in entry.S and passed it to hardware, which is why both its physical and virtual address are needed. To avoid every caller having to do the virtual-physical conversion, add a new function sie64a() which converts the virtual address to physical.
--- arch/s390/include/asm/kvm_host.h | 8 +++++++- arch/s390/include/asm/stacktrace.h | 1 + arch/s390/kernel/asm-offsets.c | 1 + arch/s390/kernel/entry.S | 26 +++++++++++++++----------- 4 files changed, 24 insertions(+), 12 deletions(-)
arch/s390/include/asm/kvm_host.h +7 -1
@@ -1017,7 +1017,13 @@ void kvm_arch_crypto_clear_masks(struct kvm *kvm);void kvm_arch_crypto_set_masks (struct kvm *kvm,unsigned long *apm,unsigned long *aqm,unsigned long *adm); -extern int sie64a (struct kvm_s390_sie_block *,u64 *); +int __sie64a (phys_addr_t sie_block_phys,struct kvm_s390_sie_block *sie_block,u64 *rsa); + +static inline int sie64a (struct kvm_s390_sie_block *sie_block,u64 *rsa) +{ +return __sie64a (virt_to_phys (sie_block), sie_block, rsa); +} +extern char sie_exit;extern int kvm_s390_gisc_register (struct kvm *kvm,u32 gisc);
arch/s390/include/asm/stacktrace.h +1 -0
@@ -46,6 +46,7 @@ struct stack_frame {unsigned long sie_savearea;unsigned long sie_reason;unsigned long sie_flags; +unsigned long sie_control_block_phys; }; };unsigned long gprs[10 ];
arch/s390/kernel/asm-offsets.c +1 -0
@@ -62,6 +62,7 @@ int main(void)OFFSET (__SF_SIE_SAVEAREA, stack_frame, sie_savearea);OFFSET (__SF_SIE_REASON, stack_frame, sie_reason);OFFSET (__SF_SIE_FLAGS, stack_frame, sie_flags); +OFFSET (__SF_SIE_CONTROL_PHYS, stack_frame, sie_control_block_phys);DEFINE (STACK_FRAME_OVERHEAD ,sizeof (struct stack_frame ));BLANK ();/* idle data offsets */
arch/s390/kernel/entry.S +15 -11
@@ -225,18 +225,20 @@ ENDPROC(__switch_to) #if IS_ENABLED(CONFIG_KVM) /* - * sie64a calling convention: - * %r2 pointer to sie control block - * %r3 guest register save area + * __sie64a calling convention: + * %r2 pointer to sie control block phys + * %r3 pointer to sie control block virt + * %r4 guest register save area */ -ENTRY(sie64a) +ENTRY(__sie64a) stmg %r6,%r14,__SF_GPRS(%r15) # save kernel registers lg %r12,__LC_CURRENT - stg %r2,__SF_SIE_CONTROL(%r15) # save control block pointer - stg %r3,__SF_SIE_SAVEAREA(%r15) # save guest register save area + stg %r2,__SF_SIE_CONTROL_PHYS(%r15) # save sie block physical.. + stg %r3,__SF_SIE_CONTROL(%r15) # ...and virtual addresses + stg %r4,__SF_SIE_SAVEAREA(%r15) # save guest register save area xc __SF_SIE_REASON(8,%r15),__SF_SIE_REASON(%r15) # reason code = 0 mvc __SF_SIE_FLAGS(8,%r15),__TI_flags(%r12) # copy thread flags - lmg %r0,%r13,0(%r3) # load guest gprs 0-13 + lmg %r0,%r13,0(%r4) # load guest gprs 0-13 lg %r14,__LC_GMAP # get gmap pointer ltgr %r14,%r14 jz .Lsie_gmap
@@ -248,6 +250,7 @@ ENTRY(sie64a) jnz .Lsie_skip TSTMSK __LC_CPU_FLAGS,_CIF_FPU jo .Lsie_skip # exit if fp/vx regs changed + lg %r14,__SF_SIE_CONTROL_PHYS(%r15) # get sie block phys addr BPEXIT __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST) .Lsie_entry: sie 0(%r14)
