Skip to contentSkip to conversation

Intel Restricted Access Region Handler

PATCH In review Ossama Othman 5 participants 15 messages Mar 14, 2010 – Mar 26, 2010 Staging

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Intel Restricted Access Region Handler

The Intel Restricted Access Region Handler provides a buffer allocation mechanism to RAR users. Since the intended usage model is to lock out CPU access to RAR (the CPU will not be able to access RAR memory), this driver does not access RAR memory, and merely keeps track of what areas of RAR memory are in use. It has it's own simple allocator that does not rely on existing kernel allocators (SLAB, etc) since those allocators are too tightly coupled with the paging mechanism, which isn't needed for the intended RAR use cases.

An mmap() implementation is provided for debugging purposes to simplify RAR memory access from the user space. However, it will effectively be a no-op when RAR access control is enabled since the CPU will not be able to access RAR.

This driver should not be confused with the rar_register driver. That driver exposes an interface to access RAR registers on the Moorestown platform. The RAR handler driver relies on the rar_register driver for low level RAR register reads and writes.

---
 drivers/staging/Kconfig                   |    2 +
 drivers/staging/Makefile                  |    1 +
 drivers/staging/memrar/Kconfig            |   15 +
 drivers/staging/memrar/Makefile           |    2 +
 drivers/staging/memrar/memrar.h           |  160 +++++
 drivers/staging/memrar/memrar_allocator.c |  432 +++++++++++++
 drivers/staging/memrar/memrar_allocator.h |  141 +++++
 drivers/staging/memrar/memrar_handler.c   |  948 +++++++++++++++++++++++++++++
 8 files changed, 1701 insertions(+), 0 deletions(-)
 create mode 100644 drivers/staging/memrar/Kconfig
 create mode 100644 drivers/staging/memrar/Makefile
 create mode 100644 drivers/staging/memrar/memrar.h
 create mode 100644 drivers/staging/memrar/memrar_allocator.c
 create mode 100644 drivers/staging/memrar/memrar_allocator.h
 create mode 100644 drivers/staging/memrar/memrar_handler.c
drivers/staging/Kconfig +2 -0
index 7696a66..a4b402e 100644
@@ -111,6 +111,8 @@ source "drivers/staging/vme/Kconfig"
 
 source "drivers/staging/rar_register/Kconfig"
 
+source "drivers/staging/memrar/Kconfig"
+
 source "drivers/staging/sep/Kconfig"
 
 source "drivers/staging/iio/Kconfig"
drivers/staging/Makefile +1 -0
index ea2e70e..a517236 100644
@@ -36,6 +36,7 @@ obj-$(CONFIG_FB_UDL)		+= udlfb/
 obj-$(CONFIG_HYPERV)		+= hv/
 obj-$(CONFIG_VME_BUS)		+= vme/
 obj-$(CONFIG_RAR_REGISTER)	+= rar_register/
+obj-$(CONFIG_MRST_RAR_HANDLER)	+= memrar/
 obj-$(CONFIG_DX_SEP)		+= sep/
 obj-$(CONFIG_IIO)		+= iio/
 obj-$(CONFIG_RAMZSWAP)		+= ramzswap/
drivers/staging/memrar/Kconfig new file +15 -0
new file mode 100644
index 0000000..a5598a8
@@ -0,0 +1,15 @@
+config MRST_RAR_HANDLER
+	tristate "RAR handler driver for Intel Moorestown platform"
+	select RAR_REGISTER
+	---help---
+	  This driver provides a memory management interface to
+	  restricted access regions (RAR) available on the Intel
+	  Moorestown platform.
+
+	  Once locked down, restricted access regions are only
+	  accessible by specific hardware on the platform.  The x86
+	  CPU is typically not one of those platforms.  As such this
+	  driver does not access RAR, and only provides a buffer
+	  allocation/bookkeeping mechanism.
+
+	  If unsure, say N.
drivers/staging/memrar/Makefile new file +2 -0
new file mode 100644
index 0000000..a3336c0
@@ -0,0 +1,2 @@
+obj-$(CONFIG_MRST_RAR_HANDLER)	+= memrar.o
+memrar-y			:= memrar_allocator.o memrar_handler.o
drivers/staging/memrar/memrar.h new file +160 -0

160 lines not on this page · raw message

drivers/staging/memrar/memrar_allocator.c new file +432 -0

432 lines not on this page · raw message

drivers/staging/memrar/memrar_allocator.h new file +141 -0

141 lines not on this page · raw message

drivers/staging/memrar/memrar_handler.c new file +948 -0

948 lines not on this page · raw message

Signature · 1 line
1.6.3.3
#1 lore raw

Hi,

Quoted 1 line

Subject: [PATCH] Intel Restricted Access Region Handler

I should mention the patch was built against the latest linux 2.6 kernel branch.

