| Message ID | 20221207214809.489070-4-oliver.upton@linux.dev (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | KVM: selftests: Fixes for ucall pool + page_fault_test | expand |
On Wed, Dec 07, 2022, Oliver Upton wrote: > diff --git a/tools/testing/selftests/kvm/lib/aarch64/processor.c b/tools/testing/selftests/kvm/lib/aarch64/processor.c > index 316de70db91d..5972a23b2765 100644 > --- a/tools/testing/selftests/kvm/lib/aarch64/processor.c > +++ b/tools/testing/selftests/kvm/lib/aarch64/processor.c > @@ -541,3 +541,13 @@ void kvm_selftest_arch_init(void) > */ > guest_modes_append_default(); > } > + > +void vm_vaddr_populate_bitmap(struct kvm_vm *vm) Add "arch" so that it's obvious this can be overidden? The "__weak" conveys that for the implementation, but not for the call site. E.g. vm_arch_vaddr_populate_bitmap(). Actually, IIUC, the issue is that the high half isn't mapped (probably the wrong terminology). I.e. the calculation for the low half stays the same, and the high half just goes away. > +{ > + /* > + * arm64 selftests use only TTBR0_EL1, meaning that the valid VA space > + * is [0, 2^(64 - TCR_EL1.T0SZ)). > + */ > + sparsebit_set_num(vm->vpages_valid, 0, > + (1ULL << vm->va_bits) >> vm->page_shift); > +} > diff --git a/tools/testing/selftests/kvm/lib/kvm_util.c b/tools/testing/selftests/kvm/lib/kvm_util.c > index e9607eb089be..c88c3ace16d2 100644 > --- a/tools/testing/selftests/kvm/lib/kvm_util.c > +++ b/tools/testing/selftests/kvm/lib/kvm_util.c > @@ -186,6 +186,15 @@ const struct vm_guest_mode_params vm_guest_mode_params[] = { > _Static_assert(sizeof(vm_guest_mode_params)/sizeof(struct vm_guest_mode_params) == NUM_VM_MODES, > "Missing new mode params?"); > > +__weak void vm_vaddr_populate_bitmap(struct kvm_vm *vm) > +{ > + sparsebit_set_num(vm->vpages_valid, > + 0, (1ULL << (vm->va_bits - 1)) >> vm->page_shift); > + sparsebit_set_num(vm->vpages_valid, > + (~((1ULL << (vm->va_bits - 1)) - 1)) >> vm->page_shift, > + (1ULL << (vm->va_bits - 1)) >> vm->page_shift); Any objection to fixing up the formatting? Actually, we can do more than just fix the indentation, e.g. the number of bits is identical, and documenting that this does a high/low split would be helpful. Together, what about? The #ifdef is a bit gross, especially around "hi_start", but it's less duplicate code. And IMO, having things bundled in the same place makes it a lot easier for newbies (to arm64 or kernel coding in general) to understand what's going on and why arm64 is different. --- tools/testing/selftests/kvm/lib/kvm_util.c | 23 +++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-) diff --git a/tools/testing/selftests/kvm/lib/kvm_util.c b/tools/testing/selftests/kvm/lib/kvm_util.c index e9607eb089be..d6f2c17e3d40 100644 --- a/tools/testing/selftests/kvm/lib/kvm_util.c +++ b/tools/testing/selftests/kvm/lib/kvm_util.c @@ -186,6 +186,23 @@ const struct vm_guest_mode_params vm_guest_mode_params[] = { _Static_assert(sizeof(vm_guest_mode_params)/sizeof(struct vm_guest_mode_params) == NUM_VM_MODES, "Missing new mode params?"); +static void vm_vaddr_populate_bitmap(struct kvm_vm *vm) +{ + /* + * All architectures supports splitting the virtual address space into + * a high and a low half. Populate both halves, except for arm64 which + * currently uses only TTBR0_EL1 (arbitrary selftests "logic"), i.e. + * only has a valid low half. + */ + sparsebit_num_t nr_va_bits = (1ULL << (vm->va_bits - 1)) >> vm->page_shift; +#ifndef __aarch64__ + sparsebit_num_t hi_start = (~((1ULL << (vm->va_bits - 1)) - 1)) >> vm->page_shift + + sparsebit_set_num(vm->vpages_valid, hi_start, nr_bits); +#endif + sparsebit_set_num(vm->vpages_valid, 0, nr_va_bits); +} + struct kvm_vm *____vm_create(enum vm_guest_mode mode) { struct kvm_vm *vm; @@ -274,11 +291,7 @@ struct kvm_vm *____vm_create(enum vm_guest_mode mode) /* Limit to VA-bit canonical virtual addresses. */ vm->vpages_valid = sparsebit_alloc(); - sparsebit_set_num(vm->vpages_valid, - 0, (1ULL << (vm->va_bits - 1)) >> vm->page_shift); - sparsebit_set_num(vm->vpages_valid, - (~((1ULL << (vm->va_bits - 1)) - 1)) >> vm->page_shift, - (1ULL << (vm->va_bits - 1)) >> vm->page_shift); + vm_vaddr_populate_bitmap(vm); /* Limit physical addresses to PA-bits. */ vm->max_gfn = vm_compute_max_gfn(vm); base-commit: 35aecc3289eebf193fd70a067ea448ae2f0bb9b9 --
