Messages in this thread Patch in this message | | | Date | Mon, 18 Jan 2021 11:06:41 +0000 | From | David Brazdil <> | Subject | Re: [PATCH v3 3/9] KVM: arm64: Enable UBSAN_BOUNDS for the both the kernel and hyp/nVHE |
| |
On Fri, Jan 15, 2021 at 05:18:24PM +0000, Elena Petrova wrote: > From: George Popescu <georgepope@google.com> > > If an out of bounds happens inside the hyp/nVHE code, the ubsan_out_of_bounds > handler stores the logging data inside the kvm_ubsan_buffer. The one responsible > for printing is the kernel ubsan_out_of_bounds handler. The process of > decapsulating the data from the buffer is straightforward. > > Signed-off-by: George Popescu <georgepope@google.com> > Signed-off-by: Elena Petrova <lenaptr@google.com> > --- > arch/arm64/include/asm/kvm_ubsan.h | 19 ++++++++++++++++++- > arch/arm64/kvm/hyp/nvhe/ubsan.c | 14 ++++++++++++-- > arch/arm64/kvm/kvm_ubsan_buffer.c | 10 ++++++++++ > 3 files changed, 40 insertions(+), 3 deletions(-) > > diff --git a/arch/arm64/include/asm/kvm_ubsan.h b/arch/arm64/include/asm/kvm_ubsan.h > index fb32c7fd65d4..4f471acb88b0 100644 > --- a/arch/arm64/include/asm/kvm_ubsan.h > +++ b/arch/arm64/include/asm/kvm_ubsan.h > @@ -9,6 +9,23 @@ > #define UBSAN_MAX_TYPE 6 > #define KVM_UBSAN_BUFFER_SIZE 1000 > > + > +struct ubsan_values { > + void *lval; > + void *rval; > + char op; > +}; > + > struct kvm_ubsan_info { > - int type; > + enum { > + UBSAN_OUT_OF_BOUNDS, > + } type; > + union { > + struct out_of_bounds_data out_of_bounds_data; > + }; > + union { > + struct ubsan_values u_val; > + };
I don't see this second union having more members later in the series. Remove it? Even the 'struct ubsan_values' seems redundant and we could just have those three fields directly here.
> }; > + > +void __ubsan_handle_out_of_bounds(void *_data, void *index); > diff --git a/arch/arm64/kvm/hyp/nvhe/ubsan.c b/arch/arm64/kvm/hyp/nvhe/ubsan.c > index 8a194fb1f6cf..55a8f6db8555 100644 > --- a/arch/arm64/kvm/hyp/nvhe/ubsan.c > +++ b/arch/arm64/kvm/hyp/nvhe/ubsan.c > @@ -13,7 +13,6 @@ > #include <asm/kvm_ubsan.h> > #include <asm/kvm_debug_buffer.h> > #include <kvm/arm_pmu.h> > -#include <ubsan.h> > > DEFINE_KVM_DEBUG_BUFFER(struct kvm_ubsan_info, kvm_ubsan_buffer, > kvm_ubsan_buff_wr_ind, KVM_UBSAN_BUFFER_SIZE); > @@ -44,7 +43,18 @@ void __ubsan_handle_type_mismatch(struct type_mismatch_data *data, void *ptr) {} > > void __ubsan_handle_type_mismatch_v1(void *_data, void *ptr) {} > > -void __ubsan_handle_out_of_bounds(void *_data, void *index) {} > +void __ubsan_handle_out_of_bounds(void *_data, void *index) > +{ > + struct kvm_ubsan_info *slot; > + struct out_of_bounds_data *data = _data; > + > + slot = kvm_ubsan_buffer_next_slot(); > + if (slot) { > + slot->type = UBSAN_OUT_OF_BOUNDS; > + slot->out_of_bounds_data = *data; > + slot->u_val.lval = index; > + } > +} > > void __ubsan_handle_shift_out_of_bounds(void *_data, void *lhs, void *rhs) {} > > diff --git a/arch/arm64/kvm/kvm_ubsan_buffer.c b/arch/arm64/kvm/kvm_ubsan_buffer.c > index 4a1959ba9f68..a1523f86be3c 100644 > --- a/arch/arm64/kvm/kvm_ubsan_buffer.c > +++ b/arch/arm64/kvm/kvm_ubsan_buffer.c > @@ -17,6 +17,15 @@ > DECLARE_KVM_DEBUG_BUFFER(struct kvm_ubsan_info, kvm_ubsan_buffer, > kvm_ubsan_buff_wr_ind, KVM_UBSAN_BUFFER_SIZE); > > +void __kvm_check_ubsan_data(struct kvm_ubsan_info *slot) > +{ > + switch (slot->type) { > + case UBSAN_OUT_OF_BOUNDS: > + __ubsan_handle_out_of_bounds(&slot->out_of_bounds_data, > + slot->u_val.lval); > + break; > + } > +} > > void iterate_kvm_ubsan_buffer(unsigned long left, unsigned long right) > { > @@ -26,6 +35,7 @@ void iterate_kvm_ubsan_buffer(unsigned long left, unsigned long right) > slot = (struct kvm_ubsan_info *) this_cpu_ptr_nvhe_sym(kvm_ubsan_buffer); > for (i = left; i < right; ++i) { > /* check ubsan data */ > + __kvm_check_ubsan_data(slot + i);
Not sure why this is breaking out into another function. The code will not be shared with any other user.
> slot[i].type = 0;
This invalidation is redundant. The buffer's cursor will be reset on next hypercall, which will implicitly invalidate all entries.
> } > } > -- > 2.30.0.296.g2bfb1c46d8-goog > > _______________________________________________ > kvmarm mailing list > kvmarm@lists.cs.columbia.edu > https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
| |