Thanks,
-Ossama

#2 lore raw
In reply to Ossama Othman
Quoted 14 lines · Ossama Othman

The Intel Restricted Access Region Handler provides a buffer allocation mechanism to RAR users. Since the intended usage model is to lock out CPU access to RAR (the CPU will not be able to access RAR memory), this driver does not access RAR memory, and merely keeps track of what areas of RAR memory are in use. It has it's own simple allocator that does not rely on existing kernel allocators (SLAB, etc) since those allocators are too tightly coupled with the paging mechanism, which isn't needed for the intended RAR use cases.

An mmap() implementation is provided for debugging purposes to simplify RAR memory access from the user space. However, it will effectively be a no-op when RAR access control is enabled since the CPU will not be able to access RAR.

Quoted 4 lines

This driver should not be confused with the rar_register driver. That driver exposes an interface to access RAR registers on the Moorestown platform. The RAR handler driver relies on the rar_register driver for low level RAR register reads and writes.

Can you also send me a patch, that adds a TODO file to the directory of this driver, that explains what you think needs to be done in order to get this code merged into the main portion of the kernel tree, and include your email address in it so that people who wish to send patches, know who to copy on them?

See the other drivers/staging/*/TODO files for an example of what is expected here.

thanks,

greg k-h

#3 lore raw

PATCH 1/1 staging: Intel Restricted Access Region Handler

This is a resubmission of the Intel RAR Handler (memrar) driver that add a TODO file, and also converts the Doxygen style comments to kernel-doc style.

The Intel Restricted Access Region Handler provides a buffer allocation mechanism to RAR users. Since the intended usage model is to lock out CPU access to RAR (the CPU will not be able to access RAR memory), this driver does not access RAR memory, and merely keeps track of what areas of RAR memory are in use. It has it's own simple allocator that does not rely on existing kernel allocators (SLAB, etc) since those allocators are too tightly coupled with the paging mechanism, which isn't needed for the intended RAR use cases.

An mmap() implementation is provided for debugging purposes to simplify RAR memory access from the user space. However, it will effectively be a no-op when RAR access control is enabled since the CPU will not be able to access RAR.

This driver should not be confused with the rar_register driver. That driver exposes an interface to access RAR registers on the Moorestown platform. The RAR handler driver relies on the rar_register driver for low level RAR register reads and writes.

This patch was generated and built against the latest linux-2.6 master branch.

---
 drivers/staging/Kconfig                   |    2 +
 drivers/staging/Makefile                  |    1 +
 drivers/staging/memrar/Kconfig            |   15 +
 drivers/staging/memrar/Makefile           |    2 +
 drivers/staging/memrar/TODO               |   43 ++
 drivers/staging/memrar/memrar.h           |  155 +++++
 drivers/staging/memrar/memrar_allocator.c |  432 +++++++++++++
 drivers/staging/memrar/memrar_allocator.h |  149 +++++
 drivers/staging/memrar/memrar_handler.c   |  937 +++++++++++++++++++++++++++++
 9 files changed, 1736 insertions(+), 0 deletions(-)
 create mode 100644 drivers/staging/memrar/Kconfig
 create mode 100644 drivers/staging/memrar/Makefile
 create mode 100644 drivers/staging/memrar/TODO
 create mode 100644 drivers/staging/memrar/memrar.h
 create mode 100644 drivers/staging/memrar/memrar_allocator.c
 create mode 100644 drivers/staging/memrar/memrar_allocator.h
 create mode 100644 drivers/staging/memrar/memrar_handler.c
drivers/staging/Kconfig +2 -0
index 7696a66..a4b402e 100644
@@ -111,6 +111,8 @@ source "drivers/staging/vme/Kconfig"
 
 source "drivers/staging/rar_register/Kconfig"
 
