X-Git-Url: http://v3vee.org/palacios/gitweb/gitweb.cgi?p=palacios.git;a=blobdiff_plain;f=palacios%2Fsrc%2Fpalacios%2Fvmm_shadow_paging.c;h=eeb4fdaf84e566f611d6652165d8d22d48fd9e2f;hp=dbc4c61eff11a1a0fae3eb33d796dcc1a4cc985b;hb=bf091be03d4a623fa1a9b3151698a4564a8339ba;hpb=d964570646e2b882848b3291d59e5cf60f568533 diff --git a/palacios/src/palacios/vmm_shadow_paging.c b/palacios/src/palacios/vmm_shadow_paging.c index dbc4c61..eeb4fda 100644 --- a/palacios/src/palacios/vmm_shadow_paging.c +++ b/palacios/src/palacios/vmm_shadow_paging.c @@ -42,6 +42,50 @@ int handle_shadow_pagefault(struct guest_info * info, addr_t fault_addr, pf_erro } } +addr_t create_new_shadow_pt32(struct guest_info * info) { + void * host_pde = 0; + + V3_AllocPages(host_pde, 1); + memset(host_pde, 0, PAGE_SIZE); + + return (addr_t)host_pde; +} + + +static int handle_pd32_nonaligned_4MB_page(struct guest_info * info, pte32_t * pt, addr_t guest_addr, pde32_4MB_t * large_shadow_pde) { + uint_t i = 0; + pte32_t * pte_cursor; + addr_t guest_pa = 0; + + for (i = 0; i < 1024; i++) { + guest_pa = guest_addr + (PAGE_SIZE * i); + host_region_type_t host_page_type = get_shadow_addr_type(info, guest_pa); + + pte_cursor = &(pt[i]); + + if (host_page_type == HOST_REGION_INVALID) { + // Currently we don't support this, but in theory we could + PrintDebug("Invalid Host Memory Type\n"); + return -1; + } else if (host_page_type == HOST_REGION_PHYSICAL_MEMORY) { + addr_t shadow_pa = get_shadow_addr(info, guest_pa); + + + pte_cursor->page_base_addr = PT32_BASE_ADDR(shadow_pa); + pte_cursor->present = 1; + pte_cursor->writable = large_shadow_pde->writable; + pte_cursor->user_page = large_shadow_pde->user_page; + pte_cursor->write_through = 0; + pte_cursor->cache_disable = 0; + pte_cursor->global_page = 0; + + } else { + PrintDebug("Unsupported Host Memory Type\n"); + return -1; + } + } + return 0; +} int handle_shadow_pagefault32(struct guest_info * info, addr_t fault_addr, pf_error_t error_code) { pde32_t * guest_pd = NULL; @@ -63,22 +107,42 @@ int handle_shadow_pagefault32(struct guest_info * info, addr_t fault_addr, pf_er // Check the guest page permissions guest_pde_access = can_access_pde32(guest_pd, fault_addr, error_code); - if (guest_pde_access != PT_ACCESS_OK) { + // Check the shadow page permissions + shadow_pde_access = can_access_pde32(shadow_pd, fault_addr, error_code); + + /* This should be redone, + but basically the reasoning is that there can be multiple reasons for a page fault: + If there is a permissions failure for a page present in the guest _BUT_ + the reason for the fault was that the page is not present in the shadow, + _THEN_ we have to map the shadow page in and reexecute, this will generate + a permissions fault which is _THEN_ valid to send to the guest + + whew... + */ + if ((guest_pde_access != PT_ACCESS_OK) && + ( (shadow_pde_access != PT_ENTRY_NOT_PRESENT) && + (guest_pde_access != PT_ENTRY_NOT_PRESENT))) { // aka (guest permission error) // inject page fault to the guest (Guest PDE fault) - info->ctrl_regs.cr2 = fault_addr; + info->ctrl_regs.cr2 = fault_addr; raise_exception_with_error(info, PF_EXCEPTION, *(uint_t *)&error_code); + PrintDebug("Injecting PDE pf to guest: (guest access error=%d) (pf error code=%d)\n", guest_pde_access, error_code); + return 0; + PrintDebug("Guest CR3=%x\n", guest_cr3); + PrintDebug("Guest PD\n"); PrintPD32(guest_pd); + PrintDebug("Shadow PD\n"); + PrintPD32(shadow_pd); - V3_ASSERT(0); - return 0; + + return -1; } - shadow_pde_access = can_access_pde32(shadow_pd, fault_addr, error_code); + //shadow_pde_access = can_access_pde32(shadow_pd, fault_addr, error_code); if (shadow_pde_access == PT_ENTRY_NOT_PRESENT) { @@ -87,13 +151,14 @@ int handle_shadow_pagefault32(struct guest_info * info, addr_t fault_addr, pf_er shadow_pde->user_page = guest_pde->user_page; shadow_pde->large_page = guest_pde->large_page; + // VMM Specific options shadow_pde->write_through = 0; shadow_pde->cache_disable = 0; shadow_pde->global_page = 0; // - guest_pde->accessed = 1; + guest_pde->accessed = 1; if (guest_pde->large_page == 0) { pte32_t * shadow_pt = NULL; @@ -113,13 +178,10 @@ int handle_shadow_pagefault32(struct guest_info * info, addr_t fault_addr, pf_er // addr_t guest_end_addr = guest_start_addr + PAGE_SIZE_4MB; // start address + 4MB - /* JRL: THIS COULD BE A PROBLEM.... - * Currently we only support large pages if the region is mapped contiguosly in shadow memory - * Lets hope this is the case... - */ - // Check that the Guest PDE entry points to valid memory // else Machine Check the guest + PrintDebug("Large Page: Page Base Addr=%x\n", guest_start_addr); + host_page_type = get_shadow_addr_type(info, guest_start_addr); if (host_page_type == HOST_REGION_INVALID) { @@ -133,17 +195,15 @@ int handle_shadow_pagefault32(struct guest_info * info, addr_t fault_addr, pf_er // Check for a large enough region in host memory mem_reg = get_shadow_region_by_addr(&(info->mem_map), guest_start_addr); + PrintDebug("Host region: host_addr=%x (guest_start=%x, end=%x)\n", + mem_reg->host_addr, mem_reg->guest_start, mem_reg->guest_end); host_start_addr = mem_reg->host_addr + (guest_start_addr - mem_reg->guest_start); region_end_addr = mem_reg->host_addr + (mem_reg->guest_end - mem_reg->guest_start); - // Check if the region is at least an additional 4MB - if (region_end_addr <= host_start_addr + PAGE_SIZE_4MB) { - PrintDebug("Large page over non contiguous host memory... Not handled\n"); - return -1; - } + PrintDebug("Host Start Addr=%x; Region End Addr=%x\n", host_start_addr, region_end_addr); + + - //4b. - large_shadow_pde->page_base_addr = PD32_4MB_BASE_ADDR(host_start_addr); //4f if (large_guest_pde->dirty == 1) { // dirty @@ -155,6 +215,38 @@ int handle_shadow_pagefault32(struct guest_info * info, addr_t fault_addr, pf_er large_shadow_pde->writable = 0; } + + // Check if the region is at least an additional 4MB + + + //4b. + if ((PD32_4MB_PAGE_OFFSET(host_start_addr) == 0) && + (region_end_addr >= host_start_addr + PAGE_SIZE_4MB)) { // if 4MB boundary + large_shadow_pde->page_base_addr = PD32_4MB_BASE_ADDR(host_start_addr); + } else { // else generate 4k pages + pte32_t * shadow_pt = NULL; + PrintDebug("Handling non aligned large page\n"); + + shadow_pde->large_page = 0; + + V3_AllocPages(shadow_pt, 1); + memset(shadow_pt, 0, PAGE_SIZE); + + if (handle_pd32_nonaligned_4MB_page(info, shadow_pt, guest_start_addr, large_shadow_pde) == -1) { + PrintDebug("Non Aligned Large Page Error\n"); + V3_Free(shadow_pt); + return -1; + } + + + /* + PrintDebug("non-aligned Shadow PT\n"); + PrintPT32(PT32_PAGE_ADDR(fault_addr), shadow_pt); + */ + shadow_pde->pt_base_addr = PD32_BASE_ADDR(shadow_pt); + } + + } else { // Handle hooked pages as well as other special pages if (handle_special_page_fault(info, fault_addr, error_code) == -1) { @@ -218,7 +310,7 @@ int handle_shadow_pagefault32(struct guest_info * info, addr_t fault_addr, pf_er } //PrintDebugPageTables(shadow_pd); - PrintDebug("Returning end of PDE function\n"); + PrintDebug("Returning end of PDE function (rip=%x)\n", info->rip); return 0; } @@ -242,19 +334,32 @@ int handle_shadow_pte32_fault(struct guest_info * info, // Check the guest page permissions guest_pte_access = can_access_pte32(guest_pt, fault_addr, error_code); + // Check the shadow page permissions + shadow_pte_access = can_access_pte32(shadow_pt, fault_addr, error_code); - if (guest_pte_access != PT_ACCESS_OK) { + /* This should be redone, + but basically the reasoning is that there can be multiple reasons for a page fault: + If there is a permissions failure for a page present in the guest _BUT_ + the reason for the fault was that the page is not present in the shadow, + _THEN_ we have to map the shadow page in and reexecute, this will generate + a permissions fault which is _THEN_ valid to send to the guest + + whew... + */ + if ((guest_pte_access != PT_ACCESS_OK) && + ((shadow_pte_access != PT_ENTRY_NOT_PRESENT) && + (guest_pte_access != PT_ENTRY_NOT_PRESENT))) { // aka (guest permission error) // Inject page fault into the guest info->ctrl_regs.cr2 = fault_addr; raise_exception_with_error(info, PF_EXCEPTION, *(uint_t *)&error_code); - PrintDebug("Access error injecting pf to guest\n"); - return 0; + PrintDebug("Access error injecting pf to guest (guest access error=%d) (pf error code=%d)\n", guest_pte_access, error_code); + return 0; } - shadow_pte_access = can_access_pte32(shadow_pt, fault_addr, error_code); + if (shadow_pte_access == PT_ACCESS_OK) { // Inconsistent state... @@ -339,14 +444,6 @@ int handle_shadow_pte32_fault(struct guest_info * info, -addr_t create_new_shadow_pt32(struct guest_info * info) { - void * host_pde = 0; - - V3_AllocPages(host_pde, 1); - memset(host_pde, 0, PAGE_SIZE); - - return (addr_t)host_pde; -} @@ -384,6 +481,17 @@ int handle_shadow_invlpg(struct guest_info * info) { addr_t first_operand; addr_t second_operand; operand_type_t addr_type; + addr_t guest_cr3 = CR3_TO_PDE32(info->shdw_pg_state.guest_cr3); + + pde32_t * guest_pd = NULL; + + if (guest_pa_to_host_va(info, guest_cr3, (addr_t*)&guest_pd) == -1) { + PrintDebug("Invalid Guest PDE Address: 0x%x\n", guest_cr3); + return -1; + } + + + index += 2; @@ -392,17 +500,25 @@ int handle_shadow_invlpg(struct guest_info * info) { if (addr_type == MEM_OPERAND) { pde32_t * shadow_pd = (pde32_t *)CR3_TO_PDE32(info->shdw_pg_state.shadow_cr3); pde32_t * shadow_pde = (pde32_t *)&shadow_pd[PDE32_INDEX(first_operand)]; + pde32_t * guest_pde; //PrintDebug("PDE Index=%d\n", PDE32_INDEX(first_operand)); //PrintDebug("FirstOperand = %x\n", first_operand); - if (shadow_pde->large_page == 1) { + PrintDebug("Invalidating page for %x\n", first_operand); + + guest_pde = (pde32_t *)&(guest_pd[PDE32_INDEX(first_operand)]); + + if (guest_pde->large_page == 1) { shadow_pde->present = 0; + PrintDebug("Invalidating Large Page\n"); } else { + if (shadow_pde->present == 1) { pte32_t * shadow_pt = (pte32_t *)PDE32_T_ADDR((*shadow_pde)); pte32_t * shadow_pte = (pte32_t *)&shadow_pt[PTE32_INDEX(first_operand)]; - + PrintDebug("Setting not present\n"); + PrintPTE32(first_operand, shadow_pte); shadow_pte->present = 0; } }