On Mon, Feb 24, 2020 at 03:55:39PM -0500, Jagannathan Raman wrote: > From: Elena Ufimtseva > > Validate the incoming commands to confirm that they would not cause any > errors in the remote process. > > Signed-off-by: Elena Ufimtseva > Signed-off-by: Jagannathan Raman > Signed-off-by: John G Johnson > --- > hw/proxy/qemu-proxy.c | 6 +++- > include/io/mpqemu-link.h | 2 ++ > io/mpqemu-link.c | 75 +++++++++++++++++++++++++++++++++++++++++++++++- > remote/remote-main.c | 4 +++ > 4 files changed, 85 insertions(+), 2 deletions(-) Please squash this into the patch(es) that introduced the code. Reviewers want to see a logical sequence of patches. Introducing unsafe code in an earlier patch and adding checks in a later patch makes it impossible to review the patches in sequence (reviewers would waste time pointing out bugs that end up getting fixed later). > diff --git a/remote/remote-main.c b/remote/remote-main.c > index 20d160e..c4aa3e0 100644 > --- a/remote/remote-main.c > +++ b/remote/remote-main.c > @@ -435,6 +435,10 @@ static void process_msg(GIOCondition cond, MPQemuChannel *chan) > if (msg->id > MAX_REMOTE_DEVICES) { > error_setg(&err, "id of the device is larger than max number of "\ > "devices per remote process."); > + } Was goto finalize_loop accidentally dropped?