remote process. This new implementation also passes all the test programs
I've written so far.
- When exceptions come from traps, no UNIX signal should evet be sent.
- Add a lock to ensure a debugger handles only one exception at a time
- Use a structure to hold flavor and behavior in exception ports, instead
of stuffing the two argument into an int.
- Implement new Mach services: thread_suspend, thread_resume and thread_abort
- Implement Darwin's ptrace PT_ATTACHEXC and PT_THUPDATE commands
- Handle NULL second argument correctly in sigprocmask.
- One mistake in the last commit (darwin_tracesig prototype)
wrong on the semantic front; the spurious wakeup confuses Darwin's gdb.
Allow vm, task and thread operations on remote processes. The code to pick up
the remote process is in mach_sys_msg_trap(), so that any Mach service can
use it.
blocked in the kernel. The task that catched the exception may unblock
it by sending a reply to the exception message (Of course it will have
to change something so that the exception is not immediatly raised again).
Handling of this reply is a bit complicated, as the kernel acts as the
client instead of the server. In this situation, we receive a message
but we will not send any reply (the message we receive is already a reply).
I have not found anything better than a special case in
mach_msg_overwrite_trap() to handle this.
A surprise: exceptions ports are preserved accross forks.
While we are there, use appropriate 64 bit types for make_memory_entry_64.
may turn into exceptions on Mach: a small message sent by the kernel to
the task that requested the exception.
On Darwin, when an exception is sent, no signal can be delivered.
TODO: more exceptions: arithmetic, bad instructions, emulation, s
software, and syscalls (plain and Mach). There is also RPC alert, but
I have no idea about what it is.
While we are there, remove some user ktrace in notification code, and add
a NODEF qualifier in mach_services.master: it will be used for notifications
and exceptions, where the kernel is always client and never server: we
don't want the message to be displayed as "unimplemented xxx" in kdump (thus
UNIMPL is not good), but we don't want to generate the server prototype
(therefore, STD is not good either). NODEF will declare it normally in the
name tables without creating the prototype.
1) make sure Mach servers will not work on data beyond the end of the
request message buffer.
2) make sure that on copying out the reply message buffer, we will not
leak kernel data located after the buffer.
3) make sure that the server will not overwrite memory beyond the end
of the reply message buffer. That check is the responsability of the
server, there is just a DIAGNOSTIC test to check everything is in
good shape. All currently implemented servers in NetBSD have been
modified to check for this condition
While we are here, build the mach services table (formerly in mach_namemap.c)
and the services prototypes automatically from mach_services.master, just
as this is done for system calls.
The next step would be to fold the message formats in the mach_services.master
file, but this tends to be difficult, as some messages are quite long and
complex.
While we are there, resolved another mystery: the unallocated port described
in the comment removed by this commit was in fact allocated by mach_task_pid().
and a body.
- If mach_init is not availabkle for boostrap requests, try to handle them
in the kernel (we don't really handle them, we just try to avoid hanging there)
- minor tweaks.
1) rights should be shared by the threads within a process. While it would
be easier to handle this with the struct proc/struct lwp split, we attempt to
do this now by sharing the right lists. Because each right holds a reference
to struct proc, this might cause some problems later.
2) in pthread_exit, really exit the thread. Also reintialize the righ tlist to
make sure we will not destroy the parent's right list
3) rights can hold multiple permissions on a port (ie: send and receive). Fix th
is.
4) first attempt on right carried by messages. We still have to do rights carrie
d in the message body (complex messages).
This does not buy us new functionnality for now, because we still have to
discover how mach_init (which acts as a name server, enabling processes to
discover each other's ports) is able to receive messages from other processes
(this is a bootstrap problem, and the bootstrap port might be the place to
search).
While we are there:
- removed a lot of debug which is now available using ktrace.
- reworked message handling to avoid mutliple copyin/copyout of the
same data. ktrace of Mach message now uses the in-kernel copy of the
message instead of copying it from userland.
- packed mach trap handlers arguments into a structure to avoid modifying
everything next time we have to add an argument.
mach_port_move_member, mach_port_set_attributes, mach_task_set_special_port,
(none do anything)
Added mach_thread_create_running, which creates a new Mach thread. It
provides the register context of the new thread. We use it in a child
function provided to fork1(). The child function is machine dependent and
is not yet implemented for i386.
The new thread crashes quickly, but at least it starts.
Check for target buffer length, and fail if it is too short
Move mach_msg_trap and mach_msg_overwrite_trap to their own file
Remove some useless debug messages now we have ktrace
Remove __P()
