[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH 5/8] ipc: provide the protected payload in ipc_kmsg_copyout_h
|
From: |
Samuel Thibault |
|
Subject: |
Re: [PATCH 5/8] ipc: provide the protected payload in ipc_kmsg_copyout_header |
|
Date: |
Wed, 1 Oct 2014 01:17:30 +0200 |
|
User-agent: |
Mutt/1.5.21+34 (58baf7c9f32f) (2010-12-30) |
Ack.
Justus Winter, le Thu 18 Sep 2014 15:22:47 +0200, a écrit :
> * ipc/ipc_kmsg.c (ipc_kmsg_copyout_header): If a protected payload is
> set for the destination port, provide it in msgh_protected_payload.
> * ipc/mach_msg.c (mach_msg_trap): Likewise in the fast paths.
> * doc/mach.texi (Message Receive): Document message semantics with
> protected payloads.
> ---
> doc/mach.texi | 19 +++++++++++++++++++
> ipc/ipc_kmsg.c | 58
> +++++++++++++++++++++++++++++++++++++++++++++-------------
> ipc/mach_msg.c | 53 ++++++++++++++++++++++++++++++++++++++++-------------
> 3 files changed, 104 insertions(+), 26 deletions(-)
>
> diff --git a/doc/mach.texi b/doc/mach.texi
> index b187888..c57e607 100644
> --- a/doc/mach.texi
> +++ b/doc/mach.texi
> @@ -1949,6 +1949,25 @@ loses the receive right after the message was dequeued
> from it, then
> right still exists, but isn't held by the caller, then
> @code{msgh_local_port} specifies @code{MACH_PORT_NULL}.
>
> +Servers usually associate some state with a receive right. To that
> +end, they might use a hash table to look up the state for the port a
> +message was sent to. To optimize this, a task may associate an opaque
> +@var{payload} with a receive right using the
> +@code{mach_port_set_protected_payload} function. Once this is done,
> +the kernel will set the @code{msgh_protected_payload} field to
> +@var{payload} when delivering a message to this right and indicate
> +this by setting the local part of @code{msgh_bits} to
> +@code{MACH_MSG_TYPE_PROTECTED_PAYLOAD}.
> +
> +The support for protected payloads was added to GNU Mach. To preserve
> +binary compatibility, the @code{msgh_local_port} and
> +@code{msgh_local_port} share the same location. This makes it
> +possible to add the payload information without increasing the size of
> +@code{mach_msg_header_t}. This is an implementation detail. Which
> +field is valid is determined by the local part of the
> +@code{msgh_bits}. Existing software is not affected. When a receive
> +right is transferred to another task, its payload is cleared.
> +
> Received messages are stamped with a sequence number, taken from the
> port from which the message was received. (Messages received from a
> port set are stamped with a sequence number from the appropriate member
> diff --git a/ipc/ipc_kmsg.c b/ipc/ipc_kmsg.c
> index 06cec72..71a0d74 100644
> --- a/ipc/ipc_kmsg.c
> +++ b/ipc/ipc_kmsg.c
> @@ -1799,9 +1799,17 @@ ipc_kmsg_copyout_header(
> } else
> ip_unlock(dest);
>
> - msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> - MACH_MSGH_BITS(0, MACH_MSG_TYPE_PORT_SEND));
> - msg->msgh_local_port = dest_name;
> + if (! ipc_port_flag_protected_payload(dest)) {
> + msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> + MACH_MSGH_BITS(0, MACH_MSG_TYPE_PORT_SEND));
