Adding upstream version 4.22.0.upstream/4.22.0

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

41 files changed, 3114 insertions, 0 deletions

diff --git a/upstream/mageia-cauldron/man9/MAJOR.9 b/upstream/mageia-cauldron/man9/MAJOR.9
new file mode 100644
index 00000000..b322981a
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/MAJOR.9
@@ -0,0 +1,87 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH MAJOR 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.\"
+.\" Turn off hyphenation and right justification for the first sections.
+.hlm 0
+.na
+.\" The name should be formatted this way for apropos to work
+.SH NAME
+MAJOR, MINOR, MKDEV \- macros to manipulate device major/minor numbers
+.\"
+.\"
+.SH SYNOPSIS
+.B #include <linux/fs.h>
+.sp
+.BI "#define MAJOR(" dev ")"
+.br
+.BI "#define MINOR(" dev ")"
+.br
+.BI "#define MKDEV(" major ", " minor ")"
+.HP
+.\"
+.\" Turn hyphenation and adjustment back on.
+.hlm 1
+.ad
+.SH DESCRIPTION
+The
+.B MAJOR
+and
+.B MINOR
+macros extract from a device id (of type
+.IR dev_t )
+the major and minor numbers respectively. It is the combined device
+number that is found in the
+.B "struct~file"
+structure.
+.PP
+The
+.B MKDEV
+macro assembles a device number from the supplied major and minor
+numbers.
+.SH "RETURN VALUE"
+Describe the return values. Enumerate all the distinct values and all
+the ranges.
+.SH AVAILABILITY
+Linux 0.0+
+.PP
+In later kernels (Linux 2.1 and later) there is movement towards using
+64bit device numbers. Although it is generally a good idea to not
+presume to know how device numbers are constructed, it is during this
+period of transition fundamentally important that you ONLY use these
+functions to manipulate device numbers.
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR mknod "(1)"
+.ad
+.hlm 1
+.PP
+Also list some source files for the
+kernel that implement the functions of the page.
+.SH AUTHOR
+Stephen Williams <steve@icarus.com>
+.SH BUGS
+These are macros, so parameters may be evaluated multiple times.
diff --git a/upstream/mageia-cauldron/man9/MAP_NR.9 b/upstream/mageia-cauldron/man9/MAP_NR.9
new file mode 100644
index 00000000..a43d3b63
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/MAP_NR.9
@@ -0,0 +1,76 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH skel 9 "$Date:$" "Linux DDI" "VM Functions"
+.\"
+.\" Turn off hyphenation and right justification for the first sections.
+.hlm 0
+.na
+.\" The name should be formatted this way for apropos to work
+.SH NAME
+MAP_NR \- get memory map index for page in kernel memory
+.\"
+.\"
+.SH SYNOPSIS
+.B #include <linux/mm.h>
+.sp
+.BI "#define MAP_NR(" page ")"
+.HP
+.\"
+.\" Turn hyphenation and adjustment back on.
+.hlm 1
+.ad
+.SH DESCRIPTION
+This macro maps the page of kernel memory to an index into the
+.B mem_map
+array. The input page is kernel-space page addresses such as those
+that come from the
+.BR get_free_page "(9)"
+function. The expression
+.I "mem_map[MAP_NR(page)]"
+returns a pointer to a
+.I mem_map_t
+for the page in question.
+.PP
+Device drivers need access to the page structure if it is implimenting
+some special kind of memory segment, i.e. cache for a virtual frame buffer.
+.SH "RETURN VALUE"
+Returns an index into the
+.B mem_map
+array.
+.SH AVAILABILITY
+Linux 2.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR get_free_pages "(9)"
+.ad
+.hlm 1
+.PP
+.I "/usr/include/asm/page.h"
+.SH AUTHOR
+Stephen Williams <steve@icarus.com>
+.SH BUGS
+There is no check that the page you are passing really is a valid page
+address.
diff --git a/upstream/mageia-cauldron/man9/MOD_INC_USE_COUNT.9 b/upstream/mageia-cauldron/man9/MOD_INC_USE_COUNT.9
new file mode 100644
index 00000000..e3509b80
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/MOD_INC_USE_COUNT.9
@@ -0,0 +1,65 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH MOD_INC_USE_COUNT 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.SH NAME
+MOD_INC_USE_COUNT, MOD_DEC_USE_COUNT \- Support reference counting of modules
+.SH SYNOPSIS
+.B #include <linux/module.h>
+.sp
+.B #define MOD_INC_USE_COUNT
+.br
+.B #define MOD_DEC_USE_COUNT
+.SH DESCRIPTION
+These macros support reference counting of modules during their
+lifetime. Each time
+.BR MOD_INC_USE_COUNT "(9)"
+is invoked, the kernel increases the number of counted references to
+the module. The
+.BR MOD_DEC_USE_COUNT "(9)"
+decreases the reference count.
+.PP
+The reference count is used by the kernel to know when a module is no
+longer in use by the device or application software. A
+.BR MOD_INC_USE_COUNT "(9)"
+is generally placed in the open entry for a driver, and a
+.BR MOD_DEC_USE_COUNT "(9)"
+in the release for the driver,
+to count the number of times the device has been opened, and thus
+prevent the module being unloaded when in use. The device driver may
+increment the use count other times, such as when memory is mapped or
+the module must remain loaded for external events.
+.PP
+If the device driver is not compiled as a module
+.IR "" ( MODULE
+is not defined) then the macro reduces to no behavior.
+.SH "RETURN VALUE"
+These macros take no parameters and return no result.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.I "/usr/include/linux/module.h"
+.SH AUTHOR
+Stephen Williams <steve@icarus.com>
+.SH BUGS
diff --git a/upstream/mageia-cauldron/man9/atomic_add.9 b/upstream/mageia-cauldron/man9/atomic_add.9
new file mode 100644
index 00000000..dce43cb1
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/atomic_add.9
@@ -0,0 +1,90 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH atomic_add 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.\"
+.\" Turn off hyphenation and right justification for the first sections.
+.hlm 0
+.na
+.\" The name should be formatted this way for apropos to work
+.SH NAME
+atomic_add, atomic_sub, atomic_inc, atomic_dec \- thread/SMP safe arithmetic on atomic data
+.\"
+.\"
+.SH SYNOPSIS
+.B #include <asm/atomic.h>
+.sp
+.BI "void atomic_add(int " i ", volatile atomic_t*" v ")"
+.br
+.BI "void atomic_sub(int " i ", volatile atomic_t*" v ")"
+.br
+.BI "void atomic_inc(volatile atomic_t*" v ")"
+.br
+.BI "void atomic_dec(volatile atomic_t*" v ")"
+.br
+.BI "int atomic_read(volatile atomic_t*" v ")"
+.br
+.BI "void atomic_set(volatile atomic_t*" v ", int " i ")"
+.br
+.BI "int atomic_dec_and_test(volatile atomic_t*" v ")"
+.HP
+.\"
+.\" Turn hyphenation and adjustment back on.
+.hlm 1
+.ad
+.SH DESCRIPTION
+These functions manipulate variables of type
+.B atomic_t
+is SMP and interrupt safe ways. These variables can be used to hold
+spin locks or SMP-safe reference counters. These functions guarantee
+that the operation that they represent is performed correctly. If
+necessary, hardware bus locking is performed to protect the operation.
+Usually, the CPU has some sort of atomic instructions that allow these
+operations to be performed quickly and safely.
+.PP
+The
+.B atomic_dec_and_test
+decrements the atomic variable, and returns true if the result is
+zero. This function is particularly useful in implementing spin locks
+on SMP systems.
+.SH "RETURN VALUE"
+The
+.B atomic_read
+function returns the integer value of the atomic variable.
+.PP
+The
+.B atomic_dec_and_test
+returns TRUE if the value becomes zero after the decrement.
+.SH AVAILABILITY
+Linux 2.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR intro "(9)"
+.ad
+.hlm 1
+.SH AUTHOR
+Stephen Williams <steve@icarus.com>
+.SH BUGS
+The read and set operations generally have no special protections.
diff --git a/upstream/mageia-cauldron/man9/cli.9 b/upstream/mageia-cauldron/man9/cli.9
new file mode 100644
index 00000000..5ec26161
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/cli.9
@@ -0,0 +1,37 @@
+.\" -*- nroff -*-
+.TH cli 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.SH NAME
+cli, sti \- disable/enable interrupts
+.SH SYNOPSIS
+.B #include  <asm/system.h>
+.sp
+.BI "extern void cli()"
+.br
+.BI "extern void sti()"
+.SH DESCRIPTION
+The
+.B cli
+function causes interrupts to be blocked on the host, so that
+following critical code may run uninterrupted. The
+.B sti
+function enables interrupts again.
+.PP
+Although it is possible to use cli/sti pairs to protect critical code,
+it is best to use
+.B cli
+with the
+.B save_flags
+macro. See
+.B save_flags "(9)."
+.SH "RETURN VALUE"
+none
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.BR save_flags "(9)"
+.PP
+.IR "include/asm-*/system.h"
+.SH AUTHOR
+Stephen Williams (steve@icarus.com)
+.SH BUGS
+The i386 ancestry of the names can be slightly confusing.
diff --git a/upstream/mageia-cauldron/man9/init_bh.9 b/upstream/mageia-cauldron/man9/init_bh.9
new file mode 100644
index 00000000..9a1aac8a
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/init_bh.9
@@ -0,0 +1,188 @@
+.\" -*- nroff -*-
+.TH INIT_BH 9 "1997/08/14 07:53:12" "Linux DDI 2.1.48" "Linux Kernel Functions"
+.\" copyright (C) 1997 Neil Moore <amethyst@maxwell.ml.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+
+.\" Turn off hyphentation
+.hlm 0
+.\" Turn off adjustment
+.na
+.SH NAME
+init_bh, remove_bh, mark_bh, disable_bh, enable_bh \- split-half interrupt handling
+.SH SYNOPSIS
+.B #include <linux/interrupt.h>
+.\" To keep all components of a declaration together, escape all white
+.\" space except after a comma.  In case the proc name plus the first
+.\" declaration is too long for one line, allow a break after '(' by
+.\" adding '\%'.
+.HP
+.BI "void\ init_bh(int " nr ", void\ (*" routine ")(void));"
+.\" We don't want an inter-paragraph space between the two declarations,
+.\" but we have to end the paragraph to stop the hanging indent.  .PD 0
+.\" gets rid of inter-paragraph spacing
+.PD 0
+.HP
+.BI "void\ remove_bh(int\ " nr ");"
+.HP
+.BI "void\ mark_bh(int\ " nr ");"
+.HP
+.BI "void\ disable_bh(int\ " nr ");"
+.HP
+.BI "void\ enable_bh(int\ " nr ");"
+.\" And set it back to normal now
+.PD 1
+.\" Back to standard adjustment
+.ad
+.\" Allow hypenation, omit if you want no hyphenation in the man page
+.hlm 1
+.SH DESCRIPTION
+.SS Theory
+.PP
+Split-half handling is a means of dividing an interrupt handler into two
+sections, one of which (known as the \(oqtop half') is time-critical and
+one of which (the \(oqbottom half') is not.
+.PP
+The top half (the handler registered with
+.BR "request_irq" "(9))"
+normally moves data between the device and a memory buffer, ensures that the
+device is in a sane state, and little else.  While the top half of a handler
+is running, the IRQ is question is blocked; if it is a fast interrupt handler
+(i.e., it has
+.B SA_INTERRUPT
+set), all interrupts are disabled.