@@ -258,13 +261,14 @@ ENTRY(sie64a) BPOFF BPENTER __SF_SIE_FLAGS(%r15),(_TIF_ISOLATE_BP|_TIF_ISOLATE_BP_GUEST) .Lsie_skip: + lg %r14,__SF_SIE_CONTROL(%r15) # get control block pointer ni __SIE_PROG0C+3(%r14),0xfe # no longer in SIE lctlg %c1,%c1,__LC_KERNEL_ASCE # load primary asce .Lsie_done: # some program checks are suppressing. C code (e.g. do_protection_exception) # will rewind the PSW by the ILC, which is often 4 bytes in case of SIE. There # are some corner cases (e.g. runtime instrumentation) where ILC is unpredictable. -# Other instructions between sie64a and .Lsie_done should not cause program +# Other instructions between __sie64a and .Lsie_done should not cause program # interrupts. So lets use 3 nops as a landing pad for all possible rewinds. .Lrewind_pad6: nopr 7
@@ -293,8 +297,8 @@ sie_exit: EX_TABLE(.Lrewind_pad4,.Lsie_fault) EX_TABLE(.Lrewind_pad2,.Lsie_fault) EX_TABLE(sie_exit,.Lsie_fault) -ENDPROC(sie64a) -EXPORT_SYMBOL(sie64a) +ENDPROC(__sie64a) +EXPORT_SYMBOL(__sie64a) EXPORT_SYMBOL(sie_exit) #endif
@@ -373,7 +377,7 @@ ENTRY(pgm_check_handler) j 3f # -> fault in user space .Lpgm_skip_asce: #if IS_ENABLED(CONFIG_KVM) - # cleanup critical section for program checks in sie64a + # cleanup critical section for program checks in __sie64a OUTSIDE %r9,.Lsie_gmap,.Lsie_done,1f SIEEXIT lghi %r10,_PIF_GUEST_FAULT
Signature · 1 line
2.37.3
PATCH 3/5 KVM: s390: sort out physical vs virtual pointers usage
In reply to Nico BoehrFix virtual vs physical address confusion (which currently are the same).
--- arch/s390/include/asm/kvm_host.h | 1 + arch/s390/kvm/intercept.c | 2 +- arch/s390/kvm/kvm-s390.c | 44 ++++++++++++++++++-------------- arch/s390/kvm/kvm-s390.h | 5 ++-- 4 files changed, 30 insertions(+), 22 deletions(-)
arch/s390/include/asm/kvm_host.h +1 -0
@@ -276,6 +276,7 @@ struct kvm_s390_sie_block {#define ECB3_AES 0x04#define ECB3_RI 0x01__u8 ecb3;/* 0x0063 */ +#define ESCA_SCAOL_MASK ~0x3fU__u32 scaol;/* 0x0064 */ __u8 sdf;/* 0x0068 */ __u8 epdx;/* 0x0069 */
arch/s390/kvm/intercept.c +1 -1
@@ -217,7 +217,7 @@ static int handle_itdb(struct kvm_vcpu *vcpu)return 0 ;if (current->thread.per_flags &PER_FLAG_NO_TE )return 0 ; - itdb = (struct kvm_s390_itdb *)vcpu->arch.sie_block->itdba; + itdb =phys_to_virt (vcpu->arch.sie_block->itdba); rc =write_guest_lc (vcpu, __LC_PGM_TDB, itdb,sizeof (*itdb));if (rc)return rc;
arch/s390/kvm/kvm-s390.c +25 -19
105 lines not on this page · raw message
arch/s390/kvm/kvm-s390.h +3 -2
@@ -23,7 +23,8 @@/* Transactional Memory Execution related macros */ #define IS_TE_ENABLED (vcpu) ((vcpu->arch.sie_block->ecb & ECB_TE))#define TDB_FORMAT1 1 -#define IS_ITDB_VALID (vcpu) ((*(char *)vcpu->arch.sie_block->itdba == TDB_FORMAT1)) +#define IS_ITDB_VALID (vcpu) \ + ((*(char *)phys_to_virt((vcpu)->arch.sie_block->itdba) == TDB_FORMAT1))extern debug_info_t *kvm_s390_dbf;extern debug_info_t *kvm_s390_dbf_uv;
@@ -233,7 +234,7 @@ static inline unsigned long kvm_s390_get_gfn_end(struct kvm_memslots *slots)static inline u32 kvm_s390_get_gisa_desc (struct kvm *kvm) { -u32 gd = (u32 )(u64 )kvm->arch.gisa_int.origin; +u32 gd =virt_to_phys (kvm->arch.gisa_int.origin);if (gd && sclp.has_gisaf) gd |=GISA_FORMAT1 ;
Signature · 1 line
2.37.3
PATCH 4/5 KVM: s390: sida: sort out physical vs virtual pointers usage
In reply to Nico BoehrAll callers of the sida_origin() macro actually expected a virtual address, so rename it to sida_addr() and hand out a virtual address.
