On 04/15/13 15:24, Laszlo Ersek wrote: > Second, 2/3 breaks the binary format exported under the > FW_CFG_ACPI_TABLES key. The format is as follows: > > * number of tables in entire fw_cfg blob : uint16_t > for each table: > * ACPI payload size belonging to this table : uint16_t > * standard ACPI header > * ACPI table contents > > whereas 2/3 changes the format to: > > * zero-filled uint16_t (i) > for each table: > * uint16_t storing value 1 in LE order (ii) > * number of bytes in blob segment belonging to this table : uint16_t
Sorry, this was mis-editing the email on my part: the _length field is computed alright, it's only the (i)<->(ii) thing that goes wrong. > * standard ACPI header > * ACPI table contents > > (ii) shouldn't exist in per-table storage, the increments should target > (i) actually. Laszlo