> + msg->msgh_local_port = dest_name;
> + } else {
> + msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> + MACH_MSGH_BITS(
> + 0, MACH_MSG_TYPE_PROTECTED_PAYLOAD));
> + msg->msgh_protected_payload =
> + dest->ip_protected_payload;
> + }
> msg->msgh_remote_port = MACH_PORT_NULL;
> return MACH_MSG_SUCCESS;
> }
> @@ -1897,10 +1905,18 @@ ipc_kmsg_copyout_header(
> } else
> ip_unlock(dest);
>
> - msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> - MACH_MSGH_BITS(MACH_MSG_TYPE_PORT_SEND_ONCE,
> - MACH_MSG_TYPE_PORT_SEND));
> - msg->msgh_local_port = dest_name;
> + if (! ipc_port_flag_protected_payload(dest)) {
> + msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> + MACH_MSGH_BITS(MACH_MSG_TYPE_PORT_SEND_ONCE,
> + MACH_MSG_TYPE_PORT_SEND));
> + msg->msgh_local_port = dest_name;
> + } else {
> + msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> + MACH_MSGH_BITS(MACH_MSG_TYPE_PORT_SEND_ONCE,
> + MACH_MSG_TYPE_PROTECTED_PAYLOAD));
> + msg->msgh_protected_payload =
> + dest->ip_protected_payload;
> + }
> msg->msgh_remote_port = reply_name;
> return MACH_MSG_SUCCESS;
> }
> @@ -1932,9 +1948,18 @@ ipc_kmsg_copyout_header(
> dest_name = MACH_PORT_NULL;
> }
>
> - msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> - MACH_MSGH_BITS(0, MACH_MSG_TYPE_PORT_SEND_ONCE));
> - msg->msgh_local_port = dest_name;
> + if (! ipc_port_flag_protected_payload(dest)) {
> + msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> + MACH_MSGH_BITS(0,
> + MACH_MSG_TYPE_PORT_SEND_ONCE));
> + msg->msgh_local_port = dest_name;
> + } else {
> + msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> + MACH_MSGH_BITS(0,
> + MACH_MSG_TYPE_PROTECTED_PAYLOAD));
> + msg->msgh_protected_payload =
> + dest->ip_protected_payload;
> + }
> msg->msgh_remote_port = MACH_PORT_NULL;
> return MACH_MSG_SUCCESS;
> }
> @@ -2224,9 +2249,16 @@ ipc_kmsg_copyout_header(
> if (IP_VALID(reply))
> ipc_port_release(reply);
>
> - msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> - MACH_MSGH_BITS(reply_type, dest_type));
> - msg->msgh_local_port = dest_name;
> + if (! ipc_port_flag_protected_payload(dest)) {
> + msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> + MACH_MSGH_BITS(reply_type, dest_type));
> + msg->msgh_local_port = dest_name;
> + } else {
> + msg->msgh_bits = (MACH_MSGH_BITS_OTHER(mbits) |
> + MACH_MSGH_BITS(reply_type,
> + MACH_MSG_TYPE_PROTECTED_PAYLOAD));
> + msg->msgh_protected_payload = dest->ip_protected_payload;
> + }
> msg->msgh_remote_port = reply_name;
> }
>
> diff --git a/ipc/mach_msg.c b/ipc/mach_msg.c
> index 01d974b..1e122c7 100644
> --- a/ipc/mach_msg.c
> +++ b/ipc/mach_msg.c
> @@ -1132,11 +1132,19 @@ mach_msg_trap(
> } else
> ip_unlock(dest_port);
>
> - kmsg->ikm_header.msgh_bits =
> - MACH_MSGH_BITS(MACH_MSG_TYPE_PORT_SEND_ONCE,
> - MACH_MSG_TYPE_PORT_SEND);
> + if (! ipc_port_flag_protected_payload(dest_port)) {
> + kmsg->ikm_header.msgh_bits = MACH_MSGH_BITS(
> + MACH_MSG_TYPE_PORT_SEND_ONCE,
> + MACH_MSG_TYPE_PORT_SEND);
> + kmsg->ikm_header.msgh_local_port = dest_name;
> + } else {
> + kmsg->ikm_header.msgh_bits = MACH_MSGH_BITS(
> + MACH_MSG_TYPE_PORT_SEND_ONCE,