At some places, the macro wasn't used, potentially creating problems if the sida size ever becomes nonzero (not currently the case), so let's start using it everywhere now while at it.
--- arch/s390/include/asm/kvm_host.h | 3 +-- arch/s390/kvm/intercept.c | 7 +++---- arch/s390/kvm/kvm-s390.c | 9 +++++---- arch/s390/kvm/priv.c | 3 +-- arch/s390/kvm/pv.c | 8 +++++--- 5 files changed, 15 insertions(+), 15 deletions(-)
arch/s390/include/asm/kvm_host.h +1 -2
@@ -142,8 +142,7 @@ struct mcck_volatile_info { CR14_EXTERNAL_DAMAGE_SUBMASK)#define SIDAD_SIZE_MASK 0xff -#define sida_origin (sie_block) \ - ((sie_block)->sidad & PAGE_MASK) +#define sida_addr (sie_block) phys_to_virt((sie_block)->sidad & PAGE_MASK)#define sida_size (sie_block) \ ((((sie_block)->sidad & SIDAD_SIZE_MASK) + 1) * PAGE_SIZE)
arch/s390/kvm/intercept.c +3 -4
@@ -409,8 +409,7 @@ int handle_sthyi(struct kvm_vcpu *vcpu) out:if (!cc) {if (kvm_s390_pv_cpu_is_protected (vcpu)) { -memcpy ((void *)(sida_origin (vcpu->arch.sie_block)), - sctns,PAGE_SIZE ); +memcpy (sida_addr (vcpu->arch.sie_block), sctns,PAGE_SIZE ); }else { r =write_guest (vcpu, addr, reg2, sctns,PAGE_SIZE );if (r) {
@@ -464,7 +463,7 @@ static int handle_operexc(struct kvm_vcpu *vcpu)static int handle_pv_spx (struct kvm_vcpu *vcpu) { -u32 pref = *(u32 *)vcpu->arch.sie_block->sidad; +u32 pref = *(u32 *)sida_addr (vcpu->arch.sie_block);kvm_s390_set_prefix (vcpu, pref);trace_kvm_s390_handle_prefix (vcpu,1 , pref);
@@ -497,7 +496,7 @@ static int handle_pv_sclp(struct kvm_vcpu *vcpu)static int handle_pv_uvc (struct kvm_vcpu *vcpu) { -struct uv_cb_share *guest_uvcb = (void *)vcpu->arch.sie_block->sidad; +struct uv_cb_share *guest_uvcb =sida_addr (vcpu->arch.sie_block);struct uv_cb_cts uvcb = { .header.cmd =UVC_CMD_UNPIN_PAGE_SHARED , .header.len =sizeof (uvcb),
arch/s390/kvm/kvm-s390.c +5 -4
@@ -5167,6 +5167,7 @@ static long kvm_s390_vcpu_sida_op(struct kvm_vcpu *vcpu, struct kvm_s390_mem_op *mop) {void __user *uaddr = (void __user *)mop->buf; +void *sida_addr;int r =0 ;if (mop->flags || !mop->size)
@@ -5178,16 +5179,16 @@ static long kvm_s390_vcpu_sida_op(struct kvm_vcpu *vcpu,if (!kvm_s390_pv_cpu_is_protected (vcpu))return -EINVAL ; + sida_addr = (char *)sida_addr (vcpu->arch.sie_block) + mop->sida_offset; +switch (mop->op) {case KVM_S390_MEMOP_SIDA_READ : -if (copy_to_user (uaddr, (void *)(sida_origin (vcpu->arch.sie_block) + - mop->sida_offset), mop->size)) +if (copy_to_user (uaddr, sida_addr, mop->size)) r = -EFAULT ;break ;case KVM_S390_MEMOP_SIDA_WRITE : -if (copy_from_user ((void *)(sida_origin (vcpu->arch.sie_block) + - mop->sida_offset), uaddr, mop->size)) +if (copy_from_user (sida_addr, uaddr, mop->size)) r = -EFAULT ;break ; }
arch/s390/kvm/priv.c +1 -2