+source "drivers/staging/memrar/Kconfig"
+
 source "drivers/staging/sep/Kconfig"
 
 source "drivers/staging/iio/Kconfig"
drivers/staging/Makefile +1 -0
index ea2e70e..a517236 100644
@@ -36,6 +36,7 @@ obj-$(CONFIG_FB_UDL)		+= udlfb/
 obj-$(CONFIG_HYPERV)		+= hv/
 obj-$(CONFIG_VME_BUS)		+= vme/
 obj-$(CONFIG_RAR_REGISTER)	+= rar_register/
+obj-$(CONFIG_MRST_RAR_HANDLER)	+= memrar/
 obj-$(CONFIG_DX_SEP)		+= sep/
 obj-$(CONFIG_IIO)		+= iio/
 obj-$(CONFIG_RAMZSWAP)		+= ramzswap/
drivers/staging/memrar/Kconfig new file +15 -0
new file mode 100644
index 0000000..a5598a8
@@ -0,0 +1,15 @@
+config MRST_RAR_HANDLER
+	tristate "RAR handler driver for Intel Moorestown platform"
+	select RAR_REGISTER
+	---help---
+	  This driver provides a memory management interface to
+	  restricted access regions (RAR) available on the Intel
+	  Moorestown platform.
+
+	  Once locked down, restricted access regions are only
+	  accessible by specific hardware on the platform.  The x86
+	  CPU is typically not one of those platforms.  As such this
+	  driver does not access RAR, and only provides a buffer
+	  allocation/bookkeeping mechanism.
+
+	  If unsure, say N.
drivers/staging/memrar/Makefile new file +2 -0
new file mode 100644
index 0000000..a3336c0
@@ -0,0 +1,2 @@
+obj-$(CONFIG_MRST_RAR_HANDLER)	+= memrar.o
+memrar-y			:= memrar_allocator.o memrar_handler.o
drivers/staging/memrar/TODO new file +43 -0

43 lines not on this page · raw message

drivers/staging/memrar/memrar.h new file +155 -0

155 lines not on this page · raw message

drivers/staging/memrar/memrar_allocator.c new file +432 -0

432 lines not on this page · raw message

drivers/staging/memrar/memrar_allocator.h new file +149 -0

149 lines not on this page · raw message

drivers/staging/memrar/memrar_handler.c new file +937 -0

937 lines not on this page · raw message

Signature · 1 line
1.6.3.3
#4 lore raw
In reply to Greg KH

Greg, Randy,

Quoted 8 lines

Can you also send me a patch, that adds a TODO file to the directory of this driver, that explains what you think needs to be done in order to get this code merged into the main portion of the kernel tree, and include your email address in it so that people who wish to send patches, know who to copy on them?

See the other drivers/staging/*/TODO files for an example of what is expected here.

I just resubmitted a consolidated patch that includes the TODO file you requested, as well as the kernel-doc related updates that Randy requested. Randy, I know that you were not in a rush to get them. I just wanted to get them out of the way.

Thanks!
-Ossama

#5 lore raw
Quoted 8 lines · Ossama

Greg, Randy,

Quoted 6 lines

Can you also send me a patch, that adds a TODO file to the directory of this driver, that explains what you think needs to be done in order to get this code merged into the main portion of the kernel tree, and include your email address in it so that people who wish to send patches, know who to copy on them?

Quoted 7 lines
Quoted 2 lines

See the other drivers/staging/*/TODO files for an example of what is expected here.

I just resubmitted a consolidated patch that includes the TODO file you requested, as well as the kernel-doc related updates that Randy requested. Randy, I know that you were not in a rush to get them. I just wanted to get them out of the way.

Looks good, I'll queue it up in a few days.

thanks,

greg k-h

#6 lore raw
Quoted 12 lines · On Mon, 15 Mar 2010 16:45:08 -0700 Greg KH <[email protected]> wrote:
Quoted 11 lines · Ossama

Greg, Randy,

Quoted 8 lines

Can you also send me a patch, that adds a TODO file to the directory of this driver, that explains what you think needs to be done in order to get this code merged into the main portion of the kernel tree, and include your email address in it so that people who wish to send patches, know who to copy on them?

See the other drivers/staging/*/TODO files for an example of what is expected here.