+.PP
+The bottom half does whatever remains to be done.  Bottom halves run with
+interrupts enabled, although a locking mechanism ensures that only one
+bottom half will be running at a given time.  Bottom halves are run by
+.BR do_bottom_half() ,
+which is called from
+.B schedule()
+and
+.BR ret_from_sys_call() ". "
+.SS Usage
+.PP
+.B init_bh()
+installs
+.I routine()
+as bottom half number
+.IR nr ". "
+It operates by adding an entry to the
+.B bh_base[]
+table, and setting the appropriate bit of the
+.B bh_mask
+vector.  Rather than specifying a number explicitly, one should add an entry
+to the anonymous enum in
+.IR "include/linux/interrupt.h" ". "
+.PP
+.B remove_bh()
+removes bottom half number
+.I nr
+from the list of bottom halves.  It removes the entry from
+.B bh_base[]
+and clears the appropriate bit of
+.BR bh_mask ". "
+.PP
+.B mark_bh()
+requests that the kernel run the specified bottom half at the first available
+opportunity.  This function is normally called from the top half of an interrupt
+handler.  It operates by setting the appropriate bit of the
+.B bh_active
+vector.
+.PP
+.B disable_bh()
+disables bottom half number
+.I nr
+by clearing the appropriate bit of
+.BR bh_mask .
+This function also increments
+.BI bh_mask_count[ nr ]\fR,
+which is used to ensure that nested calls to
+.B disable_bh()
+must be matched by an equal number of calls to
+.BR enable_bh() .
+.PP
+.B enable_bh()
+enables a bottom half previously disabled by
+.BR disable_bh() .
+This function decrements
+.BI bh_mask_count[ nr ]\fR.
+Then, if that value is zero, the specified bottom half is enabled
+by setting the appropriate bit of
+.BR bh_mask .
+.SH "RETURN VALUE"
+No value is returned.
+.SH AVAILABILITY
+Linux 2.0+.
+.B init_bh()
+and
+.B remove_bh()
+were not present in older versions on Linux.  Under those versions,
+.B bh_base[]
+and
+.B bh_mask
+must be modified by hand.
+.SH "SEE ALSO"
+.BR request_irq "(9), " queue_task "(9)"
+.PP
+.IR include/asm*/softirq.h ", " include/linux/interrupt.h ", " kernel/softirq.c
+.PP
+\(lqKernel Korner\(rq in issue 26 of
+.I The Linux Journal
+includes a discussion of split-half interrupts under Linux.  An online copy of
+this article can be found at
+.BR http://www.ssc.com/lj/issue26/interrupt.html .
+.SH AUTHOR
+.RI "Neil Moore <" amethyst@maxwell.ml.org ">"
+.SH BUGS
+Only 32 bottom halves are allowed.  Increasing this number requires changing
+the size of
+.B bh_base[]
+and
+.BI bh_mask_count[]
+in
+.IR kernel/softirq.c,
+and changing
+.B bh_active
+and
+.B bh_mask
+(in the same file) to a larger type.  A better solution, however, would be to
+consolidate multiple bottom halves into a single one by using task queues.
+See
+.BR queue_task (9)
+for details.
+.\"
+.\" init_bh.9,v
+.\" Revision 1.4  1997/08/14 07:53:12  amethyst
+.\" Formatting changes.
+.\"
+.\" Revision 1.3  1997/08/08 16:40:50  amethyst
+.\" Updated for 2.1.48 -- init_bh() and friends are now available again on the PPC.
+.\"
+.\" Revision 1.2  1997/07/27 09:39:45  amethyst
+.\" Changed some wording.
+.\" Fixed a few typographical warts.
+.\" Added email address to copyright notice.
+.\"
+.\" Revision 1.1  1997/07/27 09:27:17  amethyst
+.\" Initial revision
+.\"
+.\"
diff --git a/upstream/mageia-cauldron/man9/init_module.9 b/upstream/mageia-cauldron/man9/init_module.9
new file mode 100644
index 00000000..c582c168
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/init_module.9
@@ -0,0 +1,90 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH init_module 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.\"
+.\" Turn off hyphenation and right justification for the first sections.
+.hlm 0
+.na
+.\" The name should be formatted this way for apropos to work
+.SH NAME
+init_module, cleanup_module \- module load and unload functions
+.\"
+.\"
+.SH SYNOPSIS
+.B #include <linux/module.h>
+.br
+.B #include <linux/modversions.h>
+.sp
+.BI "int init_module(void);"
+.br
+.BI "void cleanup_module(void);"
+.HP
+.\"
+.\" Turn hyphenation and adjustment back on.
+.hlm 1
+.ad
+.SH DESCRIPTION
+These functions are not part of the kernel but entry points into
+loadable modules. These are the only symbols that must be externally
+defined in order to load a module into a running kernel.
+.PP
+When a module is loaded into a running kernel, the
+.BR init_module "(9)"
+function within that object file is called to set up the module. The
+implementation of that function initializes local features and
+uses functions such as
+.BR register_chrdev "(9)"
+to attach itself to the kernel. It then returns zero(0) if it
+succeeds. If there is a problem or the module decides that it cannot
+be loaded, it returns instead an error code (i.e.
+.IR -ENODEV )
+and the kernel releases the module again.
+.PP
+Once loaded, the
+.BR cleanup_module "(9)"
+function of a module is used by the kernel to remove the module
+again. The module detaches itself from the kernel and returns.
+.SH "RETURN VALUE"
+The
+.BR init_module "(9)"
+function returns 0 on success, or an error code <0 if the module
+cannot be initialized.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR MOD_INC_USE_COUNT "(9), " insmod "(1)"
+.ad
+.hlm 1
+.PP
+Also list some source files for the
+kernel that implement the functions of the page.
+.SH AUTHOR
+Stephen Williams <steve@icarus.com>
+.SH BUGS
+It is entirely up to the driver to be sure it is detached from the
+kernel when unloaded. If a module makes an error in this regard, bad
+things may happen.
diff --git a/upstream/mageia-cauldron/man9/intro.9 b/upstream/mageia-cauldron/man9/intro.9
new file mode 100644
index 00000000..1b64e096
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/intro.9
@@ -0,0 +1,177 @@
+.\" -*- nroff -*-
+.TH intro 9 "July 1997" "Linux DDI" "Linux DDI"
+
+.\" Copyright 1997 Stephen Williams
+
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\"
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\"
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+
+.SH NAME
+intro \- Introduction to kernel interface
+.SH SYNOPSIS
+.B #include <linux/version.h>
+.SH DESCRIPTION
+This section documents the functions available to device driver
+writers and kernel level modules. The functions are of interest mainly
+to device driver writers, although anyone considering running code in
+linux kernel mode may need to be familiar with these interfaces.
+.PP
+Some of the functions of the DDI exist only in certain versions of the
+kernel. Use the
+.B LINUX_VERSION_CODE
+macro to test for specific versions of the kernel. For example, to use
+a feature that is new to 2.1, say:
+.PP
+.nf
+#if LINUX_VERSION_CODE >= 0x020100
+    ... use new stuff ...
+#else
+    ... do it the old way ...
+#endif
+.fi
+.PP
+The following is a list of the man pages, divided roughly into
+function groups.
+.SS "Kernel Functions"
+These are general kernel functions.
+.RS
+.TP
+MAJOR
+.TP
+MOD_INC_USE_COUNT
+.TP
+cli
+.TP
+init_bh
+.TP
+init_module
+.TP
+kmalloc
+.TP
+poll_wait
+.TP
+printk
+.TP
+probe_irq_on
+.TP
+register_chrdev
+.TP
+register_console
+.TP
+request_irq
+.TP
+save_flags
+.TP
+sleep_on
+.TP
+wake_up
+.RE
+.SS "/proc functions"
+These functions relate to manipulation of the
+.B /proc
+filesystem.
+.RS
+.TP
+proc_dir_entry
+.TP
+proc_net_register
+.TP
+proc_scsi_register
+.RE
+.SS "BIOS32 functions"
+These are specific to PCI (BIOS32) support.
+.RS
+.TP
+pcibios_find_class
+.TP
+pcibios_present
+.TP
+pcibios_read_config_byte
+.TP
+pcibios_read_config_dword
+.TP
+pcibios_read_config_word
+.TP
+pcibios_strerror
+.TP
+pcibios_write_config_byte
+.TP
+pcibios_write_config_dword
+.TP
+pcibios_write_config_word
+.RE
+.SS "VM functions"
+These are functions that support manipulating the virtual memory
+subsystem.
+.RS
+.TP
+MAP_NR
+.TP
+mem_map_reserve
+.RE
+.SS "Network Functions"
+.RS
+.TP
+skb_dequeue
+.TP
+skb_insert
+.TP
+skb_peek
+.TP
+skb_queue_empty
+.TP
+skb_queue_head
+.TP
+skb_queue_head_init
+.TP
+skb_queue_len
+.TP
+skb_queue_tail
+.TP
+skb_unlink
+.RE
+.SH AVAILABILITY
+Each man page attempts to list the kernel versions where the function
+is available. If the form of the function changes, this section tells
+when the described form applies.
+.SH "SEE ALSO"
+This section lists other man pages that may be of interest. Also,
+interesting source files in the linux kernel may be listed here.
+.SH AUTHORS
+Each man page has a section like this one that lists the author(s) who
+contributed significantly to that page. Other unnamed individuals may
+also have contributed corrections, editorial, etc.
+.PP
+Major contributors are (in alphabetical order)
+Cyrus Durgin <cider@speakeasy.org>,
+Niel Moore <amethyst@maxwell.ml.org>,
+Keith Owens <kaos@ocs.com.au>,
+Kirk Petersen <kirk@speakeasy.org>,
+and Stephen Williams <steve@icarus.com>.
+.PP
+Editorial, and this intro page, were done by Stephen Williams
+<steve@icarus.com>.
+.SH BUGS
+The living linux kernel is a moving target, and the kernel functions
+are unique to linux. Therefore, although the editor and contributers
+make a good effort to be as accurate as possible, errors may
+exist. The source codes of the linux kernel are the ultimate authority
+on the behavior of any function and should be considered the final
+word.
diff --git a/upstream/mageia-cauldron/man9/kmalloc.9 b/upstream/mageia-cauldron/man9/kmalloc.9
new file mode 100644
index 00000000..5bde08e5
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/kmalloc.9
@@ -0,0 +1,39 @@
+.\" -*- nroff -*-
+.TH kmalloc 9 "Aug 3, 1997" "Linux" "Kernel Functions"
+.SH NAME
+kmalloc, kfree \- allocate and free pieces of memory
+.SH SYNOPSIS
+.B #include <linux/malloc.h>
+.sp
+.BI "void * kmalloc (size_t " size ", int " priority ");"
+.br
+.BI "void kfree (void * " __ptr ");"
+.SH DESCRIPTION
+The
+.B kmalloc
+function allocates a piece of memory.
+.PP
+The parameter
+.I size
+is the number of bytes that will be allocated.  The parameter
+.I priority
+indicates the importance and type of the memory request.  Some
+possible values are GFP_DMA, GFP_ATOMIC, GFP_BUFFER, and GFP_NFS.
+.PP
+The
+.B kfree
+function releases a piece of memory that is passed as the
+.I __ptr
+parameter.
+.SH "RETURN VALUE"
+On success,
+.B kmalloc
+returns a pointer to the newly allocated memory. 
+.PP
+If there is an error, NULL is returned instead.