> + MACH_MSG_TYPE_PROTECTED_PAYLOAD);
> + kmsg->ikm_header.msgh_protected_payload =
> + dest_port->ip_protected_payload;
> + }
> kmsg->ikm_header.msgh_remote_port = reply_name;
> - kmsg->ikm_header.msgh_local_port = dest_name;
> goto fast_put;
>
> abort_request_copyout:
> @@ -1170,11 +1178,19 @@ mach_msg_trap(
> dest_name = MACH_PORT_NULL;
> }
>
> - kmsg->ikm_header.msgh_bits =
> - MACH_MSGH_BITS(0,
> - MACH_MSG_TYPE_PORT_SEND_ONCE);
> + if (! ipc_port_flag_protected_payload(dest_port)) {
> + kmsg->ikm_header.msgh_bits = MACH_MSGH_BITS(
> + 0,
> + MACH_MSG_TYPE_PORT_SEND_ONCE);
> + kmsg->ikm_header.msgh_local_port = dest_name;
> + } else {
> + kmsg->ikm_header.msgh_bits = MACH_MSGH_BITS(
> + 0,
> + MACH_MSG_TYPE_PROTECTED_PAYLOAD);
> + kmsg->ikm_header.msgh_protected_payload =
> + dest_port->ip_protected_payload;
> + }
> kmsg->ikm_header.msgh_remote_port = MACH_PORT_NULL;
> - kmsg->ikm_header.msgh_local_port = dest_name;
> goto fast_put;
> }
>
> @@ -1204,12 +1220,23 @@ mach_msg_trap(
> dest_name = MACH_PORT_NULL;
> }
>
> - kmsg->ikm_header.msgh_bits =
> - MACH_MSGH_BITS_COMPLEX |
> - MACH_MSGH_BITS(0,
> - MACH_MSG_TYPE_PORT_SEND_ONCE);
> + if (! ipc_port_flag_protected_payload(dest_port)) {
> + kmsg->ikm_header.msgh_bits =
> + MACH_MSGH_BITS_COMPLEX
> + | MACH_MSGH_BITS(
> + 0,
> + MACH_MSG_TYPE_PORT_SEND_ONCE);
> + kmsg->ikm_header.msgh_local_port = dest_name;
> + } else {
> + kmsg->ikm_header.msgh_bits =
> + MACH_MSGH_BITS_COMPLEX
> + | MACH_MSGH_BITS(
> + 0,
> + MACH_MSG_TYPE_PROTECTED_PAYLOAD);
> + kmsg->ikm_header.msgh_protected_payload =
> + dest_port->ip_protected_payload;
> + }
> kmsg->ikm_header.msgh_remote_port = MACH_PORT_NULL;
> - kmsg->ikm_header.msgh_local_port = dest_name;
>
> mr = ipc_kmsg_copyout_body(
> (vm_offset_t) (&kmsg->ikm_header + 1),
> --
> 2.1.0
>
--
Samuel
> C'est vrai. FreeBSD pourrait être bien s'il avait une licence
> correcte.
Oh, mon troll-ô-meter vient d'exploser...
-+- OR in Guide du linuxien pervers - "Bien regler son troll-ô-meter"
- Re: [PATCH 5/5] ipc: provide the protected payload in ipc_kmsg_copyout_header, Justus Winter, 2014/09/18
- [PATCH 1/8] ipc: add protected payload, Justus Winter, 2014/09/18
- [PATCH 2/8] ipc: implement mach_port_{set,clear}_protected_payload, Justus Winter, 2014/09/18
- [PATCH 4/8] include: define MACH_MSG_TYPE_PROTECTED_PAYLOAD, Justus Winter, 2014/09/18
- [PATCH 3/8] include: add msgh_protected_payload to mach_msg_header_t, Justus Winter, 2014/09/18
- [PATCH 5/8] ipc: provide the protected payload in ipc_kmsg_copyout_header, Justus Winter, 2014/09/18
- Re: [PATCH 5/8] ipc: provide the protected payload in ipc_kmsg_copyout_header,
Samuel Thibault <=
- [PATCH 6/8] include: add a payload-aware intran mutator for memory_object_t, Justus Winter, 2014/09/18
- [PATCH 7/8] include: add a payload-aware intran mutator for notify_port_t, Justus Winter, 2014/09/18
- [PATCH 8/8] include: add a payload-aware intran mutator for device_t, Justus Winter, 2014/09/18
- Re: [PATCH 1/8] ipc: add protected payload, Samuel Thibault, 2014/09/30
- Re: [PATCH 5/5] ipc: provide the protected payload in ipc_kmsg_copyout_header, Samuel Thibault, 2014/09/30