On Thu, Dec 08, 2022 at 12:18:07AM +0000, Sean Christopherson wrote: [...] > Together, what about? The #ifdef is a bit gross, especially around "hi_start", > but it's less duplicate code. And IMO, having things bundled in the same place > makes it a lot easier for newbies (to arm64 or kernel coding in general) to > understand what's going on and why arm64 is different. I'd rather we not go this route. We really shouldn't make any attempt to de-dupe something that is inherently architecture specific. For example: > + /* > + * All architectures supports splitting the virtual address space into > + * a high and a low half. Populate both halves, except for arm64 which > + * currently uses only TTBR0_EL1 (arbitrary selftests "logic"), i.e. > + * only has a valid low half. > + */ > + sparsebit_num_t nr_va_bits = (1ULL << (vm->va_bits - 1)) >> vm->page_shift; This is still wrong for arm64. When we say the VA space is 48 bits, we really do mean that TTBR0 is able to address a full 48 bits. So this truncates the MSB for the addressing mode. With the code living in the arm64 side of the shop, I can also tailor the comment to directly match the architecture to provide breadcrumbs tying it back to the Arm ARM. -- Thanks, Oliver
On Thu, Dec 08, 2022, Oliver Upton wrote: > On Thu, Dec 08, 2022 at 12:18:07AM +0000, Sean Christopherson wrote: > > [...] > > > Together, what about? The #ifdef is a bit gross, especially around "hi_start", > > but it's less duplicate code. And IMO, having things bundled in the same place > > makes it a lot easier for newbies (to arm64 or kernel coding in general) to > > understand what's going on and why arm64 is different. > > I'd rather we not go this route. We really shouldn't make any attempt to > de-dupe something that is inherently architecture specific. > > For example: > > > + /* > > + * All architectures supports splitting the virtual address space into > > + * a high and a low half. Populate both halves, except for arm64 which > > + * currently uses only TTBR0_EL1 (arbitrary selftests "logic"), i.e. > > + * only has a valid low half. > > + */ > > + sparsebit_num_t nr_va_bits = (1ULL << (vm->va_bits - 1)) >> vm->page_shift; > > This is still wrong for arm64. When we say the VA space is 48 bits, we > really do mean that TTBR0 is able to address a full 48 bits. So this > truncates the MSB for the addressing mode. Ah, I missed the lack of a "-1" in the arm64 code. > With the code living in the arm64 side of the shop, I can also tailor > the comment to directly match the architecture to provide breadcrumbs > tying it back to the Arm ARM. The main reason why I don't like splitting the code this way is that it makes it harder for non-arm64 folks to understand what makes arm64 different. Case in point, my overlooking of the "-1". I read the changelog and the comment and still missed that small-but-important detail, largely because I am completely unfamiliar with how TTBR{0,1}_EL1 works. Actually, before we do anything, we should get confirmation from the s390 and RISC-V folks on whether they have a canonical hole like x86, i.e. maybe x86 is the oddball. Anyways, assuming one architecture is the oddball (I'm betting it's x86), I have no objection to bleeding some of the details into the common code, including a large comment to document the gory details. If every architecture manges to be different, then yeah, a hook is probably warranted. That said, I also don't mind shoving a bit of abstraction into arch code if that avoids some #ifdef ugliness or allows for better documentation, flexibility, etc. What I don't like is duplicating the logic of turning "VA bits" into the bitmap. E.g. something like this would also be an option. Readers would obviously need to track down has_split_va_space, but that should be fairly easy and can come with a big arch-specific comment, and meanwhile the core logic of how selftests populate the va bitmaps is common. Or if arm64 is the only arch without a split, invert the flag and have arm64 set the vm->has_combined_va_space or whatever. static void vm_vaddr_populate_bitmap(struct kvm_vm *vm) { unsigned int eff_va_bits = vm->va_bits; sparsebit_num_t nr_bits; /* blah blah blah */ if (vm->has_split_va_space) eff_va_bits--; nr_bits = (1ULL << eff_va_bits) >> vm->page_shift; sparsebit_set_num(vm->vpages_valid, 0, nr_va_bits); if (vm->has_split_va_space) sparsebit_set_num(vm->vpages_valid, (~((1ULL << eff_va_bits) - 1)) >> vm->page_shift, nr_bits); }