+.PP
+.SH AVAILABILITY
+Linux 2.0
+.SH AUTHOR
+Kirk Petersen (kirk@speakeasy.org)
+.SH BUGS
diff --git a/upstream/mageia-cauldron/man9/mem_map_reserve.9 b/upstream/mageia-cauldron/man9/mem_map_reserve.9
new file mode 100644
index 00000000..d46dae87
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/mem_map_reserve.9
@@ -0,0 +1,77 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH skel 9 "$Date:$" "Linux DDI" "VM Functions"
+.\"
+.\" Turn off hyphenation and right justification for the first sections.
+.hlm 0
+.na
+.\" The name should be formatted this way for apropos to work
+.SH NAME
+mem_map_reserve, mem_map_unreserve \- Manipulate flags of virtual memory pages.
+.\"
+.\"
+.SH SYNOPSIS
+.B #include <linux/mm.h>
+.br
+.B #include <linux/wrapper.h>
+.sp
+.BI "#define mem_map_reserve(" page_nr ")"
+.br
+.BI "#define mem_map_unreserve(" page_nr ")"
+.HP
+.\"
+.\" Turn hyphenation and adjustment back on.
+.hlm 1
+.ad
+.SH DESCRIPTION
+These macros cause a page to become reserved/unreserved. A reserved
+page is reserved from any further consideration by the linux kernel,
+meaning it is not scanned as potentially pageable, or available for
+page allocation. The kernel treats reserved pages as memory-mapped
+hardware.
+.PP
+It makes sense for a driver to mark a page reserved, for example, if
+the driver supports
+.BR mmap "(2)"
+with dynamically allocated pages that the target device can access via
+DMA.
+.SH "RETURN VALUE"
+None
+.SH AVAILABILITY
+Linux 2.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR MAP_NR "(9), " get_free_pages "(9), " mmap "(2)"
+.ad
+.hlm 1
+.SH AUTHOR
+Stephen Williams <steve@icarus.com>
+.SH BUGS
+If a page is marked reserved, a call to
+.BR free_page "(9)"
+will silenty ignore it. If the page was originally allocated by
+get_free_page, you must remember to unmark the page before releasing
+it to the system. Otherwise, the page is lost.
diff --git a/upstream/mageia-cauldron/man9/pcibios_find_class.9 b/upstream/mageia-cauldron/man9/pcibios_find_class.9
new file mode 100644
index 00000000..b18741e2
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/pcibios_find_class.9
@@ -0,0 +1,82 @@
+.\" -*- nroff -*-
+.TH pcibios_find_class 9 "August 5, 1997" "Linux DDI" "BIOS32 Functions"
+.hlm 0
+.na
+.SH NAME
+pcibios_find_class, pcibios_find_device \- find a PCI class or device
+.SH SYNOPSIS
+.B #include  <linux/bios32.h>
+.sp
+.TP
+.BI "int pcibios_find_class(unsigned int " class_code \
+", unsigned short " index ", unsigned char* " bus \
+", unsigned char* " device_fn ");"
+.TP
+.BI "int pcibios_find_device(unsigned\ short\ "\ vendor \
+", unsigned\ short\ " device_id\ \
+", unsigned\ short\ " index ", unsigned\ char*\ " bus \
+", unsigned\ char*\ " device_fn ");"
+.hlm 1
+.ad
+.SH DESCRIPTION
+The
+.B pcibios_find_class
+function searches for a certain class of device on the PCI bus.  In this
+instance, class is used to represent a few different categories of
+devices.  Some examples are storage, network, display, and memory
+classes. The classes are defined by the PCI Specification, and the
+.I class_code
+is matched with the dword in the configuration space of the device at
+offset 0x08.
+.PP
+The
+.B pcibios_find_device
+function finds a device on the PCI bus that has the matching device
+and vendor ids. The vendor ids are assigned to the PCI Sig to vendors,
+who in turn assign device ids to the devices they develop. The vendor
+and device ids are encoded in dword 0x00 of the configuration space.
+.PP
+Both
+.B pcibios_find_class
+and
+.B pcibios_find_device
+include an extra
+.I index
+parameter, which is used to select the specific
+device if there are multiple matches. An
+.I index
+of 0 matches the firts located board, 1 the second, and so on.
+.PP
+The
+.I bus
+and
+.I device_fn
+parameters are PCI specific cookies that are passed to other pcibios
+functions to access the configuration space of the located device.
+.SH "RETURN VALUE"
+On success, the functions return
+.IR PCIBIOS_SUCCESSFUL "."
+Otherwise, one of the following error codes is returned:
+.RS
+.TP
+PCIBIOS_DEVICE_NOT_FOUND
+There is no device that matches the search criteria,
+.TP
+PCIBIOS_BAD_VENDOR_ID
+The vendor id is invalid. (0xffff is not a valid id.)
+.TP
+PCIBIOS_FUNC_NOT_SUPPRTED
+The PCI subsystem is not available
+.RE
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.BR pcibios_present "(9)"
+.PP
+.BR /usr/include/linux/pci.h ", "
+.BR /usr/include/linux/bios32.h
+.PP
+.I "PCI Local Bus Specification"
+.SH AUTHOR
+Kirk Petersen (kirk@speakeasy.org)
+.SH BUGS
diff --git a/upstream/mageia-cauldron/man9/pcibios_present.9 b/upstream/mageia-cauldron/man9/pcibios_present.9
new file mode 100644
index 00000000..a9941d0f
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/pcibios_present.9
@@ -0,0 +1,28 @@
+.\" -*- nroff -*-
+.TH pcibios_present 9 "August 5, 1997" "Linux DDI" "BIOS32 Functions"
+.SH NAME
+pcibios_present \- determine whether a PCI bus is available
+.SH SYNOPSIS
+.B #include  <linux/bios32.h>
+.sp
+.BI "int pcibios_present(void);"
+.SH DESCRIPTION
+The
+.B pcibios_present
+tests for the presence of PCI support on the local host. It returns
+true (!0) if PCI support exists, false otherwise.
+.SH "RETURN VALUE"
+Zero if PCI support is NOT present.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR "/usr/include/linux/bios32.h" ","
+.BR "/usr/src/linux/arch/*/kernel/bios32.c"
+.ad
+.hlm 1
+.SH AUTHOR
+Kirk Petersen (kirk@speakeasy.org)
+.SH BUGS
+Bugs?! This is way to trivial to have bugs.
diff --git a/upstream/mageia-cauldron/man9/pcibios_read_config_byte.9 b/upstream/mageia-cauldron/man9/pcibios_read_config_byte.9
new file mode 100644
index 00000000..13bb6311
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/pcibios_read_config_byte.9
@@ -0,0 +1,35 @@
+.\" -*- nroff -*-
+.TH pcibios_read_config_byte 9 "August 5, 1997" "Linux DDI" "BIOS32 Functions"
+.SH NAME
+pcibios_read_config_byte \- read one byte of data from the configuration
+space of the PCI bus
+.SH SYNOPSIS
+.B #include  <linux/bios32.h>
+.sp
+.BI "int pcibios_read_config_byte (unsigned char bus, unsigned char device_fn,\
+unsigned char where, unsigned char * value);"
+.SH DESCRIPTION
+The
+.B pcibios_read_config_byte
+function reads one byte from the configuration space of the PCI bus.  The
+bus can be specified with the
+.B bus
+parameters.  The
+.B device_fn
+paramter determines which ? to use.
+.B where
+is set to the ?.  After the call is made,
+.B value
+points to the value that was read.
+.SH "RETURN VALUE"
+The return value is taken from the PCI controller in a way that I'm
+not quite sure of.
+.PP
+If the PCI system is unavailable, PCIBIOS_FUNC_NOT_SUPPORTED is
+returned.
+.SH AVAILABILITY
+Linux 1.0?
+.SH AUTHOR
+Kirk Petersen (kirk@speakeasy.org)
+.SH BUGS
+
diff --git a/upstream/mageia-cauldron/man9/pcibios_read_config_dword.9 b/upstream/mageia-cauldron/man9/pcibios_read_config_dword.9
new file mode 100644
index 00000000..45335114
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/pcibios_read_config_dword.9
@@ -0,0 +1,34 @@
+.\" -*- nroff -*-
+.TH pcibios_read_config_dword 9 "August 5, 1997" "Linux DDI" "BIOS32 Functions"
+.SH NAME
+pcibios_read_config_dword \- read a double word of data from the configuration
+space of the PCI bus
+.SH SYNOPSIS
+.B #include  <linux/bios32.h>
+.sp
+.BI "int pcibios_read_config_dword (unsigned char bus, unsigned char device_fn,\
+unsigned char where, unsigned int * value);"
+.SH DESCRIPTION
+The
+.B pcibios_read_config_dword
+function reads a double word from the configuration space of the PCI bus.  The
+bus can be specified with the
+.B bus
+parameters.  The
+.B device_fn
+paramter determines which ? to use.
+.B where
+is set to the ?.  After the call is made,
+.B value
+points to the value that was read.
+.SH "RETURN VALUE"
+The return value is taken from the PCI controller in a way that I'm
+not quite sure of.
+.PP
+If the PCI system is unavailable, PCIBIOS_FUNC_NOT_SUPPORTED is
+returned.
+.SH AVAILABILITY
+Linux 1.0?
+.SH AUTHOR
+Kirk Petersen (kirk@speakeasy.org)
+.SH BUGS
diff --git a/upstream/mageia-cauldron/man9/pcibios_read_config_word.9 b/upstream/mageia-cauldron/man9/pcibios_read_config_word.9
new file mode 100644
index 00000000..faed7b42
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/pcibios_read_config_word.9
@@ -0,0 +1,34 @@
+.\" -*- nroff -*-
+.TH pcibios_read_config_word 9 "August 5, 1997" "Linux DDI" "BIOS32 Functions"
+.SH NAME
+pcibios_read_config_word \- read a word of data from the configuration space
+of the PCI bus
+.SH SYNOPSIS
+.B #include  <linux/bios32.h>
+.sp
+.BI "int pcibios_read_config_word (unsigned char bus, unsigned char device_fn,\
+unsigned char where, unsigned short * value);"
+.SH DESCRIPTION
+The
+.B pcibios_read_config_word
+function reads a word from the configuration space of the PCI bus.  The
+bus can be specified with the
+.B bus
+parameters.  The
+.B device_fn
+paramter determines which ? to use.
+.B where
+is set to the ?.  After the call is made,
+.B value
+points to the value that was read.
+.SH "RETURN VALUE"
+The return value is taken from the PCI controller in a way that I'm
+not quite sure of.
+.PP
+If the PCI system is unavailable, PCIBIOS_FUNC_NOT_SUPPORTED is
+returned.
+.SH AVAILABILITY
+Linux 1.0?
+.SH AUTHOR
+Kirk Petersen (kirk@speakeasy.org)
+.SH BUGS
diff --git a/upstream/mageia-cauldron/man9/pcibios_strerror.9 b/upstream/mageia-cauldron/man9/pcibios_strerror.9
new file mode 100644
index 00000000..290819ce
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/pcibios_strerror.9
@@ -0,0 +1,30 @@
+.\" -*- nroff -*-
+.TH pcibios_strerror 9 "August 5, 1997" "Linux DDI" "BIOS32 Functions"
+.SH NAME
+pcibios_strerror \- Convert BIOS32 return codes to strings
+.SH SYNOPSIS
+.B #include  <linux/bios32.h>
+.sp
+.BI "const\ char*\ pcibios_strerror(\%int\ " error\%  ");"
+.SH DESCRIPTION
+The
+.B pcibios_strerror
+function converts a PCI BIOS32 error number into a human readable error
+message. The bios32 error codes are defined by the PCI standard.
+.SH "RETURN VALUE"
+The return value is a pointer to the error message.