Quoted 6 lines
Quoted 4 lines

I just resubmitted a consolidated patch that includes the TODO file you requested, as well as the kernel-doc related updates that Randy requested. Randy, I know that you were not in a rush to get them. I just wanted to get them out of the way.

Looks good, I'll queue it up in a few days.

Then what happens? I guess I still don't understand the -staging process.

Does the code get re-posted for re-review? If so, who does this? How do we ensure that suitably experienced/motivated reviewers and testers take a look at this code?

Or does it just fester in -staging until we get bored of it then it gets merged anyway?

Regarding the code itself: it appears to implement a userspace interface via some /dev node. But there is no description of this interface at all in the changelog and there is no documentation provided. But the userspace-facing interface is the most important part of the entire feature, because it is something we cannot ever change. It should be exhaustively described right up-front in the changelog so that reviewers can easily and fully understand the proposed API.

#7 lore raw
Quoted 20 lines · Andrew Morton
Quoted 17 lines · On Mon, 15 Mar 2010 16:45:08 -0700 Greg KH <[email protected]> wrote:
Quoted 15 lines · Ossama

Greg, Randy,

Quoted 8 lines

Can you also send me a patch, that adds a TODO file to the directory of this driver, that explains what you think needs to be done in order to get this code merged into the main portion of the kernel tree, and include your email address in it so that people who wish to send patches, know who to copy on them?

See the other drivers/staging/*/TODO files for an example of what is expected here.

I just resubmitted a consolidated patch that includes the TODO file you requested, as well as the kernel-doc related updates that Randy requested. Randy, I know that you were not in a rush to get them. I just wanted to get them out of the way.

Quoted 7 lines
Quoted 1 line

Looks good, I'll queue it up in a few days.

Then what happens? I guess I still don't understand the -staging process.

Does the code get re-posted for re-review? If so, who does this? How do we ensure that suitably experienced/motivated reviewers and testers take a look at this code?

The code lives in the drivers/staging tree for a while, constantly being cleaned up and the issues in the original TODO file addressed. When those are finished, and me and the original authors feel the driver is ready to move into the main portion of the kernel tree, the code is then posted for review for the subsystem that the driver lives in.

After a few rounds of fixing stuff up there, then the code moves into that subsystem and everyone lives happily ever after.

Quoted 2 lines

Or does it just fester in -staging until we get bored of it then it gets merged anyway?

No, if it festers and no one pays attention to it, I delete it from the drivers/staging tree. Lots of code has been recently removed this way.

Quoted 8 lines

Regarding the code itself: it appears to implement a userspace interface via some /dev node. But there is no description of this interface at all in the changelog and there is no documentation provided. But the userspace-facing interface is the most important part of the entire feature, because it is something we cannot ever change. It should be exhaustively described right up-front in the changelog so that reviewers can easily and fully understand the proposed API.

I agree, that is one thing that always needs to be closely reviewed.

thanks,

greg k-h

#8 lore raw
In reply to Andrew Morton

Andrew, Greg,

Quoted 8 lines

Regarding the code itself: it appears to implement a userspace interface via some /dev node. But there is no description of this interface at all in the changelog and there is no documentation provided. But the userspace-facing interface is the most important part of the entire feature, because it is something we cannot ever change. It should be exhaustively described right up-front in the changelog so that reviewers can easily and fully understand the proposed API.

Would a README file of some sort be suitable to describe the driver interface or should I just embed additional documentation in the driver's interface header?

Thanks,
-Ossama

#9 lore raw
Quoted 11 lines · Ossama

Andrew, Greg,

Quoted 8 lines

Regarding the code itself: it appears to implement a userspace interface via some /dev node. But there is no description of this interface at all in the changelog and there is no documentation provided. But the userspace-facing interface is the most important part of the entire feature, because it is something we cannot ever change. It should be exhaustively described right up-front in the changelog so that reviewers can easily and fully understand the proposed API.

Quoted 3 lines

Would a README file of some sort be suitable to describe the driver interface or should I just embed additional documentation in the driver's interface header?

