The current vmrunkernel sets up 1 Gbyte of 2M pages. It is best if we can start the kernel with the minimum amount of page tables set up. That way, if some very bad addressing happens in early assembly, we'll see it right away as a triple fault.
We now only set up enough 2M pages to cover the size of the file we read in. We had thought we might want to dial it back to 4k pages, but it turns out guests are pretty careful not to turn off paging attributes that are enabled; rather, they carefully set up page tables and then enable what they need, not disabling anything that might have been set up. Change-Id: I1c9f458f6147e81cd7bf0c51745167a6ff17db79 Signed-off-by: Ronald G. Minnich <[email protected]> --- tests/vmm/vmrunkernel.c | 47 ++++++++++++++++++++++++++--------------------- 1 file changed, 26 insertions(+), 21 deletions(-) diff --git a/tests/vmm/vmrunkernel.c b/tests/vmm/vmrunkernel.c index 116baed..61a9bb1 100644 --- a/tests/vmm/vmrunkernel.c +++ b/tests/vmm/vmrunkernel.c @@ -335,6 +335,7 @@ int main(int argc, char **argv) int vmmflags = 0; // Disabled probably forever. VMM_VMCALL_PRINTF; uint64_t entry = 0x1200000, kerneladdress = 0x1200000; int ret; + uintptr_t size; void * xp; int kfd = -1; static char cmd[512]; @@ -429,10 +430,10 @@ int main(int argc, char **argv) perror(argv[0]); exit(1); } - // read in the kernel. + // read in the kernel, one 2M page at a time. xp = (void *)kerneladdress; for(;;) { - amt = read(kfd, xp, 1048576); + amt = read(kfd, xp, PML2_PTE_REACH); if (amt < 0) { perror("read"); exit(1); @@ -442,7 +443,8 @@ int main(int argc, char **argv) } xp += amt; } - fprintf(stderr, "Read in %d bytes\n", xp-kerneladdress); + size = ROUNDUP((uintptr_t)xp-kerneladdress, PML2_PTE_REACH); + fprintf(stderr, "Read in %d bytes\n", size); close(kfd); // The low 1m so we can fill in bullshit like ACPI. */ @@ -626,33 +628,36 @@ int main(int argc, char **argv) ret = vmm_init(vm, vmmflags); assert(!ret); - - ret = posix_memalign((void **)&p512, 4096, 3*4096); - fprintf(stderr, "memalign is %p\n", p512); + /* Allocate 3 pages for page table pages: a page of 512 GiB + * PTEs with only one entry filled to point to a page of 1 GiB + * PTEs; a page of 1 GiB PTEs with only one entry filled to + * point to a page of 2 MiB PTEs; and a page of 2 MiB PTEs, + * only a subset of which will be filled. */ + ret = posix_memalign((void **)&p512, PGSIZE, 3*PGSIZE); if (ret) { perror("ptp alloc"); exit(1); } - p1 = &p512[512]; - p2m = &p512[1024]; - uint64_t kernbase = 0; //0xffffffff80000000; - uint64_t highkernbase = 0xffffffff80000000; - p512[PML4(kernbase)] = (unsigned long long)p1 | 7; - p1[PML3(kernbase)] = /*0x87; */(unsigned long long)p2m | 7; - p512[PML4(highkernbase)] = (unsigned long long)p1 | 7; - p1[PML3(highkernbase)] = /*0x87; */(unsigned long long)p2m | 7; -#define _2MiB (0x200000) - - for (i = 0; i < 512; i++) { - p2m[PML2(kernbase + i * _2MiB)] = 0x87 | i * _2MiB; + + /* Set up a 1:1 page mapping from physical to virtual + * addresses derived from the (host virtual) + * `kerneladdress`. Be aware that the values stored in the + * table are physical addresses. This is subtle and mistakes + * are easily disguised due to the identity mapping, so take + * care when manipulating these mappings. */ + p1 = &p512[NPTENTRIES]; + p2m = &p512[2*NPTENTRIES]; + + p512[PML4(kerneladdress)] = (uint64_t)p1 | PTE_KERN_RW; + p1[PML3(kerneladdress)] = (uint64_t)p2m | PTE_KERN_RW; + for (uintptr_t i = 0; i < size; i += PML2_PTE_REACH) { + p2m[PML2(kerneladdress + i)] = + (uint64_t)(kerneladdress + i) | PTE_KERN_RW | PTE_PS; } - kernbase >>= (0+12); - kernbase <<= (0 + 12); uint8_t *kernel = (void *)GKERNBASE; //write_coreboot_table(coreboot_tables, ((void *)VIRTIOBASE) /*kernel*/, KERNSIZE + 1048576); hexdump(stdout, coreboot_tables, 512); - fprintf(stderr, "kernbase for pml4 is 0x%llx and entry is %llx\n", kernbase, entry); fprintf(stderr, "p512 %p p512[0] is 0x%lx p1 %p p1[0] is 0x%x\n", p512, p512[0], p1, p1[0]); vmm_run_task(vm, timer_thread, 0); -- 2.8.0.rc3.226.g39d4020 -- You received this message because you are subscribed to the Google Groups "Akaros" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. For more options, visit https://groups.google.com/d/optout.