On Thu, Dec 08, 2022 at 01:09:38AM +0000, Sean Christopherson wrote: ... > Actually, before we do anything, we should get confirmation from the s390 and > RISC-V folks on whether they have a canonical hole like x86, i.e. maybe x86 is > the oddball. riscv splits like x86. Thanks, drew
diff --git a/tools/testing/selftests/kvm/include/kvm_util_base.h b/tools/testing/selftests/kvm/include/kvm_util_base.h index 6cd86da698b3..fbc2a79369b8 100644 --- a/tools/testing/selftests/kvm/include/kvm_util_base.h +++ b/tools/testing/selftests/kvm/include/kvm_util_base.h @@ -420,6 +420,7 @@ void vm_mem_region_set_flags(struct kvm_vm *vm, uint32_t slot, uint32_t flags); void vm_mem_region_move(struct kvm_vm *vm, uint32_t slot, uint64_t new_gpa); void vm_mem_region_delete(struct kvm_vm *vm, uint32_t slot); struct kvm_vcpu *__vm_vcpu_add(struct kvm_vm *vm, uint32_t vcpu_id); +void vm_populate_vaddr_bitmap(struct kvm_vm *vm); vm_vaddr_t vm_vaddr_unused_gap(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_min); vm_vaddr_t vm_vaddr_alloc(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_min); vm_vaddr_t __vm_vaddr_alloc(struct kvm_vm *vm, size_t sz, vm_vaddr_t vaddr_min, diff --git a/tools/testing/selftests/kvm/lib/aarch64/processor.c b/tools/testing/selftests/kvm/lib/aarch64/processor.c index 316de70db91d..5972a23b2765 100644 --- a/tools/testing/selftests/kvm/lib/aarch64/processor.c +++ b/tools/testing/selftests/kvm/lib/aarch64/processor.c @@ -541,3 +541,13 @@ void kvm_selftest_arch_init(void) */ guest_modes_append_default(); } + +void vm_vaddr_populate_bitmap(struct kvm_vm *vm) +{ + /* + * arm64 selftests use only TTBR0_EL1, meaning that the valid VA space + * is [0, 2^(64 - TCR_EL1.T0SZ)). + */ + sparsebit_set_num(vm->vpages_valid, 0, + (1ULL << vm->va_bits) >> vm->page_shift); +} diff --git a/tools/testing/selftests/kvm/lib/kvm_util.c b/tools/testing/selftests/kvm/lib/kvm_util.c index e9607eb089be..c88c3ace16d2 100644 --- a/tools/testing/selftests/kvm/lib/kvm_util.c +++ b/tools/testing/selftests/kvm/lib/kvm_util.c @@ -186,6 +186,15 @@ const struct vm_guest_mode_params vm_guest_mode_params[] = { _Static_assert(sizeof(vm_guest_mode_params)/sizeof(struct vm_guest_mode_params) == NUM_VM_MODES, "Missing new mode params?"); +__weak void vm_vaddr_populate_bitmap(struct kvm_vm *vm) +{ + sparsebit_set_num(vm->vpages_valid, + 0, (1ULL << (vm->va_bits - 1)) >> vm->page_shift); + sparsebit_set_num(vm->vpages_valid, + (~((1ULL << (vm->va_bits - 1)) - 1)) >> vm->page_shift, + (1ULL << (vm->va_bits - 1)) >> vm->page_shift); +} + struct kvm_vm *____vm_create(enum vm_guest_mode mode) { struct kvm_vm *vm; @@ -274,11 +283,7 @@ struct kvm_vm *____vm_create(enum vm_guest_mode mode) /* Limit to VA-bit canonical virtual addresses. */ vm->vpages_valid = sparsebit_alloc(); - sparsebit_set_num(vm->vpages_valid, - 0, (1ULL << (vm->va_bits - 1)) >> vm->page_shift); - sparsebit_set_num(vm->vpages_valid, - (~((1ULL << (vm->va_bits - 1)) - 1)) >> vm->page_shift, - (1ULL << (vm->va_bits - 1)) >> vm->page_shift); + vm_vaddr_populate_bitmap(vm); /* Limit physical addresses to PA-bits. */ vm->max_gfn = vm_compute_max_gfn(vm);
An interesting feature of the Arm architecture is that the stage-1 MMU supports two distinct VA regions, controlled by TTBR{0,1}_EL1. As KVM selftests on arm64 only uses TTBR0_EL1, the VA space is constrained to [0, 2^(va_bits-1)). This is different from other architectures that allow for addressing low and high regions of the VA space from a single page table. KVM selftests' VA space allocator presumes the valid address range is split between low and high memory based the MSB, which of course is a poor match for arm64's TTBR0 region. Allow architectures to override the default VA space layout. Make use of the override to align vpages_valid with the behavior of TTBR0 on arm64. Signed-off-by: Oliver Upton <oliver.upton@linux.dev> --- .../testing/selftests/kvm/include/kvm_util_base.h | 1 + .../testing/selftests/kvm/lib/aarch64/processor.c | 10 ++++++++++ tools/testing/selftests/kvm/lib/kvm_util.c | 15 ++++++++++----- 3 files changed, 21 insertions(+), 5 deletions(-)