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR /usr/include/linux/bios32.h ,
+.BR /usr/src/linux/arch/*/kernel/bios32.c
+.PP
+.I "PCI Local Bus -- PCI BIOS SPECIFICATION"
+.ad
+.SH AVAILABILITY
+Linux 1.0?
+.SH AUTHOR
+Kirk Petersen (kirk@speakeasy.org)
+.SH BUGS
+The result may or may not be overwritten by a subsequent call to this
+function. it is best to assume it does.
diff --git a/upstream/mageia-cauldron/man9/pcibios_write_config_byte.9 b/upstream/mageia-cauldron/man9/pcibios_write_config_byte.9
new file mode 100644
index 00000000..6c111c7c
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/pcibios_write_config_byte.9
@@ -0,0 +1,35 @@
+.\" -*- nroff -*-
+.TH pcibios_write_config_byte 9 "August 5, 1997" "Linux DDI" "BIOS32 Functions"
+.SH NAME
+pcibios_write_config_byte \- write one byte of data to the configuration
+space of the PCI bus
+.SH SYNOPSIS
+.B #include  <linux/bios32.h>
+.sp
+.BI "int pcibios_write_config_byte(unsigned char bus, unsigned char device_fn,\
+unsigned char where, unsigned char value);"
+.SH DESCRIPTION
+The
+.B pcibios_write_config_byte
+function writes one byte to the configuration space of the PCI bus.  The
+bus can be specified with the
+.B bus
+parameters.  The
+.B device_fn
+paramter determines which ? to use.
+.B where
+is set to the ?.  The byte to be written to the configuration space is
+stored in the
+.B value
+parameter.
+.SH "RETURN VALUE"
+The return value is taken from the PCI controller in a way that I'm
+not quite sure of.
+.PP
+If the PCI system is unavailable, PCIBIOS_FUNC_NOT_SUPPORTED is
+returned.
+.SH AVAILABILITY
+Linux 1.0?
+.SH AUTHOR
+Kirk Petersen (kirk@speakeasy.org)
+.SH BUGS
diff --git a/upstream/mageia-cauldron/man9/pcibios_write_config_dword.9 b/upstream/mageia-cauldron/man9/pcibios_write_config_dword.9
new file mode 100644
index 00000000..5e317b03
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/pcibios_write_config_dword.9
@@ -0,0 +1,35 @@
+.\" -*- nroff -*-
+.TH pcibios_write_config_dword 9 "August 5, 1997" "Linux DDI" "BIOS32 Functions"
+.SH NAME
+pcibios_write_config_dword \- write one double word of data to the
+configuration space of the PCI bus
+.SH SYNOPSIS
+.B #include  <linux/bios32.h>
+.sp
+.BI "int pcibios_write_config_dword (unsigned char bus, unsigned char device_fn,\
+unsigned char where, unsigned int value);"
+.SH DESCRIPTION
+The
+.B pcibios_write_config_dword
+function writes one double word to the configuration space of the PCI bus.  The
+bus can be specified with the
+.B bus
+parameters.  The
+.B device_fn
+paramter determines which ? to use.
+.B where
+is set to the ?.  The word to be written to the configuration space is
+stored in the
+.B value
+parameter.
+.SH "RETURN VALUE"
+The return value is taken from the PCI controller in a way that I'm
+not quite sure of.
+.PP
+If the PCI system is unavailable, PCIBIOS_FUNC_NOT_SUPPORTED is
+returned.
+.SH AVAILABILITY
+Linux 1.0?
+.SH AUTHOR
+Kirk Petersen (kirk@speakeasy.org)
+.SH BUGS
diff --git a/upstream/mageia-cauldron/man9/pcibios_write_config_word.9 b/upstream/mageia-cauldron/man9/pcibios_write_config_word.9
new file mode 100644
index 00000000..d1d9516b
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/pcibios_write_config_word.9
@@ -0,0 +1,35 @@
+.\" -*- nroff -*-
+.TH pcibios_write_config_word 9 "August 5, 1997" "Linux DDI" "BIOS32 Functions"
+.SH NAME
+pcibios_write_config_word \- write one word of data to the configuration
+space of the PCI bus
+.SH SYNOPSIS
+.B #include  <linux/bios32.h>
+.sp
+.BI "int pcibios_write_config_word (unsigned char bus, unsigned char device_fn,\
+unsigned char where, unsigned short value);"
+.SH DESCRIPTION
+The
+.B pcibios_write_config_word
+function writes one word to the configuration space of the PCI bus.  The
+bus can be specified with the
+.B bus
+parameters.  The
+.B device_fn
+paramter determines which ? to use.
+.B where
+is set to the ?.  The word to be written to the configuration space is
+stored in the
+.B value
+parameter.
+.SH "RETURN VALUE"
+The return value is taken from the PCI controller in a way that I'm
+not quite sure of.
+.PP
+If the PCI system is unavailable, PCIBIOS_FUNC_NOT_SUPPORTED is
+returned.
+.SH AVAILABILITY
+Linux 1.0?
+.SH AUTHOR
+Kirk Petersen (kirk@speakeasy.org)
+.SH BUGS
diff --git a/upstream/mageia-cauldron/man9/poll_wait.9 b/upstream/mageia-cauldron/man9/poll_wait.9
new file mode 100644
index 00000000..9fc3d5a9
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/poll_wait.9
@@ -0,0 +1,76 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH poll_wait 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.\"
+.\" Turn off hyphenation and right justification for the first sections.
+.hlm 0
+.na
+.\" The name should be formatted this way for apropos to work
+.SH NAME
+poll_wait \- wait for selectable event to be ready
+.\"
+.\"
+.SH SYNOPSIS
+.B #include <linux/poll.h>
+.PP
+.BI "void poll_wait(struct\%wait_queue**" sync ", poll_table*" pt ")"
+.HP
+.\"
+.\" Turn hyphenation and adjustment back on.
+.hlm 1
+.ad
+.SH DESCRIPTION
+This function is used in support of the poll device driver entry
+point. The intent is for a device driver to put
+.I sync
+into the poll table immediately entering the device poll routine, then
+returning a bit mask of events that are currently ready. The kernel
+looks at the mask of events to see if something it needs is ready, and
+suspends the process if not.
+.PP
+.nf
+static unsigned int xxpoll(struct file*file, poll_table*pt)
+{
+    poll_wait(&data_in_sync, pt);
+    if (data_in.cnt > 0) return POLLIN | POLLRDNORM;
+    else return 0;
+}
+.fi
+.PP
+This example shows the basic use of
+.BR poll_wait "(9)"
+in a device driver poll function.
+.SH AVAILABILITY
+Linux kernel 2.1.34+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR  intro "(9)"
+.ad
+.hlm 1
+.SH AUTHOR
+Stephen Williams <steve@icarus.com>
+.SH BUGS
+To early to say.
diff --git a/upstream/mageia-cauldron/man9/printk.9 b/upstream/mageia-cauldron/man9/printk.9
new file mode 100644
index 00000000..892812d9
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/printk.9
@@ -0,0 +1,78 @@
+.\" -*- nroff -*-
+.TH printk 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.SH NAME
+printk \- print messages to console log
+.SH SYNOPSIS
+.B #include <linux/kernel.h>
+.sp
+.BI "int printk(const char*" fmt ", ...)"
+.SH DESCRIPTION
+Print a formatted message to the kernel console, much like the
+.B printf
+function of the stdio library. Normally, the message is written to the
+physical console device of the computer, although this behavior can be
+changed with the
+.B register_console
+function. Messages are also stored in a message log book.
+.PP
+The generated string may also start with a message priority code,
+which sets the priority of the message. The priority code strings are
+of the form
+.I "<n>"
+where n is a number from 0 - 7. The following macros are defined in
+the
+.I "<linux/kernel.h>"
+header file:
+.RS
+.TP
+KERN_EMERG
+System is unuseable
+.TP
+KERN_ALERT
+Action must be taken immediately
+.TP
+KERN_CRIT
+Critical conditions
+.TP
+KERN_ERR
+Error conditions
+.TP
+KERN_WARNING
+Warning conditions
+.TP
+KERN_NOTICE
+Normal but significant condition
+.TP
+KERN_INFO
+Informational
+.TP
+KERN_DEBUG
+Debug-level messages
+.RE
+
+For example
+.nf
+
+	printk(KERN_NOTICE "Hello, world.\\n");
+
+.fi
+does the expected thing.
+.SH "RETURN VALUE"
+Returns the number of characters written to the log.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.BR register_console "(9), " syslog "(2)"
+.PP
+.IR "kernel/printk.c"
+.SH AUTHOR
+Stephen Williams (steve@icarus.com)
+.SH BUGS
+float and double formats are not supported. Floats and doubles do not
+belong inside the kernel anyhow.
+.PP
+The
+.B printk
+implementation protects itself from interruption, so in principle it
+can be used in interrupts handlers and critical sections. However,
+there are no guarantees about the console function that is registered.
diff --git a/upstream/mageia-cauldron/man9/probe_irq_on.9 b/upstream/mageia-cauldron/man9/probe_irq_on.9
new file mode 100644
index 00000000..2eee7461
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/probe_irq_on.9
@@ -0,0 +1,144 @@
+.\" -*- nroff -*-
+.TH PROBE_IRQ_ON 9 "1997/08/14 07:53:32" "Linux DDI 2.1.47" "Linux Kernel Functions"
+.\" Copyright (C) 1997 Neil Moore <amethyst@maxwell.ml.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+
+.\" Turn off hyphentation
+.hlm 0
+.\" Turn off adjustment
+.na
+.SH NAME
+probe_irq_on, probe_irq_off \- safe probing for IRQs
+.SH SYNOPSIS
+.B #include <linux/interrupt.h>
+.\" To keep all components of a declaration together, escape all white
+.\" space except after a comma.  In case the proc name plus the first
+.\" declaration is too long for one line, allow a break after '(' by
+.\" adding '\%'.
+.HP
+.B "unsigned\ long\ probe_irq_on(void)"
+.\" We don't want an inter-paragraph space between the two declarations,
+.\" but we have to end the paragraph to stop the hanging indent.  .PD 0
+.\" gets rid of inter-paragraph spacing
+.PD 0
+.HP
+.BI "int\ probe_irq_off(unsigned\ long " irqs ");"
+.\" And set it back to normal now
+.PD 1
+.\" Back to standard adjustment
+.ad
+.\" Allow hypenation, omit if you want no hyphenation in the man page
+.hlm 1
+.SH DESCRIPTION
+.SS Usage
+.PP
+.B probe_irq_on()
+turns on IRQ detection.  It operates by enabling all interrupts which have
+no handlers, while keeping the handlers for those interrupts NULL.  The
+kernel's generic interrupt handling routine will disable these IRQs when
+an interrupt is received on them.  
+.B probe_irq_on()
+adds each of these IRQ numbers to a vector which it will return.  It waits
+approximately 100ms for any spurious interrupts that may occur, and masks
+these from its vector; it then returns this vector to its caller.
+.PP
+.B probe_irq_off()
+tests an internal list of enabled IRQs against its
+.I irqs
+parameter, which should be the value returned by the last
+.BR probe_irq_on() .
+This function basically detects which IRQs have been switched off, and thus
+which ones have received interrupts.
