Hi all,
after reviewing the patch "fix multiply issue for int and uint types"
with Ian Jackson, we realized that cpu_ioreq_pio and cpu_ioreq_move are
in much need for a simplification as well as removal of a possible
integer overflow.
This patch series tries to accomplish both switching to two new helper
functions and using a more obvious arithmetic. Doing so it should also
fix the original problem that Dongxiao was experiencing. The C language
can be a nasty backstabber when signed and unsigned integers are
involved.
The current patch series if for qemu-xen-traditional but if the patches
are deemed correct I'll submit an equivalent set for QEMU upstream (with
the appropriate code style changes).
Stefano Stabellini (2):
i should be uint32_t rather than int
introduce read_physical_offset and write_physical_offset
i386-dm/helper2.c | 66 +++++++++++++++++++++++++++++++++-------------------
1 files changed, 42 insertions(+), 24 deletions(-)
Cheers,
Stefano
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
after reviewing the patch "fix multiply issue for int and uint types"
with Ian Jackson, we realized that cpu_ioreq_pio and cpu_ioreq_move are
in much need for a simplification as well as removal of a possible
integer overflow.
This patch series tries to accomplish both switching to two new helper
functions and using a more obvious arithmetic. Doing so it should also
fix the original problem that Dongxiao was experiencing. The C language
can be a nasty backstabber when signed and unsigned integers are
involved.
The current patch series if for qemu-xen-traditional but if the patches
are deemed correct I'll submit an equivalent set for QEMU upstream (with
the appropriate code style changes).
Stefano Stabellini (2):
i should be uint32_t rather than int
introduce read_physical_offset and write_physical_offset
i386-dm/helper2.c | 66 +++++++++++++++++++++++++++++++++-------------------
1 files changed, 42 insertions(+), 24 deletions(-)
Cheers,
Stefano
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel