Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
/*-
|
|
|
|
* Copyright (c) 2008 Ed Schouten <ed@FreeBSD.org>
|
|
|
|
* All rights reserved.
|
|
|
|
*
|
|
|
|
* Portions of this software were developed under sponsorship from Snow
|
|
|
|
* B.V., the Netherlands.
|
|
|
|
*
|
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
* modification, are permitted provided that the following conditions
|
|
|
|
* are met:
|
|
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer.
|
|
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
|
|
* documentation and/or other materials provided with the distribution.
|
|
|
|
*
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
|
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
|
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
|
|
* SUCH DAMAGE.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <sys/cdefs.h>
|
|
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
|
|
|
|
#include <sys/param.h>
|
|
|
|
#include <sys/kernel.h>
|
|
|
|
#include <sys/lock.h>
|
|
|
|
#include <sys/queue.h>
|
|
|
|
#include <sys/sysctl.h>
|
|
|
|
#include <sys/systm.h>
|
|
|
|
#include <sys/tty.h>
|
|
|
|
#include <sys/uio.h>
|
|
|
|
|
|
|
|
#include <vm/uma.h>
|
|
|
|
|
|
|
|
/*
|
|
|
|
* TTY output queue buffering.
|
|
|
|
*
|
|
|
|
* The previous design of the TTY layer offered the so-called clists.
|
|
|
|
* These clists were used for both the input queues and the output
|
|
|
|
* queue. We don't use certain features on the output side, like quoting
|
|
|
|
* bits for parity marking and such. This mechanism is similar to the
|
|
|
|
* old clists, but only contains the features we need to buffer the
|
|
|
|
* output.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* Statistics. */
|
2009-02-26 10:28:32 +00:00
|
|
|
static unsigned long ttyoutq_nfast = 0;
|
|
|
|
SYSCTL_ULONG(_kern, OID_AUTO, tty_outq_nfast, CTLFLAG_RD,
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
&ttyoutq_nfast, 0, "Unbuffered reads to userspace on output");
|
2009-02-26 10:28:32 +00:00
|
|
|
static unsigned long ttyoutq_nslow = 0;
|
|
|
|
SYSCTL_ULONG(_kern, OID_AUTO, tty_outq_nslow, CTLFLAG_RD,
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
&ttyoutq_nslow, 0, "Buffered reads to userspace on output");
|
|
|
|
|
|
|
|
struct ttyoutq_block {
|
2009-02-03 19:58:28 +00:00
|
|
|
struct ttyoutq_block *tob_next;
|
|
|
|
char tob_data[TTYOUTQ_DATASIZE];
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
static uma_zone_t ttyoutq_zone;
|
|
|
|
|
2009-02-03 19:58:28 +00:00
|
|
|
#define TTYOUTQ_INSERT_TAIL(to, tob) do { \
|
|
|
|
if (to->to_end == 0) { \
|
|
|
|
tob->tob_next = to->to_firstblock; \
|
|
|
|
to->to_firstblock = tob; \
|
|
|
|
} else { \
|
|
|
|
tob->tob_next = to->to_lastblock->tob_next; \
|
|
|
|
to->to_lastblock->tob_next = tob; \
|
|
|
|
} \
|
|
|
|
to->to_nblocks++; \
|
|
|
|
} while (0)
|
|
|
|
|
|
|
|
#define TTYOUTQ_REMOVE_HEAD(to) do { \
|
|
|
|
to->to_firstblock = to->to_firstblock->tob_next; \
|
|
|
|
to->to_nblocks--; \
|
|
|
|
} while (0)
|
|
|
|
|
|
|
|
#define TTYOUTQ_RECYCLE(to, tob) do { \
|
|
|
|
if (to->to_quota <= to->to_nblocks) \
|
|
|
|
uma_zfree(ttyoutq_zone, tob); \
|
|
|
|
else \
|
|
|
|
TTYOUTQ_INSERT_TAIL(to, tob); \
|
|
|
|
} while(0)
|
|
|
|
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
void
|
|
|
|
ttyoutq_flush(struct ttyoutq *to)
|
|
|
|
{
|
|
|
|
|
|
|
|
to->to_begin = 0;
|
|
|
|
to->to_end = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
|
|
|
ttyoutq_setsize(struct ttyoutq *to, struct tty *tp, size_t size)
|
|
|
|
{
|
|
|
|
struct ttyoutq_block *tob;
|
|
|
|
|
2008-08-30 09:18:27 +00:00
|
|
|
to->to_quota = howmany(size, TTYOUTQ_DATASIZE);
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
|
2008-08-30 09:18:27 +00:00
|
|
|
while (to->to_quota > to->to_nblocks) {
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
/*
|
|
|
|
* List is getting bigger.
|
|
|
|
* Add new blocks to the tail of the list.
|
|
|
|
*
|
|
|
|
* We must unlock the TTY temporarily, because we need
|
|
|
|
* to allocate memory. This won't be a problem, because
|
|
|
|
* in the worst case, another thread ends up here, which
|
|
|
|
* may cause us to allocate too many blocks, but this
|
|
|
|
* will be caught by the loop below.
|
|
|
|
*/
|
|
|
|
tty_unlock(tp);
|
|
|
|
tob = uma_zalloc(ttyoutq_zone, M_WAITOK);
|
|
|
|
tty_lock(tp);
|
|
|
|
|
2009-02-03 19:58:28 +00:00
|
|
|
TTYOUTQ_INSERT_TAIL(to, tob);
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
}
|
2008-08-30 09:18:27 +00:00
|
|
|
}
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
|
2008-08-30 09:18:27 +00:00
|
|
|
void
|
|
|
|
ttyoutq_free(struct ttyoutq *to)
|
|
|
|
{
|
|
|
|
struct ttyoutq_block *tob;
|
|
|
|
|
|
|
|
ttyoutq_flush(to);
|
|
|
|
to->to_quota = 0;
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
|
2009-02-03 19:58:28 +00:00
|
|
|
while ((tob = to->to_firstblock) != NULL) {
|
|
|
|
TTYOUTQ_REMOVE_HEAD(to);
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
uma_zfree(ttyoutq_zone, tob);
|
|
|
|
}
|
2008-08-30 09:18:27 +00:00
|
|
|
|
|
|
|
MPASS(to->to_nblocks == 0);
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
size_t
|
|
|
|
ttyoutq_read(struct ttyoutq *to, void *buf, size_t len)
|
|
|
|
{
|
|
|
|
char *cbuf = buf;
|
|
|
|
|
|
|
|
while (len > 0) {
|
|
|
|
struct ttyoutq_block *tob;
|
|
|
|
size_t cbegin, cend, clen;
|
|
|
|
|
|
|
|
/* See if there still is data. */
|
|
|
|
if (to->to_begin == to->to_end)
|
|
|
|
break;
|
2009-02-03 19:58:28 +00:00
|
|
|
tob = to->to_firstblock;
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
if (tob == NULL)
|
|
|
|
break;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The end address should be the lowest of these three:
|
|
|
|
* - The write pointer
|
|
|
|
* - The blocksize - we can't read beyond the block
|
|
|
|
* - The end address if we could perform the full read
|
|
|
|
*/
|
|
|
|
cbegin = to->to_begin;
|
|
|
|
cend = MIN(MIN(to->to_end, to->to_begin + len),
|
|
|
|
TTYOUTQ_DATASIZE);
|
|
|
|
clen = cend - cbegin;
|
|
|
|
|
2009-02-03 19:58:28 +00:00
|
|
|
/* Copy the data out of the buffers. */
|
|
|
|
memcpy(cbuf, tob->tob_data + cbegin, clen);
|
|
|
|
cbuf += clen;
|
|
|
|
len -= clen;
|
|
|
|
|
|
|
|
if (cend == to->to_end) {
|
|
|
|
/* Read the complete queue. */
|
|
|
|
to->to_begin = 0;
|
|
|
|
to->to_end = 0;
|
|
|
|
} else if (cend == TTYOUTQ_DATASIZE) {
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
/* Read the block until the end. */
|
2009-02-03 19:58:28 +00:00
|
|
|
TTYOUTQ_REMOVE_HEAD(to);
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
to->to_begin = 0;
|
2009-02-03 19:58:28 +00:00
|
|
|
to->to_end -= TTYOUTQ_DATASIZE;
|
|
|
|
TTYOUTQ_RECYCLE(to, tob);
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
} else {
|
|
|
|
/* Read the block partially. */
|
|
|
|
to->to_begin += clen;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return (cbuf - (char *)buf);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* An optimized version of ttyoutq_read() which can be used in pseudo
|
|
|
|
* TTY drivers to directly copy data from the outq to userspace, instead
|
|
|
|
* of buffering it.
|
|
|
|
*
|
|
|
|
* We can only copy data directly if we need to read the entire block
|
|
|
|
* back to the user, because we temporarily remove the block from the
|
|
|
|
* queue. Otherwise we need to copy it to a temporary buffer first, to
|
|
|
|
* make sure data remains in the correct order.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
ttyoutq_read_uio(struct ttyoutq *to, struct tty *tp, struct uio *uio)
|
|
|
|
{
|
|
|
|
|
|
|
|
while (uio->uio_resid > 0) {
|
|
|
|
int error;
|
|
|
|
struct ttyoutq_block *tob;
|
|
|
|
size_t cbegin, cend, clen;
|
|
|
|
|
|
|
|
/* See if there still is data. */
|
|
|
|
if (to->to_begin == to->to_end)
|
|
|
|
return (0);
|
2009-02-03 19:58:28 +00:00
|
|
|
tob = to->to_firstblock;
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
if (tob == NULL)
|
|
|
|
return (0);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The end address should be the lowest of these three:
|
|
|
|
* - The write pointer
|
|
|
|
* - The blocksize - we can't read beyond the block
|
|
|
|
* - The end address if we could perform the full read
|
|
|
|
*/
|
|
|
|
cbegin = to->to_begin;
|
|
|
|
cend = MIN(MIN(to->to_end, to->to_begin + uio->uio_resid),
|
|
|
|
TTYOUTQ_DATASIZE);
|
|
|
|
clen = cend - cbegin;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We can prevent buffering in some cases:
|
|
|
|
* - We need to read the block until the end.
|
|
|
|
* - We don't need to read the block until the end, but
|
|
|
|
* there is no data beyond it, which allows us to move
|
|
|
|
* the write pointer to a new block.
|
|
|
|
*/
|
|
|
|
if (cend == TTYOUTQ_DATASIZE || cend == to->to_end) {
|
|
|
|
atomic_add_long(&ttyoutq_nfast, 1);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Fast path: zero copy. Remove the first block,
|
|
|
|
* so we can unlock the TTY temporarily.
|
|
|
|
*/
|
2009-02-03 19:58:28 +00:00
|
|
|
TTYOUTQ_REMOVE_HEAD(to);
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
to->to_begin = 0;
|
2009-02-03 19:58:28 +00:00
|
|
|
if (to->to_end <= TTYOUTQ_DATASIZE)
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
to->to_end = 0;
|
2009-02-03 19:58:28 +00:00
|
|
|
else
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
to->to_end -= TTYOUTQ_DATASIZE;
|
|
|
|
|
|
|
|
/* Temporary unlock and copy the data to userspace. */
|
|
|
|
tty_unlock(tp);
|
|
|
|
error = uiomove(tob->tob_data + cbegin, clen, uio);
|
|
|
|
tty_lock(tp);
|
|
|
|
|
|
|
|
/* Block can now be readded to the list. */
|
2009-02-03 19:58:28 +00:00
|
|
|
TTYOUTQ_RECYCLE(to, tob);
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
} else {
|
|
|
|
char ob[TTYOUTQ_DATASIZE - 1];
|
|
|
|
atomic_add_long(&ttyoutq_nslow, 1);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Slow path: store data in a temporary buffer.
|
|
|
|
*/
|
|
|
|
memcpy(ob, tob->tob_data + cbegin, clen);
|
|
|
|
to->to_begin += clen;
|
|
|
|
MPASS(to->to_begin < TTYOUTQ_DATASIZE);
|
|
|
|
|
|
|
|
/* Temporary unlock and copy the data to userspace. */
|
|
|
|
tty_unlock(tp);
|
|
|
|
error = uiomove(ob, clen, uio);
|
|
|
|
tty_lock(tp);
|
|
|
|
}
|
2008-08-30 09:18:27 +00:00
|
|
|
|
|
|
|
if (error != 0)
|
|
|
|
return (error);
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
size_t
|
|
|
|
ttyoutq_write(struct ttyoutq *to, const void *buf, size_t nbytes)
|
|
|
|
{
|
|
|
|
const char *cbuf = buf;
|
|
|
|
struct ttyoutq_block *tob;
|
|
|
|
unsigned int boff;
|
|
|
|
size_t l;
|
|
|
|
|
|
|
|
while (nbytes > 0) {
|
|
|
|
boff = to->to_end % TTYOUTQ_DATASIZE;
|
|
|
|
|
|
|
|
if (to->to_end == 0) {
|
|
|
|
/* First time we're being used or drained. */
|
|
|
|
MPASS(to->to_begin == 0);
|
2009-02-03 19:58:28 +00:00
|
|
|
tob = to->to_firstblock;
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
if (tob == NULL) {
|
|
|
|
/* Queue has no blocks. */
|
|
|
|
break;
|
|
|
|
}
|
2009-02-03 19:58:28 +00:00
|
|
|
to->to_lastblock = tob;
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
} else if (boff == 0) {
|
|
|
|
/* We reached the end of this block on last write. */
|
2009-02-03 19:58:28 +00:00
|
|
|
tob = to->to_lastblock->tob_next;
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
if (tob == NULL) {
|
|
|
|
/* We've reached the watermark. */
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
to->to_lastblock = tob;
|
2009-02-03 19:58:28 +00:00
|
|
|
} else {
|
|
|
|
tob = to->to_lastblock;
|
Integrate the new MPSAFE TTY layer to the FreeBSD operating system.
The last half year I've been working on a replacement TTY layer for the
FreeBSD kernel. The new TTY layer was designed to improve the following:
- Improved driver model:
The old TTY layer has a driver model that is not abstract enough to
make it friendly to use. A good example is the output path, where the
device drivers directly access the output buffers. This means that an
in-kernel PPP implementation must always convert network buffers into
TTY buffers.
If a PPP implementation would be built on top of the new TTY layer
(still needs a hooks layer, though), it would allow the PPP
implementation to directly hand the data to the TTY driver.
- Improved hotplugging:
With the old TTY layer, it isn't entirely safe to destroy TTY's from
the system. This implementation has a two-step destructing design,
where the driver first abandons the TTY. After all threads have left
the TTY, the TTY layer calls a routine in the driver, which can be
used to free resources (unit numbers, etc).
The pts(4) driver also implements this feature, which means
posix_openpt() will now return PTY's that are created on the fly.
- Improved performance:
One of the major improvements is the per-TTY mutex, which is expected
to improve scalability when compared to the old Giant locking.
Another change is the unbuffered copying to userspace, which is both
used on TTY device nodes and PTY masters.
Upgrading should be quite straightforward. Unlike previous versions,
existing kernel configuration files do not need to be changed, except
when they reference device drivers that are listed in UPDATING.
Obtained from: //depot/projects/mpsafetty/...
Approved by: philip (ex-mentor)
Discussed: on the lists, at BSDCan, at the DevSummit
Sponsored by: Snow B.V., the Netherlands
dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Don't copy more than was requested. */
|
|
|
|
l = MIN(nbytes, TTYOUTQ_DATASIZE - boff);
|
|
|
|
MPASS(l > 0);
|
|
|
|
memcpy(tob->tob_data + boff, cbuf, l);
|
|
|
|
|
|
|
|
cbuf += l;
|
|
|
|
nbytes -= l;
|
|
|
|
to->to_end += l;
|
|
|
|
}
|
|
|
|
|
|
|
|
return (cbuf - (const char *)buf);
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
ttyoutq_write_nofrag(struct ttyoutq *to, const void *buf, size_t nbytes)
|
|
|
|
{
|
|
|
|
size_t ret;
|
|
|
|
|
|
|
|
if (ttyoutq_bytesleft(to) < nbytes)
|
|
|
|
return (-1);
|
|
|
|
|
|
|
|
/* We should always be able to write it back. */
|
|
|
|
ret = ttyoutq_write(to, buf, nbytes);
|
|
|
|
MPASS(ret == nbytes);
|
|
|
|
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
ttyoutq_startup(void *dummy)
|
|
|
|
{
|
|
|
|
|
|
|
|
ttyoutq_zone = uma_zcreate("ttyoutq", sizeof(struct ttyoutq_block),
|
|
|
|
NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
SYSINIT(ttyoutq, SI_SUB_DRIVERS, SI_ORDER_FIRST, ttyoutq_startup, NULL);
|