On 03/05/19 12:32, nadav.amit@xxxxxxxxx wrote: > From: Nadav Amit <nadav.amit@xxxxxxxxx> > > For reproducibility, it is best to zero pages before they are used. > There are hidden assumptions on memory being zeroed (by BIOS/KVM), which > might be broken at any given moment. The full argument appears in the > first patch commit log. > > Following the first patch that zeros the memory, the rest of the > patch-set removes redundant zeroing do to the additional zeroing. > > This patch-set is only tested on x86. > > v1->v2: > * Change alloc_pages() as well > * Remove redundant page zeroing [Andrew] > > Nadav Amit (4): > lib/alloc_page: Zero allocated pages > x86: Remove redeundant page zeroing > lib: Remove redeundant page zeroing > arm: Remove redeundant page zeroing > > lib/alloc_page.c | 4 ++++ > lib/arm/asm/pgtable.h | 2 -- > lib/arm/mmu.c | 1 - > lib/arm64/asm/pgtable.h | 1 - > lib/virtio-mmio.c | 1 - > lib/x86/intel-iommu.c | 5 ----- > x86/eventinj.c | 1 - > x86/hyperv_connections.c | 4 ---- > x86/vmx.c | 10 ---------- > x86/vmx_tests.c | 11 ----------- > 10 files changed, 4 insertions(+), 36 deletions(-) > Queued, thanks. Paolo