@@ -924,8 +924,7 @@ static int handle_stsi(struct kvm_vcpu *vcpu)return -EREMOTE ; }if (kvm_s390_pv_cpu_is_protected (vcpu)) { -memcpy ((void *)sida_origin (vcpu->arch.sie_block), (void *)mem, -PAGE_SIZE ); +memcpy (sida_addr (vcpu->arch.sie_block), (void *)mem,PAGE_SIZE ); rc =0 ; }else { rc =write_guest (vcpu, operand2, ar, (void *)mem,PAGE_SIZE );
arch/s390/kvm/pv.c +5 -3
@@ -44,7 +44,7 @@ int kvm_s390_pv_destroy_cpu(struct kvm_vcpu *vcpu, u16 *rc, u16 *rrc)free_pages (vcpu->arch.pv.stor_base,get_order (uv_info.guest_cpu_stor_len)); -free_page (sida_origin (vcpu->arch.sie_block)); +free_page ((unsigned long )sida_addr (vcpu->arch.sie_block)); vcpu->arch.sie_block->pv_handle_cpu =0 ; vcpu->arch.sie_block->pv_handle_config =0 ;memset (&vcpu->arch.pv,0 ,sizeof (vcpu->arch.pv));
@@ -66,6 +66,7 @@ int kvm_s390_pv_create_cpu(struct kvm_vcpu *vcpu, u16 *rc, u16 *rrc) .header.cmd = UVC_CMD_CREATE_SEC_CPU, .header.len = sizeof(uvcb), }; +void *sida_addr;int cc;if (kvm_s390_pv_cpu_get_handle (vcpu))
@@ -83,12 +84,13 @@ int kvm_s390_pv_create_cpu(struct kvm_vcpu *vcpu, u16 *rc, u16 *rrc) uvcb.stor_origin = (u64 )vcpu->arch.pv.stor_base;/* Alloc Secure Instruction Data Area Designation */ - vcpu->arch.sie_block->sidad =__get_free_page (GFP_KERNEL_ACCOUNT | __GFP_ZERO); -if (!vcpu->arch.sie_block->sidad) { + sida_addr = (void *)__get_free_page (GFP_KERNEL_ACCOUNT | __GFP_ZERO); +if (!sida_addr) {free_pages (vcpu->arch.pv.stor_base,get_order (uv_info.guest_cpu_stor_len));return -ENOMEM ; } + vcpu->arch.sie_block->sidad =virt_to_phys (sida_addr); cc =uv_call (0 , (u64)&uvcb); *rc = uvcb.header.rc;
Signature · 1 line
2.37.3
PATCH 5/5 KVM: s390: pv: sort out physical vs virtual pointers usage
In reply to Nico BoehrFix virtual vs physical address confusion (which currently are the same).
--- arch/s390/kvm/pv.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-)
arch/s390/kvm/pv.c +5 -4
@@ -80,8 +80,8 @@ int kvm_s390_pv_create_cpu(struct kvm_vcpu *vcpu, u16 *rc, u16 *rrc)/* Input */ uvcb.guest_handle =kvm_s390_pv_get_handle (vcpu->kvm); uvcb.num = vcpu->arch.sie_block->icpua; - uvcb.state_origin = (u64 )vcpu->arch.sie_block; - uvcb.stor_origin = (u64 )vcpu->arch.pv.stor_base; + uvcb.state_origin =virt_to_phys (vcpu->arch.sie_block); + uvcb.stor_origin =virt_to_phys ((void *)vcpu->arch.pv.stor_base);/* Alloc Secure Instruction Data Area Designation */ sida_addr = (void *)__get_free_page (GFP_KERNEL_ACCOUNT | __GFP_ZERO);
@@ -228,8 +228,9 @@ int kvm_s390_pv_init_vm(struct kvm *kvm, u16 *rc, u16 *rrc) uvcb.guest_stor_origin =0 ;/* MSO is 0 for KVM */ uvcb.guest_stor_len = kvm->arch.pv.guest_len; uvcb.guest_asce = kvm->arch.gmap->asce; - uvcb.guest_sca = (unsigned long )kvm->arch.sca; - uvcb.conf_base_stor_origin = (u64 )kvm->arch.pv.stor_base; + uvcb.guest_sca =virt_to_phys (kvm->arch.sca); + uvcb.conf_base_stor_origin = +virt_to_phys ((void *)kvm->arch.pv.stor_base); uvcb.conf_virt_stor_origin = (u64 )kvm->arch.pv.stor_var; cc =uv_call_sched (0 , (u64)&uvcb);
Signature · 1 line
2.37.3