You are going to have to create a Documentation/ABI/ style file eventually, so you might as well do it now and add it to the driver directory.

thanks,

greg k-h

#10 lore raw

PATCH 1/1 staging: Initial memrar ABI document

---
 drivers/staging/memrar/memrar-abi |   89 +++++++++++++++++++++++++++++++++++++
 1 files changed, 89 insertions(+), 0 deletions(-)
 create mode 100644 drivers/staging/memrar/memrar-abi
drivers/staging/memrar/memrar-abi new file +89 -0
new file mode 100644
index 0000000..98a6bb1
@@ -0,0 +1,89 @@
+What:		/dev/memrar
+Date:		March 2010
+KernelVersion:	Kernel version this feature first showed up in.
+Contact:	Ossama Othman <[email protected]>
+Description:	The Intel Moorestown Restricted Access Region (RAR)
+		Handler driver exposes an ioctl() based interface that
+		allows a user to reserve and release blocks of RAR
+		memory.
+
+		Note:  A sysfs based one was not appropriate for the
+		RAR handler's usage model.
+
+		=========================================================
+				ioctl() Requests
+		=========================================================
+		RAR_HANDLER_RESERVE
+		-------------------
+		Description:	Reserve RAR block.
+		Type:		struct RAR_block_info
+		Direction:	in/out
+		Errors:		EINVAL (invalid RAR type or size)
+				ENOMEM (not enough RAR memory)
+
+		RAR_HANDLER_STAT
+		----------------
+		Description:	Get RAR statistics.
+		Type:		struct RAR_stat
+		Direction:	in/out
+		Errors:		EINVAL (invalid RAR type)
+
+		RAR_HANDLER_RELEASE
+		-------------------
+		Description:	Release previously reserved RAR block.
+		Type:		32 bit unsigned integer
+				(e.g. uint32_t), i.e the RAR "handle".
+		Direction:	in
+		Errors:		EINVAL (invalid RAR handle)
+
+
+		=========================================================
+			ioctl() Request Parameter Types
+		=========================================================
+		The structures referred to above are defined as
+		follows:
+
+		/**
+		 * struct RAR_block_info - user space struct that
+		 *			   describes RAR buffer
+		 * @type:	Type of RAR memory (e.g.,
+		 *		RAR_TYPE_VIDEO or RAR_TYPE_AUDIO) [in]
+		 * @size:	Requested size of a block in bytes to
+		 *		be reserved in RAR. [in]
+		 * @handle:	Handle that can be used to refer to
+		 *		reserved block. [out]
+		 *
+		 * This is the basic structure exposed to the user
+		 * space that describes a given RAR buffer.  It used
+		 * as the parameter for the RAR_HANDLER_RESERVE ioctl.
+		 * The buffer's underlying bus address is not exposed
+		 * to the user.  User space code refers to the buffer
+		 * entirely by "handle".
+		 */
+		struct RAR_block_info {
+			__u32 type;
+			__u32 size;
+			__u32 handle;
+		};
+
+		/**
+		 * struct RAR_stat - RAR statistics structure
+		 * @type:		Type of RAR memory (e.g.,
+		 *			RAR_TYPE_VIDEO or
+		 *			RAR_TYPE_AUDIO) [in]
+		 * @capacity:		Total size of RAR memory
+		 *			region. [out]
+		 * @largest_block_size:	Size of the largest reservable
+		 *			block. [out]
+		 *
+		 * This structure is used for RAR_HANDLER_STAT ioctl.
+		 */
+		struct RAR_stat {
+			__u32 type;
+			__u32 capacity;
+			__u32 largest_block_size;
+		};
+
+		Lastly, the RAR_HANDLER_RELEASE ioctl expects a
+		"handle" to the RAR block of memory.  It is a 32 bit
+		unsigned integer.
Signature · 1 line
1.6.3.3
#11 lore raw
In reply to Ossama Othman
Quoted 75 lines · Ossama Othman

The Intel Restricted Access Region Handler provides a buffer allocation mechanism to RAR users. Since the intended usage model is to lock out CPU access to RAR (the CPU will not be able to access RAR memory), this driver does not access RAR memory, and merely keeps track of what areas of RAR memory are in use. It has it's own simple allocator that does not rely on existing kernel allocators (SLAB, etc) since those allocators are too tightly coupled with the paging mechanism, which isn't needed for the intended RAR use cases.

An mmap() implementation is provided for debugging purposes to simplify RAR memory access from the user space. However, it will effectively be a no-op when RAR access control is enabled since the CPU will not be able to access RAR.

