diff options
Diffstat (limited to 'arch/sparc64/mm/ultra.S')
| -rw-r--r-- | arch/sparc64/mm/ultra.S | 374 |
1 files changed, 293 insertions, 81 deletions
diff --git a/arch/sparc64/mm/ultra.S b/arch/sparc64/mm/ultra.S index e4c9151fa116..f8479fad4047 100644 --- a/arch/sparc64/mm/ultra.S +++ b/arch/sparc64/mm/ultra.S @@ -15,6 +15,7 @@ #include <asm/head.h> #include <asm/thread_info.h> #include <asm/cacheflush.h> +#include <asm/hypervisor.h> /* Basically, most of the Spitfire vs. Cheetah madness * has to do with the fact that Cheetah does not support @@ -29,16 +30,18 @@ .text .align 32 .globl __flush_tlb_mm -__flush_tlb_mm: /* %o0=(ctx & TAG_CONTEXT_BITS), %o1=SECONDARY_CONTEXT */ +__flush_tlb_mm: /* 18 insns */ + /* %o0=(ctx & TAG_CONTEXT_BITS), %o1=SECONDARY_CONTEXT */ ldxa [%o1] ASI_DMMU, %g2 cmp %g2, %o0 bne,pn %icc, __spitfire_flush_tlb_mm_slow mov 0x50, %g3 stxa %g0, [%g3] ASI_DMMU_DEMAP stxa %g0, [%g3] ASI_IMMU_DEMAP + sethi %hi(KERNBASE), %g3 + flush %g3 retl - flush %g6 - nop + nop nop nop nop @@ -51,7 +54,7 @@ __flush_tlb_mm: /* %o0=(ctx & TAG_CONTEXT_BITS), %o1=SECONDARY_CONTEXT */ .align 32 .globl __flush_tlb_pending -__flush_tlb_pending: +__flush_tlb_pending: /* 26 insns */ /* %o0 = context, %o1 = nr, %o2 = vaddrs[] */ rdpr %pstate, %g7 sllx %o1, 3, %o1 @@ -72,7 +75,8 @@ __flush_tlb_pending: brnz,pt %o1, 1b nop stxa %g2, [%o4] ASI_DMMU - flush %g6 + sethi %hi(KERNBASE), %o4 + flush %o4 retl wrpr %g7, 0x0, %pstate nop @@ -82,7 +86,8 @@ __flush_tlb_pending: .align 32 .globl __flush_tlb_kernel_range -__flush_tlb_kernel_range: /* %o0=start, %o1=end */ +__flush_tlb_kernel_range: /* 16 insns */ + /* %o0=start, %o1=end */ cmp %o0, %o1 be,pn %xcc, 2f sethi %hi(PAGE_SIZE), %o4 @@ -94,8 +99,11 @@ __flush_tlb_kernel_range: /* %o0=start, %o1=end */ membar #Sync brnz,pt %o3, 1b sub %o3, %o4, %o3 -2: retl - flush %g6 +2: sethi %hi(KERNBASE), %o3 + flush %o3 + retl + nop + nop __spitfire_flush_tlb_mm_slow: rdpr %pstate, %g1 @@ -105,7 +113,8 @@ __spitfire_flush_tlb_mm_slow: stxa %g0, [%g3] ASI_IMMU_DEMAP flush %g6 stxa %g2, [%o1] ASI_DMMU - flush %g6 + sethi %hi(KERNBASE), %o1 + flush %o1 retl wrpr %g1, 0, %pstate @@ -181,7 +190,7 @@ __flush_dcache_page: /* %o0=kaddr, %o1=flush_icache */ .previous /* Cheetah specific versions, patched at boot time. */ -__cheetah_flush_tlb_mm: /* 18 insns */ +__cheetah_flush_tlb_mm: /* 19 insns */ rdpr %pstate, %g7 andn %g7, PSTATE_IE, %g2 wrpr %g2, 0x0, %pstate @@ -196,12 +205,13 @@ __cheetah_flush_tlb_mm: /* 18 insns */ stxa %g0, [%g3] ASI_DMMU_DEMAP stxa %g0, [%g3] ASI_IMMU_DEMAP stxa %g2, [%o2] ASI_DMMU - flush %g6 + sethi %hi(KERNBASE), %o2 + flush %o2 wrpr %g0, 0, %tl retl wrpr %g7, 0x0, %pstate -__cheetah_flush_tlb_pending: /* 26 insns */ +__cheetah_flush_tlb_pending: /* 27 insns */ /* %o0 = context, %o1 = nr, %o2 = vaddrs[] */ rdpr %pstate, %g7 sllx %o1, 3, %o1 @@ -225,7 +235,8 @@ __cheetah_flush_tlb_pending: /* 26 insns */ brnz,pt %o1, 1b nop stxa %g2, [%o4] ASI_DMMU - flush %g6 + sethi %hi(KERNBASE), %o4 + flush %o4 wrpr %g0, 0, %tl retl wrpr %g7, 0x0, %pstate @@ -245,7 +256,76 @@ __cheetah_flush_dcache_page: /* 11 insns */ nop #endif /* DCACHE_ALIASING_POSSIBLE */ -cheetah_patch_one: + /* Hypervisor specific versions, patched at boot time. */ +__hypervisor_tlb_tl0_error: + save %sp, -192, %sp + mov %i0, %o0 + call hypervisor_tlbop_error + mov %i1, %o1 + ret + restore + +__hypervisor_flush_tlb_mm: /* 10 insns */ + mov %o0, %o2 /* ARG2: mmu context */ + mov 0, %o0 /* ARG0: CPU lists unimplemented */ + mov 0, %o1 /* ARG1: CPU lists unimplemented */ + mov HV_MMU_ALL, %o3 /* ARG3: flags */ + mov HV_FAST_MMU_DEMAP_CTX, %o5 + ta HV_FAST_TRAP + brnz,pn %o0, __hypervisor_tlb_tl0_error + mov HV_FAST_MMU_DEMAP_CTX, %o1 + retl + nop + +__hypervisor_flush_tlb_pending: /* 16 insns */ + /* %o0 = context, %o1 = nr, %o2 = vaddrs[] */ + sllx %o1, 3, %g1 + mov %o2, %g2 + mov %o0, %g3 +1: sub %g1, (1 << 3), %g1 + ldx [%g2 + %g1], %o0 /* ARG0: vaddr + IMMU-bit */ + mov %g3, %o1 /* ARG1: mmu context */ + mov HV_MMU_ALL, %o2 /* ARG2: flags */ + srlx %o0, PAGE_SHIFT, %o0 + sllx %o0, PAGE_SHIFT, %o0 + ta HV_MMU_UNMAP_ADDR_TRAP + brnz,pn %o0, __hypervisor_tlb_tl0_error + mov HV_MMU_UNMAP_ADDR_TRAP, %o1 + brnz,pt %g1, 1b + nop + retl + nop + +__hypervisor_flush_tlb_kernel_range: /* 16 insns */ + /* %o0=start, %o1=end */ + cmp %o0, %o1 + be,pn %xcc, 2f + sethi %hi(PAGE_SIZE), %g3 + mov %o0, %g1 + sub %o1, %g1, %g2 + sub %g2, %g3, %g2 +1: add %g1, %g2, %o0 /* ARG0: virtual address */ + mov 0, %o1 /* ARG1: mmu context */ + mov HV_MMU_ALL, %o2 /* ARG2: flags */ + ta HV_MMU_UNMAP_ADDR_TRAP + brnz,pn %o0, __hypervisor_tlb_tl0_error + mov HV_MMU_UNMAP_ADDR_TRAP, %o1 + brnz,pt %g2, 1b + sub %g2, %g3, %g2 +2: retl + nop + +#ifdef DCACHE_ALIASING_POSSIBLE + /* XXX Niagara and friends have an 8K cache, so no aliasing is + * XXX possible, but nothing explicit in the Hypervisor API + * XXX guarantees this. + */ +__hypervisor_flush_dcache_page: /* 2 insns */ + retl + nop +#endif + +tlb_patch_one: 1: lduw [%o1], %g1 stw %g1, [%o0] flush %o0 @@ -264,22 +344,22 @@ cheetah_patch_cachetlbops: or %o0, %lo(__flush_tlb_mm), %o0 sethi %hi(__cheetah_flush_tlb_mm), %o1 or %o1, %lo(__cheetah_flush_tlb_mm), %o1 - call cheetah_patch_one - mov 18, %o2 + call tlb_patch_one + mov 19, %o2 sethi %hi(__flush_tlb_pending), %o0 or %o0, %lo(__flush_tlb_pending), %o0 sethi %hi(__cheetah_flush_tlb_pending), %o1 or %o1, %lo(__cheetah_flush_tlb_pending), %o1 - call cheetah_patch_one - mov 26, %o2 + call tlb_patch_one + mov 27, %o2 #ifdef DCACHE_ALIASING_POSSIBLE sethi %hi(__flush_dcache_page), %o0 or %o0, %lo(__flush_dcache_page), %o0 sethi %hi(__cheetah_flush_dcache_page), %o1 or %o1, %lo(__cheetah_flush_dcache_page), %o1 - call cheetah_patch_one + call tlb_patch_one mov 11, %o2 #endif /* DCACHE_ALIASING_POSSIBLE */ @@ -295,16 +375,14 @@ cheetah_patch_cachetlbops: * %g1 address arg 1 (tlb page and range flushes) * %g7 address arg 2 (tlb range flush only) * - * %g6 ivector table, don't touch - * %g2 scratch 1 - * %g3 scratch 2 - * %g4 scratch 3 - * - * TODO: Make xcall TLB range flushes use the tricks above... -DaveM + * %g6 scratch 1 + * %g2 scratch 2 + * %g3 scratch 3 + * %g4 scratch 4 */ .align 32 .globl xcall_flush_tlb_mm -xcall_flush_tlb_mm: +xcall_flush_tlb_mm: /* 21 insns */ mov PRIMARY_CONTEXT, %g2 ldxa [%g2] ASI_DMMU, %g3 srlx %g3, CTX_PGSZ1_NUC_SHIFT, %g4 @@ -316,9 +394,19 @@ xcall_flush_tlb_mm: stxa %g0, [%g4] ASI_IMMU_DEMAP stxa %g3, [%g2] ASI_DMMU retry + nop + nop + nop + nop + nop + nop + nop + nop + nop + nop .globl xcall_flush_tlb_pending -xcall_flush_tlb_pending: +xcall_flush_tlb_pending: /* 21 insns */ /* %g5=context, %g1=nr, %g7=vaddrs[] */ sllx %g1, 3, %g1 mov PRIMARY_CONTEXT, %g4 @@ -341,9 +429,10 @@ xcall_flush_tlb_pending: nop stxa %g2, [%g4] ASI_DMMU retry + nop .globl xcall_flush_tlb_kernel_range -xcall_flush_tlb_kernel_range: +xcall_flush_tlb_kernel_range: /* 25 insns */ sethi %hi(PAGE_SIZE - 1), %g2 or %g2, %lo(PAGE_SIZE - 1), %g2 andn %g1, %g2, %g1 @@ -360,14 +449,30 @@ xcall_flush_tlb_kernel_range: retry nop nop + nop + nop + nop + nop + nop + nop + nop + nop + nop /* This runs in a very controlled environment, so we do * not need to worry about BH races etc. */ .globl xcall_sync_tick xcall_sync_tick: - rdpr %pstate, %g2 + +661: rdpr %pstate, %g2 wrpr %g2, PSTATE_IG | PSTATE_AG, %pstate + .section .sun4v_2insn_patch, "ax" + .word 661b + nop + nop + .previous + rdpr %pil, %g2 wrpr %g0, 15, %pil sethi %hi(109f), %g7 @@ -390,8 +495,15 @@ xcall_sync_tick: */ .globl xcall_report_regs xcall_report_regs: - rdpr %pstate, %g2 + +661: rdpr %pstate, %g2 wrpr %g2, PSTATE_IG | PSTATE_AG, %pstate + .section .sun4v_2insn_patch, "ax" + .word 661b + nop + nop + .previous + rdpr %pil, %g2 wrpr %g0, 15, %pil sethi %hi(109f), %g7 @@ -453,62 +565,96 @@ xcall_flush_dcache_page_spitfire: /* %g1 == physical page address nop nop - .data - -errata32_hwbug: - .xword 0 - - .text - - /* These two are not performance critical... */ - .globl xcall_flush_tlb_all_spitfire -xcall_flush_tlb_all_spitfire: - /* Spitfire Errata #32 workaround. */ - sethi %hi(errata32_hwbug), %g4 - stx %g0, [%g4 + %lo(errata32_hwbug)] - - clr %g2 - clr %g3 -1: ldxa [%g3] ASI_DTLB_DATA_ACCESS, %g4 - and %g4, _PAGE_L, %g5 - brnz,pn %g5, 2f - mov TLB_TAG_ACCESS, %g7 - - stxa %g0, [%g7] ASI_DMMU - membar #Sync - stxa %g0, [%g3] ASI_DTLB_DATA_ACCESS + /* %g5: error + * %g6: tlb op + */ +__hypervisor_tlb_xcall_error: + mov %g5, %g4 + mov %g6, %g5 + ba,pt %xcc, etrap + rd %pc, %g7 + mov %l4, %o0 + call hypervisor_tlbop_error_xcall + mov %l5, %o1 + ba,a,pt %xcc, rtrap_clr_l6 + + .globl __hypervisor_xcall_flush_tlb_mm +__hypervisor_xcall_flush_tlb_mm: /* 21 insns */ + /* %g5=ctx, g1,g2,g3,g4,g7=scratch, %g6=unusable */ + mov %o0, %g2 + mov %o1, %g3 + mov %o2, %g4 + mov %o3, %g1 + mov %o5, %g7 + clr %o0 /* ARG0: CPU lists unimplemented */ + clr %o1 /* ARG1: CPU lists unimplemented */ + mov %g5, %o2 /* ARG2: mmu context */ + mov HV_MMU_ALL, %o3 /* ARG3: flags */ + mov HV_FAST_MMU_DEMAP_CTX, %o5 + ta HV_FAST_TRAP + mov HV_FAST_MMU_DEMAP_CTX, %g6 + brnz,pn %o0, __hypervisor_tlb_xcall_error + mov %o0, %g5 + mov %g2, %o0 + mov %g3, %o1 + mov %g4, %o2 + mov %g1, %o3 + mov %g7, %o5 membar #Sync + retry - /* Spitfire Errata #32 workaround. */ - sethi %hi(errata32_hwbug), %g4 - stx %g0, [%g4 + %lo(errata32_hwbug)] - -2: ldxa [%g3] ASI_ITLB_DATA_ACCESS, %g4 - and %g4, _PAGE_L, %g5 - brnz,pn %g5, 2f - mov TLB_TAG_ACCESS, %g7 - - stxa %g0, [%g7] ASI_IMMU - membar #Sync - stxa %g0, [%g3] ASI_ITLB_DATA_ACCESS + .globl __hypervisor_xcall_flush_tlb_pending +__hypervisor_xcall_flush_tlb_pending: /* 21 insns */ + /* %g5=ctx, %g1=nr, %g7=vaddrs[], %g2,%g3,%g4,g6=scratch */ + sllx %g1, 3, %g1 + mov %o0, %g2 + mov %o1, %g3 + mov %o2, %g4 +1: sub %g1, (1 << 3), %g1 + ldx [%g7 + %g1], %o0 /* ARG0: virtual address */ + mov %g5, %o1 /* ARG1: mmu context */ + mov HV_MMU_ALL, %o2 /* ARG2: flags */ + srlx %o0, PAGE_SHIFT, %o0 + sllx %o0, PAGE_SHIFT, %o0 + ta HV_MMU_UNMAP_ADDR_TRAP + mov HV_MMU_UNMAP_ADDR_TRAP, %g6 + brnz,a,pn %o0, __hypervisor_tlb_xcall_error + mov %o0, %g5 + brnz,pt %g1, 1b + nop + mov %g2, %o0 + mov %g3, %o1 + mov %g4, %o2 membar #Sync - - /* Spitfire Errata #32 workaround. */ - sethi %hi(errata32_hwbug), %g4 - stx %g0, [%g4 + %lo(errata32_hwbug)] - -2: add %g2, 1, %g2 - cmp %g2, SPITFIRE_HIGHEST_LOCKED_TLBENT - ble,pt %icc, 1b - sll %g2, 3, %g3 - flush %g6 retry - .globl xcall_flush_tlb_all_cheetah -xcall_flush_tlb_all_cheetah: - mov 0x80, %g2 - stxa %g0, [%g2] ASI_DMMU_DEMAP - stxa %g0, [%g2] ASI_IMMU_DEMAP + .globl __hypervisor_xcall_flush_tlb_kernel_range +__hypervisor_xcall_flush_tlb_kernel_range: /* 25 insns */ + /* %g1=start, %g7=end, g2,g3,g4,g5,g6=scratch */ + sethi %hi(PAGE_SIZE - 1), %g2 + or %g2, %lo(PAGE_SIZE - 1), %g2 + andn %g1, %g2, %g1 + andn %g7, %g2, %g7 + sub %g7, %g1, %g3 + add %g2, 1, %g2 + sub %g3, %g2, %g3 + mov %o0, %g2 + mov %o1, %g4 + mov %o2, %g7 +1: add %g1, %g3, %o0 /* ARG0: virtual address */ + mov 0, %o1 /* ARG1: mmu context */ + mov HV_MMU_ALL, %o2 /* ARG2: flags */ + ta HV_MMU_UNMAP_ADDR_TRAP + mov HV_MMU_UNMAP_ADDR_TRAP, %g6 + brnz,pn %o0, __hypervisor_tlb_xcall_error + mov %o0, %g5 + sethi %hi(PAGE_SIZE), %o2 + brnz,pt %g3, 1b + sub %g3, %o2, %g3 + mov %g2, %o0 + mov %g4, %o1 + mov %g7, %o2 + membar #Sync retry /* These just get rescheduled to PIL vectors. */ @@ -527,4 +673,70 @@ xcall_capture: wr %g0, (1 << PIL_SMP_CAPTURE), %set_softint retry + .globl xcall_new_mmu_context_version +xcall_new_mmu_context_version: + wr %g0, (1 << PIL_SMP_CTX_NEW_VERSION), %set_softint + retry + #endif /* CONFIG_SMP */ + + + .globl hypervisor_patch_cachetlbops +hypervisor_patch_cachetlbops: + save %sp, -128, %sp + + sethi %hi(__flush_tlb_mm), %o0 + or %o0, %lo(__flush_tlb_mm), %o0 + sethi %hi(__hypervisor_flush_tlb_mm), %o1 + or %o1, %lo(__hypervisor_flush_tlb_mm), %o1 + call tlb_patch_one + mov 10, %o2 + + sethi %hi(__flush_tlb_pending), %o0 + or %o0, %lo(__flush_tlb_pending), %o0 + sethi %hi(__hypervisor_flush_tlb_pending), %o1 + or %o1, %lo(__hypervisor_flush_tlb_pending), %o1 + call tlb_patch_one + mov 16, %o2 + + sethi %hi(__flush_tlb_kernel_range), %o0 + or %o0, %lo(__flush_tlb_kernel_range), %o0 + sethi %hi(__hypervisor_flush_tlb_kernel_range), %o1 + or %o1, %lo(__hypervisor_flush_tlb_kernel_range), %o1 + call tlb_patch_one + mov 16, %o2 + +#ifdef DCACHE_ALIASING_POSSIBLE + sethi %hi(__flush_dcache_page), %o0 + or %o0, %lo(__flush_dcache_page), %o0 + sethi %hi(__hypervisor_flush_dcache_page), %o1 + or %o1, %lo(__hypervisor_flush_dcache_page), %o1 + call tlb_patch_one + mov 2, %o2 +#endif /* DCACHE_ALIASING_POSSIBLE */ + +#ifdef CONFIG_SMP + sethi %hi(xcall_flush_tlb_mm), %o0 + or %o0, %lo(xcall_flush_tlb_mm), %o0 + sethi %hi(__hypervisor_xcall_flush_tlb_mm), %o1 + or %o1, %lo(__hypervisor_xcall_flush_tlb_mm), %o1 + call tlb_patch_one + mov 21, %o2 + + sethi %hi(xcall_flush_tlb_pending), %o0 + or %o0, %lo(xcall_flush_tlb_pending), %o0 + sethi %hi(__hypervisor_xcall_flush_tlb_pending), %o1 + or %o1, %lo(__hypervisor_xcall_flush_tlb_pending), %o1 + call tlb_patch_one + mov 21, %o2 + + sethi %hi(xcall_flush_tlb_kernel_range), %o0 + or %o0, %lo(xcall_flush_tlb_kernel_range), %o0 + sethi %hi(__hypervisor_xcall_flush_tlb_kernel_range), %o1 + or %o1, %lo(__hypervisor_xcall_flush_tlb_kernel_range), %o1 + call tlb_patch_one + mov 25, %o2 +#endif /* CONFIG_SMP */ + + ret + restore |