On Fri, 22 Mar 2019 at 10:11, P J P <ppan...@redhat.com> wrote: > > +-- On Fri, 22 Mar 2019, Peter Maydell wrote --+ > | This document is specific to aarch64, but the part of > | QEMU's device tree code being modified here is > | architecture independent. > | > | Cc'ing David Gibson who will probably know if there is > | an architecture-independent limit on DTB size we should > | be enforcing, or whether we are better just to have a check > | that avoids the overflow. > > Thank you for CC'ing David. It seems Agraf did not receive email @suse.de.
Yes, Alex's email has changed (I've updated the cc list). > Current limit defined by FDT_MAX_SIZE is ~1MB. But currently this is only used when creating a DT from scratch. > Proposed check (dt_size > FDT_MAX_SIZE) in this patch is to enforce same size > limit as used in create_device_tree() and avoid loading large files and the > said integer overflow. My worry is that this might possibly break existing working use cases which load a device tree that is larger than 1MB. Unless there's a cross-architecture justification for the 1MB limit it seems quite a low one to be enforcing (especially since the one limit we've found so far for aarch64 is 2MB, not 1MB). thanks -- PMM