This driver should not be confused with the rar_register driver. That driver exposes an interface to access RAR registers on the Moorestown platform. The RAR handler driver relies on the rar_register driver for low level RAR register reads and writes.

---
 drivers/staging/Kconfig                   |    2 +
 drivers/staging/Makefile                  |    1 +
 drivers/staging/memrar/Kconfig            |   15 +
 drivers/staging/memrar/Makefile           |    2 +
 drivers/staging/memrar/memrar.h           |  160 +++++
 drivers/staging/memrar/memrar_allocator.c |  432 +++++++++++++
 drivers/staging/memrar/memrar_allocator.h |  141 +++++
 drivers/staging/memrar/memrar_handler.c   |  948 +++++++++++++++++++++++++++++
 8 files changed, 1701 insertions(+), 0 deletions(-)
 create mode 100644 drivers/staging/memrar/Kconfig
 create mode 100644 drivers/staging/memrar/Makefile
 create mode 100644 drivers/staging/memrar/memrar.h
 create mode 100644 drivers/staging/memrar/memrar_allocator.c
 create mode 100644 drivers/staging/memrar/memrar_allocator.h
 create mode 100644 drivers/staging/memrar/memrar_handler.c
drivers/staging/memrar/Makefile new file +2 -0
new file mode 100644
index 0000000..a3336c0
@@ -0,0 +1,2 @@
+obj-$(CONFIG_MRST_RAR_HANDLER)	+= memrar.o
+memrar-y			:= memrar_allocator.o memrar_handler.o
drivers/staging/memrar/memrar.h new file +24 -0
new file mode 100644
index 0000000..7a09e7d
@@ -0,0 +1,160 @@
+/*
+ *      RAR Handler (/dev/memrar) internal driver API.
+ *      Copyright (C) 2009 Intel Corporation. All rights reserved.
+ *
+ *      This program is free software; you can redistribute it and/or
+ *      modify it under the terms of version 2 of the GNU General
+ *      Public License as published by the Free Software Foundation.
+ *
+ *      This program is distributed in the hope that it will be
+ *      useful, but WITHOUT ANY WARRANTY; without even the implied
+ *      warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
+ *      PURPOSE.  See the GNU General Public License for more details.
+ *      You should have received a copy of the GNU General Public
+ *      License along with this program; if not, write to the Free
+ *      Software Foundation, Inc., 59 Temple Place - Suite 330,
+ *      Boston, MA  02111-1307, USA.
+ *      The full GNU General Public License is included in this
+ *      distribution in the file called COPYING.
+ */
+
+
+#ifndef _MEMRAR_H
+#define _MEMRAR_H
+

drivers/staging/memrar/memrar.h line 34

Quoted 10 lines
drivers/staging/memrar/memrar.h new file +10 -0
+#include <linux/ioctl.h>
+#include <linux/types.h>
+
+
+/*
+ * @struct RAR_stat
+ *
+ * @brief This structure is used for @c RAR_HANDLER_STAT ioctl and for
+ *        @c RAR_get_stat() user space wrapper function.
+ */

Don't use doxygen or whatever that is in the Linux kernel.
Use kernel-doc notation as described in Documentation/kernel-doc-nano-HOWTO.txt.

In many places/multiple files.

drivers/staging/memrar/memrar.h line 46

Quoted 12 lines
drivers/staging/memrar/memrar.h new file +12 -0
+struct RAR_stat {
+	/* Type of RAR memory (e.g., audio vs. video) */
+	__u32 type;
+
+	/*
+	* Total size of RAR memory region.
+	*/
+	__u32 capacity;
+
+	/* Size of the largest reservable block. */
+	__u32 largest_block_size;
+};

Thanks.

Signature · 1 line
~Randy
#12 lore raw

Hi Randy,

Quoted 5 lines

Don't use doxygen or whatever that is in the Linux kernel. Use kernel-doc notation as described in Documentation/kernel-doc-nano- HOWTO.txt.

In many places/multiple files.

I'll fix this immediately.

Thanks!
-Ossama

#14 lore raw
Quoted 2 lines · Alan Cox

Randy - this is a staging driver submission so this sort of stuff is going to get fixed anyway

OK.

I didn't require that it be fixed immediately.

Signature · 1 line
~Randy