+.SS Example
+This explanation may seem a bit confusing, so here is an example of code
+the mythical FUBAR 2000 driver could use to probe for IRQs:
+.RS
+.PP
+unsigned long irqs;
+.br
+int irq;
+.sp
+irqs = probe_irq_on();
+.sp
+outb(FB2K_GIVE_ME_AN_INTERRUPT_OR_GIVE_ME_DEATH,
+     FB2K_CONTROL_PORT);
+.br
+  /* the interrupt could take a while to occur */
+.br
+udelay(1000);
+.sp
+irq = probe_irq_off(irqs);
+.sp
+if (irq == 0) {
+.br
+	printk("fb2k: could not detect IRQ.\en");
+.br
+	printk("fb2k: Installation failed.\en");
+.br
+} else if (irq == -1) {
+.br
+	printk("fb2k: multiple IRQs detected.\en");
+.br
+	printk("fb2k: Installation failed.\en");
+.br
+} else {
+.br
+	fb2k_dev->irq = irq;
+.br
+	printk("fb2k: using probed IRQ %d.\en", irq);
+.br
+}
+.br
+.SH "RETURN VALUE"
+.B probe_irq_on()
+returns a bitmap of all unhandled IRQs (except those which are receiving
+spurious interrupts).  This value should only be used as a parameter to
+the next call to
+.BR probe_irq_off() .
+.PP
+.B probe_irq_off()
+returns the IRQ number of whichever unhandled interrupt has occurred since
+the last
+.BR probe_irq_on() .
+If no interrupts have occurred on any of the marked IRQs, 0 is returned;
+if interrupts have occurred on more than one of these IRQs, -1 is returned.
+.SH AVAILABILITY
+Linux 1.2+.  These functions are not available on m68k-based machines.
+.SH "SEE ALSO"
+.BR request_irq "(9)"
+.PP
+.I arch/*/kernel/irq.c
+.SH AUTHOR
+.RI "Neil Moore <" amethyst@maxwell.ml.org ">"
+.SH BUGS
+As mentioned above, these functions are not available on m68k-based machines.
+.PP
+This manpage is way too confusing.
+.\"
+.\" probe_irq_on.9,v
+.\" Revision 1.2  1997/08/14 07:53:32  amethyst
+.\" Formatting changes.
+.\"
+.\" Revision 1.1  1997/07/28 10:28:34  amethyst
+.\" Initial revision
+.\"
+.\"
diff --git a/upstream/mageia-cauldron/man9/proc_dir_entry.9 b/upstream/mageia-cauldron/man9/proc_dir_entry.9
new file mode 100644
index 00000000..88e95e6c
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/proc_dir_entry.9
@@ -0,0 +1,163 @@
+.\"
+.TH proc_dir_entry 9 "July 1997" "Linux 2.0.30" "/proc Functions"
+.LO 9
+.hlm 0
+.na
+.SH NAME
+proc_dir_entry, proc_register, proc_register_dynamic, proc_unregister \- register entries in the /proc filesystem.
+.SH SYNOPSIS
+.B #include <linux/proc_fs.h>
+.TP
+.BI "int\ proc_register(\%struct\ proc_dir_entry\ *\ " "parent" ", struct\ proc_dir_entry\ *\ " "child");
+.TP
+.BI "int\ proc_unregister(\%struct\ proc_dir_entry\ *\ " "parent" ", int\ " "inode");
+.TP
+.BI "int\ proc_register_dynamic(\%struct\ proc_dir_entry\ *\ " "parent" ", struct\ proc_dir_entry\ *\ " "child");
+.ad
+.SH DESCRIPTION
+The
+.B proc_register
+functions add file or directory entries to the /proc file system.  They
+associate processing routines with each node of the /proc tree.
+The structure
+.B proc_dir_entry
+is defined as
+.nf
+
+.B
+struct proc_dir_entry {
+.B
+        unsigned short low_ino;
+.B
+        unsigned short namelen;
+.B
+        const char *name;
+.B
+        mode_t mode;
+.B
+        nlink_t nlink;
+.B
+        uid_t uid;
+.B
+        gid_t gid;
+.B
+        unsigned long size;
+.B
+        struct inode_operations * ops;
+.B
+        int (*get_info)(char *buffer, char **start,
+.B
+                off_t offset, int length, int unused);
+.B
+        void (*fill_inode)(struct inode *);
+.B
+        struct proc_dir_entry *next, *parent, *subdir;
+.B
+        void *data;
+.B
+};
+.fi
+
+.IP low_ino 8
+The inode number of this directory entry.  For
+.B proc_register
+this  number should be unique within the /proc filesystem, values are
+defined in
+.IR <linux/proc_fs.h> .
+For
+.B proc_register_dynamic
+the inode number is dynamically assigned.
+.IP namelen
+The length of the name, excluding the trailing null.
+.IP name
+The name of this node.
+.IP mode
+.ne 3
+The node's type and permissions.  Drawn from
+.IR <linux/stat.h> .
+.IP nlink
+Number of links to the node.  Initialise to 2 if mode includes
+S_IFDIR, 1 otherwise.
+.IP uid
+The uid that owns the node, normally 0.
+.IP gid
+The gid that owns the node. normally 0.
+.IP size
+Sets the size of the node, the value will appear as the inode size in
+listings and be returned by
+.BR stat .
+Unless you really need a size, set this to zero.
+.IP ops
+Defines the set of inode operations to perform for your /proc node.
+For a directory node, use
+.I &proc_dir_inode_operations
+unless you have special requirements.  For a leaf node, set to NULL
+unless you have special requirements.
+.IP get_info
+If defined, this proc is called when the node is read.  Should be NULL
+for directory nodes.
+.B NOTE:
+If you need to return large amounts of data, the proc must return the
+data in chunks and reposition itself on the next call, using the
+.I offset
+variable.  See
+.I ip_masq_procinfo
+for example code with large output.
+.IP fill_inode
+Dynamically fill in the inode characteristics during directory
+operations.  Not normally required and set to NULL.  See
+proc_pid_fill_inode for example code.
+.IP "next, parent, subdir
+Maintained by /proc routines.  Initial value is irrelevant, set to NULL.
+.IP data
+An opaque pointer which can be used by proc handlers to pass local data
+around.  Set to whatever you like when calling
+.BR proc_register ,
+normally NULL.  This pointer is copied into the inode u.ip_generic
+field (by proc_get_inode) so it is available to any proc routines that
+are passed an inode.
+.PP
+.B proc_register
+adds the
+.B child
+as a node under the
+.BR parent .
+.PP
+.B proc_register_dynamic
+dynamically assigns an inode number then adds the
+.B child
+as a node under the
+.BR parent .
+.PP
+.B proc_unregister
+scans the inode list under the
+.B parent
+for the specified
+.B inode
+number and removes the matching entry.
+.SH "RETURN VALUE"
+.PP
+.ne 2
+.B proc_register
+always returns 0.
+.PP
+.B proc_register_dynamic
+.ne 2
+returns 0 for success or
+.B -EAGAIN
+if there are no free dynamic inode numbers.
+.PP
+.B proc_unregister
+.ne 2
+returns 0 for success or
+.B -EINVAL
+if the node was not found.
+.SH "SEE ALSO"
+.na
+.BR proc_net_register "(9), " proc_net_unregister "(9), " proc_scsi_register "(9), " proc_scsi_unregister "(9), " stat "(2)."
+.ad
+.SH AUTHOR
+Keith Owens <kaos@ocs.com.au>
+.SH BUGS
+The uniqueness of /proc inode numbers is assumed, not enforced.  It is
+possible to add two nodes with the same inode number.
diff --git a/upstream/mageia-cauldron/man9/proc_net_register.9 b/upstream/mageia-cauldron/man9/proc_net_register.9
new file mode 100644
index 00000000..436620e6
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/proc_net_register.9
@@ -0,0 +1,30 @@
+.\"
+.TH proc_net_register 9 "July 1997" "Linux 2.0.30" "/proc Functions"
+.LO 9
+.hlm 0
+.na
+.SH NAME
+proc_net_register, proc_net_unregister \- register network entries in the /proc filesystem
+.SH SYNOPSIS
+.B #include <linux/proc_fs.h>
+.TP
+.BI "int\ proc_net_register(\%struct\ proc_dir_entry\ *\ " "child");
+.TP
+.BI "int\ proc_net_unregister(\%int\ " "inode");
+.ad
+
+.SH DESCRIPTION
+These are wrapper functions around
+.B proc_register
+and
+.BR proc_unregister .
+They always use a parent of proc_net.
+.SH "RETURN VALUE"
+As for
+.B proc_register
+and
+.BR proc_unregister .
+.SH "SEE ALSO"
+.BR proc_register "(9), " proc_unregister "(9)"
+.SH AUTHOR
+Keith Owens <kaos@ocs.com.au>
diff --git a/upstream/mageia-cauldron/man9/proc_scsi_register.9 b/upstream/mageia-cauldron/man9/proc_scsi_register.9
new file mode 100644
index 00000000..cccda817
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/proc_scsi_register.9
@@ -0,0 +1,50 @@
+.\"
+.TH proc_scsi_register 9 "July 1997" "Linux 2.0.30" "/proc Functions"
+.LO 9
+.hlm 0
+.na
+.SH NAME
+proc_scsi_register, proc_scsi_unregister \- register SCSI entries in the /proc filesystem
+.SH SYNOPSIS
+.B #include <linux/proc_fs.h>
+.TP
+.BI "int\ proc_scsi_register(\%struct\ proc_dir_entry\ *\ " "driver" ", struct\ proc_dir_entry\ *\ " "child");
+.TP
+.BI "int\ proc_scsi_unregister(\%struct\ proc_dir_entry\ *\ " "driver" ", int\ " "inode");
+.ad
+
+.SH DESCRIPTION
+These are wrapper functions around
+.B proc_register
+and
+.BR proc_unregister .
+.PP
+.B proc_scsi_register
+always sets the
+.I ops
+field of
+.B child
+to proc_scsi_inode_operations.  If the child inode number is less than
+PROC_SCSI_FILE, the child is registered with a parent of proc_scsi and
+.B driver
+is ignored.
+Otherwise the child is registered with a parent of
+.BR driver .
+.PP
+.BR proc_scsi_unregister .
+If the child inode number is less than PROC_SCSI_FILE, the child is
+unregistered with a parent of proc_scsi and
+.B driver
+is ignored.
+Otherwise the child is unregistered with a parent of
+.B driver
+and scsi_init_free is called on the child.
+.SH "RETURN VALUE"
+As for
+.B proc_register
+and
+.BR proc_unregister .
+.SH "SEE ALSO"
+.BR proc_register "(9), " proc_unregister "(9)"
+.SH AUTHOR
+Keith Owens <kaos@ocs.com.au>
diff --git a/upstream/mageia-cauldron/man9/register_chrdev.9 b/upstream/mageia-cauldron/man9/register_chrdev.9
new file mode 100644
index 00000000..c357aaa4
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/register_chrdev.9
@@ -0,0 +1,68 @@
+.\" -*- nroff -*-
+.TH register_chrdev 9 "$Date:$" "Linux 2.0" "Kernel Functions"
+.SH NAME
+register_chrdev, unregister_chrdev \- register a driver major number
+.SH SYNOPSIS
+.B #include <linux/fs.h>
+.sp
+.BI "int register_chrdev(unsigned int " major ", const char*" name ","
+.BI "struct file_operations*" ops ");"
+.br
+.BI "int unregister_chrdev(unsigned int " major ", const char *" name ");"
+.SH DESCRIPTION
+The
+.B register_chrdev
+function associates a character major number with set of driver entry
+points. The file_operations structure contains pointers to functions
+that the the driver uses to implement the kernel interface to the
+driver.
+.PP
+The paramter
+.I major
+is the character major number assigned to the device driver and to be
+mapped to the function table. The
+.I name
+parameter is a short name for the device and is displayed in the
+The
+.I "/proc/devices"
+list. It also must exactly match the name passed to
+.B unregister_chrdev
+function when releasing the functions.
+.PP
+A device driver module may register as many different major numbers as
+it supports, though this is not typically done.
+.PP
+The
+.B unregister_chrdev
+function releases the major number, and is normally called in the
+module_cleanup function to remove the driver from the kernel.
+.SH "RETURN VALUE"
+On success,
+.B register_chrdev
+returns 0 if
+.I major
+is a number other then 0, otherwise Linux will choose a major number
+and return the chosen value.
+.PP
+If there is an error, one of the following codes is returned instead:
+.RS
+.TP
+-EINVAL
+The specified number is not valid (> MAX_CHRDEV)
+.TP
+-EBUSY
+The major number is busy
+.RE
+.PP
+The
+.B unregister_chrdev
+function will return 0 if successful, or
+.B "-EINVAL"
+if the major number is not registered with the matching name.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.BR register_blkdev "(9) " mknod "(2)"
+.SH AUTHOR
+Stephen Williams (steve@icarus.com)
+.SH BUGS
diff --git a/upstream/mageia-cauldron/man9/register_console.9 b/upstream/mageia-cauldron/man9/register_console.9
new file mode 100644
index 00000000..a48fdf28
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/register_console.9
@@ -0,0 +1,52 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH register_console 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.SH NAME
+register_console \- take on kernel console duties
+.SH SYNOPSIS
+.BI "extern void register_console(void(*" proc ")(const char*))"
+.SH DESCRIPTION
+This function registers with the kernel a function to display all the
+kernel messages printed by the
+.BR printk "(9)"
+function.
+.PP
+This function is normally called once at boot time by the device
+driver for the console device, whatever that turns out to be for the
+machine. The console procedure is only called to print messages that
+are to be printed, the filtering for log level is done already.
+.SH "RETURN VALUE"
+None.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.BR printk "(9)"
+.PP
+.I "/usr/src/linux/kernel/printk.c"
+.SH AUTHOR
+Stephen Williams <steve@icarus.com>
+.SH BUGS
+Only one console can exist at a time, and there is no way to know if
+any other device has registered as a console.
diff --git a/upstream/mageia-cauldron/man9/request_irq.9 b/upstream/mageia-cauldron/man9/request_irq.9
new file mode 100644
index 00000000..c55e72c8
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/request_irq.9
@@ -0,0 +1,305 @@
+.\" -*- nroff -*-
+.TH REQUEST_IRQ 9 "1997/08/14 07:53:47" "Linux DDI 2.1.47" "Linux Kernel Functions"
+.\" copyright (C) 1997 Neil Moore <amethyst@maxwell.ml.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+
+.\" Turn off hyphentation
+.hlm 0
+.\" Turn off adjustment
+.na
+.SH NAME
+request_irq, free_irq \- register an interrupt handler
+.SH SYNOPSIS
+.B #include <asm/irq.h>
+.br
+.B #include <linux/signal.h>
+.br
+.B #include <linux/sched.h>
+.br
+.B #include <linux/interrupt.h>
+.\" To keep all components of a declaration together, escape all white
+.\" space except after a comma.  In case the proc name plus the first
+.\" declaration is too long for one line, allow a break after '(' by
+.\" adding '\%'.
+.HP
+.BI "int request_irq(unsigned\ int\ " irq ","
+.BI "void\ (*" handler ")(int,\ void\ *,\ struct\ pt_regs\ *),"
+.BI "unsigned\ long " irqflags ", const\ char\ *" devname ","
+.BI "void\ *" dev_id ");"
+.\" We don't want an inter-paragraph space between the two declarations,
+.\" but we have to end the paragraph to stop the hanging indent.  .PD 0
+.\" gets rid of inter-paragraph spacing
+.PD 0
+.PP
+.BI "void\ free_irq(unsigned\ int " irq ", void\ *" dev_id ");"
+.\" And set it back to normal now
+.PD 1
+.\" Back to standard adjustment
+.ad
+.\" Allow hypenation, omit if you want no hyphenation in the man page
+.hlm 1
+.SH DESCRIPTION
+.SS Usage
+The
+.B request_irq()
+function requests that a specified function (the handler) be called
+whenever the kernel receives a given interrupt.  The handler may
+in turn register a bottom half, which is usually a slower part of
+the handler which does not need to be executed as soon as the interrupt
+is received.  See
+.BR init_bh "(9) "
+for more information on bottom halves.
+.PP
+The
+.I irq
+parameter is the interrupt number you want to handle.  It must be less
+than
+.B NR_IRQS
+(16 on Intel systems), and there may be additional limitations on the
+value.  See
+.IR arch/*/kernel/irq.c " (" intr.c
+on m68k machines) for
+more information.
+.PP
+.I handler
+is a pointer to the a pointer to the function that will handle the
+interrupt.  The handler is passed the following parameters:
+.RS
+.TP
+.BI int " irq"
+The interrupt number.  By testing the value of this parameter, it is
+possible for a single function to handle multiple IRQs.
+.TP
+.BI "void *" dev_id
+The device ID of this handler (see below).
+.TP
+.BI "struct pt_regs *" regs
+The registers stored on the stack of the process that was interrupted.
+Normally, one shouldn't mess with these, although they can be read to
+determine, for example, whether the interrupted process was in kernel
+or user mode.
+.RE
+.PP
+.I irqflags
+is, as the name suggests, a bitmask of flags pertaining to this interrupt
+handler.  Legal bits are:
+.RS
+.TP
+.B SA_INTERRUPT
+This bit indicates that you are registering a fast interrupt handler.
+The semantics of this are discussed below.
+.TP
+.B SA_SHIRQ
+This bit indicates that your handler supports sharing an IRQ with other
+handlers (see also
+.RI * dev_id
+below).
+.TP
+.B SA_SAMPLE_RANDOM
+This bit indicates that this IRQ may be used as an entropy source for
+.I /dev/random
+and
+.I /dev/urandom
+(see
+.IR "drivers/char/random.c" ")."
+.TP
+.B SA_PROBE
+This bit indicates that the IRQ is being probed and that the handler
+being installed is not a real one.  It was intended that this value
+be used internally by
+.B probe_irq_on()
+(q.v.), but it is no longer used in 2.1.x kernels.  In fact, even with
+2.0.x kernels, it is only used by the MIPS architecture.  You should not
+be using this value unless you know what you are doing.
+.TP
+.B SA_STATIC_ALLOC
+(Sparc/Sparc64 only)  This bit requests that your
+.B struct irqaction
+(see below) be added to a statically allocated array of four handlers,
+rather than the normal
+.B irq_action[]
+table.  This is used for IRQs that must be requested early in the boot
+process, before
+.B kmalloc_init()
+has been called.
+.RE
+.PP
+The
+.I devname
+parameter is a short name for the device and is displayed in the
+.I "/proc/interrupts"
+list.
+.PP
+The
+.I dev_id
+parameter is the device ID.  This parameter is usually set to NULL, but should
+be non-null if you wish to do IRQ sharing.  This doesn't matter when hooking
+the interrupt, but is required so that, when
+.B free_irq()
+is called, the correct driver is unhooked.  Since this is a
+.BR "void *" ","
+it can point to anything (such as a device-specific structure, or even empty
+space), but make sure you pass the same pointer to
+.BR free_irq() "."
+.PP
+The
+.B free_irq()
+function releases an interrupt handler from operation.  It takes as parameters
+the IRQ to unregister, and the device ID of the handler to be unregistered.
+You should pass the same values here as you did to
+.BR request_irq() ". "
+You probably shouldn't unregister other people's interrupt handlers unless you
+know what you are doing.
+.SS Operation
+.PP
+For most architectures,
+.B request_irq()
+operates by allocating memory for a
+.BR "struct irqaction" ", "
+filling out the fields thereof, and adding it to the
+.B irq_action[]
+table.
+.B enable_irq()
+is then called, which simply tells the kernel to start delivering interrupts
+to the installed handler.  This process is vastly different on m68k machines,
+where it varies depending on what type of machine (Amiga, Atari, etc.) one is
+using.
+.B free_irq()
+simply removes the entries that
+.B request_irq()
+added. Note that some of these names differ depending on the architecture (for
+example,
+.B "struct irqaction"
+is known as
+.B "struct irq_action"
+on the Power PC).  If you need to know more about the internal workings of
+these functions, you are best off reading the source, as some of this
+information may have changed by the time you read this (if so, tell me, so
+I can try to update this page).
+.SS
+Fast Interrupt Handlers
+.PP
+A `fast' interrupt handler (one with
+.B SA_INTERRUPT
+set) has the following
+differences from normal `slow' interrupt handlers:
+.PP
+.RS
+On the ix86 and MIPS, the handler is called with interrupts disabled
+(they are enabled by default on these machines; on other machines,
+they are disabled by default).
+.PP
+On the MIPS, a faster return is used.
+.PP
+On the Alpha, MIPS, Sparc, and Sparc64, a fast and a slow handler
+may not share the same IRQ.
+.PP
+On all architectures except the m68k and the ix86, a `+' is displayed
+between the interrupt count and the device name in
+.IR /proc/interrupts ". "
+.RE
+.PP
+The slow-versus-fast interrupt distinction is slowly being phased out.  For
+example, under 2.0.x on the ix86,
+.B SA_INTERRUPT
+enabled a fast return as it still does on the MIPS; this distiction was
+removed in 2.1.x.
+.SH "RETURN VALUE"
+On success,
+.B request_irq()
+returns 0 if everything goes as planned.  Your interrupt handler will start
+receiving its interrupts immediately.  On failure,
+.B request_irq()
+returns:
+.RS
+.TP
+.B -EINVAL
+The IRQ number you requested was either invalid or reserved, or your passed a NULL
+pointer for the
+.I handler()
+parameter.
+.TP
+.B -ENOMEM
+.B request_irq()
+could not allocate enough memory for something (probably the
+.BR "struct irqaction" ")."
+.TP
+.B -EBUSY
+The IRQ you requested is already being handled, and the IRQ cannot be
+shared.  This can occur because either the handler being registered or
+the handler already present does not have
+.B SA_SHIRQ
+in its
+.I irqflags
+field.
+In addition, on most architectures, all handlers sharing a single IRQ must be
+of the same speed; i.e., either all or none of them may have the
+.B SA_INTERRUPT
+flag set.  Finally, it is possible that your architecture may not support sharing
+of the IRQ you are attempting to use.
+.TP
+.B -ENXIO
+The m68k returns this value for an invalid IRQ number.
+.RE
+.PP
+.B free_irq()
+does not return a value.
+.SH AVAILABILITY
+Linux 2.1+.  The information on this page should work for 2.0.x, but there
+may be subtle differences (for example, the semantics of
+.B SA_INTERRUPT
+on Intel-based machines).  Versions earlier than 2.0 had these functions,
+but the
+.I dev_id
+parameter was missing.  If you want your code to work with versions both earlier
+and later than 2.0, you should protect your code with preprocessor macros using
+.BR LINUX_VERSION_CODE ". "
+.SH "SEE ALSO"
+.BR init_bh "(9), " probe_irq_on "(9), "
+.IR "arch/*/kernel/irq.c" ", " "arch/*/kernel/entry.S" ", "
+.IR "include/linux/interrupt.h" ", " "include/asm*/signal.h" "."
+.SH AUTHOR
+.RI "Neil Moore <" amethyst@maxwell.ml.org ">"
+.SH BUGS
+It's not exactly a bug, but
+.B request_irq()
+on the m68k is
+.I very
+strange compared to the same function on the other supported architectures.
+You should really read
+.IR "arch/m68k/kernel/ints.c" ", " "arch/m68k/atari/ataints.c" ", "
+.IR "arch/m68k/amiga/amiints.c" ", and " "arch/m68k/amiga/cia.c"
+if you plan on writing drivers for any of these systems.
+.\" request_irq.9,v
+.\" Revision 1.3  1997/08/14 07:53:47  amethyst
+.\" Formatting changes.
+.\"
+.\" Revision 1.2  1997/07/27 09:52:24  amethyst
+.\" Changed some wording.
+.\" Fixed a few typographical warts.
+.\" Added email address to copyright notice.
+.\" Changed one occurence of ``i386'' to ``ix86''.
+.\" Changed ``Parameters'' subsection to ``Usage''.
+.\"
+.\" Revision 1.1  1997/07/27 09:28:25  amethyst
+.\" Initial revision
+.\"
diff --git a/upstream/mageia-cauldron/man9/save_flags.9 b/upstream/mageia-cauldron/man9/save_flags.9
new file mode 100644
index 00000000..da290dd7
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/save_flags.9
@@ -0,0 +1,63 @@
+.\" -*- nroff -*-
+.TH save_flags 9 "$Date:$" "Linux DDI" "Linux Kernel Functions"
+.SH NAME
+save_flags, restore_flags \- save/restore processor state and irq mask
+.SH SYNOPSIS
+.B #include <asm/system.h>
+.sp
+.BI "unsigned long " flags ";"
+.br
+.BI "void save_flags(" flags ")"
+.br
+.BI "void restore_flags(" flags ")"
+.SH DESCRIPTION
+The
+.B save_flags
+and
+.B restore_flags
+macros cooperate with the
+.B cli
+function to provide interrupt protection to critical sections of
+code. The
+.B save_flags
+function saves the current processor state, most specifically the
+interrupt priority level, in the
+.I flags
+value (which must be an l-value). The flags must be passed to a
+subsequent call to
+.B restore_flags
+to restore the processor state.
+.PP
+The
+.B save_flags
+macro does not affect the processor state, but drivers may use the
+.BR cli (9)
+and
+.B sti
+functions to disable and enable interrupts after saving the current
+flags. When the critical section of code is passed, the
+.B restore_flags
+returns the processor to its state at the point where the matching
+.B save_flags
+was called.
+.PP
+Only values returned by
+.B save_flags
+can be passed to
+.B restore_flags.
+Any other values may cause unpredictable results, and are
+certainly not portable.
+.SH "RETURN VALUE"
+The
+.B restore_flags
+macro saves the processor state in the
+.I flags
+parameter, which must be a non-const l-value.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.BR "cli" (9)
+.SH AUTHOR
+Stephen Williams (steve@icarus.com)
+.SH BUGS
+How does all this react with symmetric multiprocessor machines?
diff --git a/upstream/mageia-cauldron/man9/skb_dequeue.9 b/upstream/mageia-cauldron/man9/skb_dequeue.9
new file mode 100644
index 00000000..64a852be
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/skb_dequeue.9
@@ -0,0 +1,65 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Cyrus Durgin <cider@speakeasy.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH SKB_DEQUEUE 9 "November 24, 1997" "Linux DDI" "Networking Functions"
+.hlm 0
+.na
+.SH NAME
+__skb_dequeue, skb_dequeue \- remove an sk_buff from the head of a list
+.SH SYNOPSIS
+.B #include <linux/skbuff.h>
+.HP
+.BI "struct sk_buff *__skb_dequeue(struct sk_buff_head " *list ");"
+.TP
+.BI "struct sk_buff *skb_dequeue(struct sk_buff_head " *list ");"
+.hlm 1
+.ad
+.SH DESCRIPTION
+The \fBskb_dequeue\fP function removes the head element of an sk_buff_head.
+It decrements the list qlen pointer, and cleanly detaches the head from
+the queue.  This function should be used instead of performing this task
+manually, as it provides a clean, standardized way of manipulating an
+sk_buff_head, and provides interrupt disabling (see \fBNOTES\fP below.)
+.SH "RETURN VALUE"
+Returns a pointer to the head element of \fIlist\fP, or \fBNULL\fP if
+\fIlist\fP is empty.
+.SH NOTES
+It is important to note the difference between not only \fB__skb_dequeue\fP
+and \fBskb_dequeue\fP, but all the \fB__skb_\fP functions and their
+\fBskb_\fP counterparts.  Essentially, the \fB__skb_\fP functions are
+non-atomic, and should only be used with interrupts disabled.  As a convenience,
+the \fBskb_\fP functions are provided, which perform interrupt disable /
+enable wrapper functionality in addition to performing their specific tasks.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR intro "(9), " skb_unlink "(9), " skb_insert "(9), " skb_queue_head "(9), " skb_queue_tail "(9)"
+.ad
+.hlm 1
+.PP
+/usr/src/linux/net/netlink.c /usr/src/linux/net/ax25/af_ax25.c /usr/src/linux/net/core/datagram.c /usr/src/linux/net/ipv4/ipmr.c /usr/src/linux/net/ipv4/tcp.c
+.SH AUTHOR
+Cyrus Durgin <cider@speakeasy.org>
diff --git a/upstream/mageia-cauldron/man9/skb_insert.9 b/upstream/mageia-cauldron/man9/skb_insert.9
new file mode 100644
index 00000000..b4af6ecb
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/skb_insert.9
@@ -0,0 +1,81 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Cyrus Durgin <cider@speakeasy.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH SKB_INSERT 9 "November 24, 1997" "Linux DDI" "Networking Functions"
+.hlm 0
+.na
+.SH NAME
+__skb_insert, skb_insert, skb_append \- insert an sk_buff into a list
+.SH SYNOPSIS
+.B #include <linux/skbuff.h>
+.HP
+.BI "void __skb_insert(struct sk_buff " *newsk ", struct sk_buff " *prev ", struct sk_buff " *next ", struct sk_buff_head " *list ")"
+.TP
+.BI "void skb_insert(struct sk_buff " *old ", struct sk_buff " *newsk ")"
+.TP
+.BI "void skb_append(struct sk_buff " *old ", struct sk_buff " *newsk ")"
+.hlm 1
+.ad
+.SH DESCRIPTION
+\fBskb_insert\fP and \fBskb_append\fP are essentially wrapper functions for
+\fB__skb_insert\fP (see \fBNOTES\fP, below.)  \fB__skb_insert\fP inserts
+\fInewsk\fP into \fIlist\fP, and resets the appropriate next and prev
+pointers.  \fIprev\fP and \fInext\fP are used to frame \fInewsk\fP in
+\fIlist\fP.  After setting the next and prev pointers in \fInewsk\fP,
+\fB__skb_insert\fP sets the prev pointer in \fInext\fP and the next pointer
+in \fIprev\fP, sets the list pointer in \fInewsk\fP, and increments the qlen
+counter in \fIlist\fP.
+
+\fBskb_insert\fP and \fBskb_append\fP should be used to add sk_buffs
+to a list rather than performing this task manually; in addition to
+performing this task in a standardized way, these functions also provide
+for interrupt diasabling and prevent list mangling.  Both of these
+functions use the list pointer in \fIold\fP to determine to which list
+\fInewsk\fP should be attached.
+
+The \fBskb_insert\fP function adds \fInewsk\fP to the list before \fIold\fP.
+
+The \fBskb_append\fP function adds \fInewsk\fP to the list after \fIold\fP.
+.SH "RETURN VALUE"
+None.
+.SH NOTES
+It is important to note the difference between not only \fBskb_insert\fP,
+\fBskb_append\fP and \fB__skb_insert\fP, but all the \fB__skb_\fP functions
+and their \fBskb_\fP counterparts.  Essentially, the \fB__skb_\fP functions
+are non-atomic, and should only be used with interrupts disabled.  As a
+convenience, the \fBskb_\fP functions are provided, which perform interrupt
+disable / enable wrapper functionality in addition to performing their specific
+tasks.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR intro "(9), " skb_queue_head "(9), " skb_queue_tail "(9)"
+.ad
+.hlm 1
+.PP
+/usr/src/linux/net/ax25/af_ax25.c /usr/src/linux/net/core/skbuff.c /usr/src/linux/net/ipv4/tcp_input.c /usr/src/linux/net/netrom/nr_in.c
+.SH AUTHOR
+Cyrus Durgin <cider@speakeasy.org>
diff --git a/upstream/mageia-cauldron/man9/skb_peek.9 b/upstream/mageia-cauldron/man9/skb_peek.9
new file mode 100644
index 00000000..d1b7bd66
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/skb_peek.9
@@ -0,0 +1,58 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Cyrus Durgin <cider@speakeasy.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH SKB_PEEK 9 "November 24, 1997" "Linux DDI" "Networking Functions"
+.hlm 0
+.na
+.SH NAME
+skb_peek \- peek an sk_buff
+.SH SYNOPSIS
+.B #include <linux/skbuff.h>
+.HP
+.BI "struct sk_buff *skb_peek(struct sk_buff_head " *list_ ");"
+.hlm 1
+.ad
+.SH DESCRIPTION
+The \fBskb_peek\fP function extracts the head element of \fIlist_\fP,
+without modifying \fIlist_\fP.  It is important to note that this is
+not necessarily a safe operation, as \fIlist_\fP maintains the original
+sk_buff, and other operations on \fIlist_\fP may effect it.  To be safe,
+either disable interrupts using \fBcli\fP(9), call \fBskb_peek\fP, copy
+the data necessary to continue, and re-enable interrupts using
+\fBsti\fP(9), or use \fBskb_dequeue\fP(9).
+.SH "RETURN VALUE"
+Returns a pointer to sk_buff if there is a next element on the \fIlist_\fP.
+Otherwise, returns NULL.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR intro "(9), " skb_unlink "(9), " skb_dequeue "(9)"
+.ad
+.hlm 1
+.PP
+/usr/src/linux/net/appletalk/ddp.c /usr/src/linux/net/core/datagram.c /usr/src/linux/net/ipv4/tcp.c /usr/src/linux/net/ipv4/udp.c /usr/src/linux/net/unix/af_unix.c
+.SH AUTHOR
+Cyrus Durgin <cider@speakeasy.org>
diff --git a/upstream/mageia-cauldron/man9/skb_queue_empty.9 b/upstream/mageia-cauldron/man9/skb_queue_empty.9
new file mode 100644
index 00000000..c6154f83
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/skb_queue_empty.9
@@ -0,0 +1,54 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Cyrus Durgin <cider@speakeasy.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH SKB_QUEUE_EMPTY 9 "November 24, 1997" "Linux DDI" "Networking Functions"
+.hlm 0
+.na
+.SH NAME
+skb_queue_empty \- detect an empty skbuff queue
+.SH SYNOPSIS
+.B #include <linux/skbuff.h>
+.HP
+.BI "int skb_queue_empty(struct sk_buff_head " *list ");"
+.hlm 1
+.ad
+.SH DESCRIPTION
+The \fBskb_queue_empty\fP function checks an skbuff queue for "emptiness".
+This function provides a quick and easy way to determine if there are
+any sk_buff elements on a given queue.
+.SH "RETURN VALUE"
+If there are any sk_buff elements on the list, then 0 is returned.
+Otherwise, 1 is returned to indicate that the list is empty.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR intro "(9), " skb_queue_len "(9)"
+.ad
+.hlm 1
+.PP
+/usr/src/linux/net/core/datagram.c /usr/src/linux/net/core/dev.c /usr/src/linux/net/core/sock.c /usr/src/linux/net/ipv4/tcp.c 
+.SH AUTHOR
+Cyrus Durgin <cider@speakeasy.org>
diff --git a/upstream/mageia-cauldron/man9/skb_queue_head.9 b/upstream/mageia-cauldron/man9/skb_queue_head.9
new file mode 100644
index 00000000..c94a9f51
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/skb_queue_head.9
@@ -0,0 +1,64 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Cyrus Durgin <cider@speakeasy.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH SKB_QUEUE_HEAD 9 "November 24, 1997" "Linux DDI" "Networking Functions"
+.hlm 0
+.na
+.SH NAME
+skb_queue_head \- insert an sk_buff at the head of a list
+.SH SYNOPSIS
+.B #include <linux/skbuff.h>
+.HP
+.BI "void  __skb_queue_head(struct sk_buff_head " *list ", struct sk_buff " *newsk ");"
+.TP
+.BI "void  skb_queue_head(struct sk_buff_head " *list ", struct sk_buff " *newsk ");"
+.HP
+.hlm 1
+.ad
+.SH DESCRIPTION
+The \fBskb_queue_head\fP function adds \fInewsk\fP to the head of \fIlist\fP.
+Specifically, it sets the list, next, and prev pointers in \fInewsk\fP, and
+increments the qlen element of \fIlist\fP.  It also rearranges the next
+and prev pointers in the existing head of \fIlist\fP before returning.
+.SH "RETURN VALUE"
+None.
+.SH NOTES
+It is important to note the difference between not only \fB__skb_queue_head\fP
+and \fBskb_queue_head\fP, but all the \fB__skb_\fP functions and their
+\fBskb_\fP counterparts.  Essentially, the \fB__skb_\fP functions are
+non-atomic, and should only be used with interrupts disabled.  As a convenience,
+the \fBskb_\fP functions are provided, which perform interrupt disable /
+enable wrapper functionality in addition to performing their specific tasks.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR intro "(9), " skb_queue_tail "(9), " skb_insert "(9)"
+.ad
+.hlm 1
+.PP
+/usr/src/linux/net/ax25/af_ax25.c /usr/src/linux/net/core/skbuff.c /usr/src/linux/net/ipv4/tcp_input.c /usr/src/linux/net/netrom/nr_in.c
+.SH AUTHOR
+Cyrus Durgin <cider@speakeasy.org>
diff --git a/upstream/mageia-cauldron/man9/skb_queue_head_init.9 b/upstream/mageia-cauldron/man9/skb_queue_head_init.9
new file mode 100644
index 00000000..c5de1074
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/skb_queue_head_init.9
@@ -0,0 +1,58 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Cyrus Durgin <cider@speakeasy.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH SKB_QUEUE_HEAD_INIT 9 "November 24, 1997" "Linux DDI" "Networking Functions"
+.hlm 0
+.na
+.SH NAME
+skb_queue_head_init \- prepare an sk_buff_head for use
+.SH SYNOPSIS
+.B #include <linux/skbuff.h>
+.HP
+.BI "void  skb_queue_head_init(struct sk_buff_head " *list ");"
+.hlm 1
+.ad
+.SH DESCRIPTION
+The \fBskb_queue_head_init\fP function prepares an sk_buff_head for
+use by the other \fBskb_\fP functions, and should be called after
+declaring and allocating memory for any sk_buff_head elements.
+Specifically, \fBskb_queue_head_init\fP sets the prev and next
+pointers of the \fIlist\fP to point back at the \fIlist\fP itself,
+and sets the \fIlist\fP length to 0.  The use of sk_buff_head elements
+which have not been initialized with this function is undocumented and
+may produce irregular results.
+.SH "RETURN VALUE"
+None.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR intro "(9)"
+.ad
+.hlm 1
+.PP
+/usr/src/linux/net/netlink.c /usr/src/linux/net/appletalk/aarp.c /usr/src/linux/net/ipv4/af_inet.c /usr/src/linux/net/ipx/af_ipx.c /usr/src/linux/net/unix/af_unix.c
+.SH AUTHOR
+Cyrus Durgin <cider@speakeasy.org>
diff --git a/upstream/mageia-cauldron/man9/skb_queue_len.9 b/upstream/mageia-cauldron/man9/skb_queue_len.9
new file mode 100644
index 00000000..be668fa1
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/skb_queue_len.9
@@ -0,0 +1,53 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Cyrus Durgin <cider@speakeasy.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH SKB_QUEUE_LEN 9 "November 24, 1997" "Linux DDI" "Networking Functions"
+.hlm 0
+.na
+.SH NAME
+skb_queue_len \- determine the length of an sk_buff_head
+.SH SYNOPSIS
+.B #include <linux/skbuff.h>
+.HP
+.BI "__u32 skb_queue_len(struct sk_buff_head " *list ");"
+.hlm 1
+.ad
+.SH DESCRIPTION
+The \fBskb_queue_len\fP function determines the number of sk_buffs assigned to
+an sk_buff_head.
+.SH "RETURN VALUE"
+\fBskb_queue_len\fP returns the number of sk_buff elements attached to an
+sk_buff_head.  If the sk_buff_head is empty, then 0 is returned.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR intro "(9), " skb_queue_head_init "(9), " skb_queue_head "(9), " skb_queue_tail "(9), " skb_insert "(9)"
+.ad
+.hlm 1
+.PP
+/usr/src/linux/net/core/datagram.c
+.SH AUTHOR
+Cyrus Durgin <cider@speakeasy.org>
diff --git a/upstream/mageia-cauldron/man9/skb_queue_tail.9 b/upstream/mageia-cauldron/man9/skb_queue_tail.9
new file mode 100644
index 00000000..2a2003eb
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/skb_queue_tail.9
@@ -0,0 +1,63 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Cyrus Durgin <cider@speakeasy.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH SKB_QUEUE_TAIL 9 "November 24, 1997" "Linux DDI" "Networking Functions"
+.hlm 0
+.na
+.SH NAME
+__skb_queue_tail, skb_queue_tail \- insert an sk_buff at the tail of a list
+.SH SYNOPSIS
+.B #include <linux/skbuff.h>
+.HP
+.BI "void  __skb_queue_tail(struct sk_buff_head " *list ", struct sk_buff " *newsk ");"
+.TP
+.BI "void  skb_queue_tail(struct sk_buff_head " *list ", struct sk_buff " *newsk ");"
+.hlm 1
+.ad
+.SH DESCRIPTION
+The \fBskb_queue_tail\fP function adds \fInewsk\fP to the tail
+of \fIlist\fP.  Specifically, it sets the list, next, and prev
+pointers in \fInewsk\fP, and increments the qlen element of \fIlist\fP.
+.SH "RETURN VALUE"
+None.
+.SH NOTES
+It is important to note the difference between not only \fB__skb_queue_tail\fP
+and \fBskb_queue_tail\fP, but all the \fB__skb_\fP functions and their
+\fBskb_\fP counterparts.  Essentially, the \fB__skb_\fP functions are
+non-atomic, and should only be used with interrupts disabled.  As a convenience,
+the \fBskb_\fP functions are provided, which perform interrupt
+disable / enable wrapper functionality in addition to performing their specific
+tasks.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR intro "(9), " skb_queue_head "(9), " skb_insert "(9)"
+.ad
+.hlm 1
+.PP
+/usr/src/linux/net/netlink.c /usr/src/linux/net/appletalk/aarp.c /usr/src/linux/net/core/skbuff.c /usr/src/linux/net/ipv4/arp.c /usr/src/linux/include/net/sock.h
+.SH AUTHOR
+Cyrus Durgin <cider@speakeasy.org>
diff --git a/upstream/mageia-cauldron/man9/skb_unlink.9 b/upstream/mageia-cauldron/man9/skb_unlink.9
new file mode 100644
index 00000000..fe11af5b
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/skb_unlink.9
@@ -0,0 +1,69 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Cyrus Durgin <cider@speakeasy.org>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH SKB_UNLINK 9 "November 24, 1997" "Linux DDI" "Networking Functions"
+.hlm 0
+.na
+.SH NAME
+__skb_unlink, skb_unlink \- remove an sk_buff from its list
+.SH SYNOPSIS
+.B #include <linux/skbuff.h>
+.HP
+.BI "void __skb_unlink(struct sk_buff " *skb ", struct sk_buff_head " *list ");"
+.TP
+.BI "void skb_unlink(struct sk_buff " *skb ");"
+.hlm 1
+.ad
+.SH DESCRIPTION
+The \fBskb_unlink\fP function is a wrapper for \fB__skb_unlink\fP.
+\fB__skb_unlink\fP removes \fIskb\fP from its sk_buff_head.
+It decrements the list qlen pointer, and cleanly detaches the sk_buff
+from its queue.  This function should always be used instead of
+performing this task manually, as it provides a clean, standardized
+way of manipulating an sk_buff_head, and provides interrupt disabling (see
+\fBNOTES\fP below.)  Most users will not call \fB__skb_unlink\fP
+directly, as it requires that two arguments be supplied and does not
+provide any interrupt handling.  \fBskb_unlink\fP determines the list
+from which \fIskb\fP should be detached, and disables interrupts.
+.SH "RETURN VALUE"
+None.
+.SH NOTES
+It is important to note the difference between not only \fB__skb_unlink\fP
+and \fBskb_unlink\fP, but all the \fB__skb_\fP functions and their
+\fBskb_\fP counterparts.  Essentially, the \fB__skb_\fP functions are
+non-atomic, and should only be used with interrupts disabled.  As a convenience,
+the \fBskb_\fP functions are provided, which perform interrupt disable /
+enable wrapper functionality in addition to performing their specific tasks.
+.SH AVAILABILITY
+Linux 1.0+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR intro "(9), " skb_dequeue "(9), " skb_insert "(9), " skb_queue_head "(9), " skb_queue_tail "(9)"
+.ad
+.hlm 1
+.PP
+/usr/src/linux/net/core/skbuff.c /usr/src/linux/net/ipv4/af_inet.c /usr/src/linux/net/ipv4/ip_output.c /usr/src/linux/net/ipv4/tcp.c
+.SH AUTHOR
+Cyrus Durgin <cider@speakeasy.org>
diff --git a/upstream/mageia-cauldron/man9/skel.9 b/upstream/mageia-cauldron/man9/skel.9
new file mode 100644
index 00000000..5a1d477c
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/skel.9
@@ -0,0 +1,65 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH skel 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.\"
+.\" Turn off hyphenation and right justification for the first sections.
+.hlm 0
+.na
+.\" The name should be formatted this way for apropos to work
+.SH NAME
+skel \- skeleton man page for section 9 entries
+.\"
+.\"
+.SH SYNOPSIS
+.B #include <linux/linux.h>
+.HP
+.\"
+.\" Turn hyphenation and adjustment back on.
+.hlm 1
+.ad
+.SH DESCRIPTION
+Describe the function(s) and its parameters. This section should not
+be considered an introduction to kernel programming, just an english
+text description of the function at hand. It is OK to presume some
+basic knowledge of driver programming.
+.SH "RETURN VALUE"
+Describe the return values. Enumerate all the distinct values and all
+the ranges.
+.SH AVAILABILITY
+List kernel versions, and if restricted to certain architectures, say so.
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR man "(1), " man "(7), " intro "(9)"
+.ad
+.hlm 1
+.PP
+Also list some source files for the
+kernel that implement the functions of the page.
+.SH AUTHOR
+Who are you?
+.SH BUGS
+Describe any misfeatures or suprises that the use of these functions
+may lead to. The may not be errors, just unfortunate side effects.
diff --git a/upstream/mageia-cauldron/man9/sleep_on.9 b/upstream/mageia-cauldron/man9/sleep_on.9
new file mode 100644
index 00000000..ec19865c
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/sleep_on.9
@@ -0,0 +1,100 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH sleep_on 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.\"
+.\" Turn off hyphenation and right justification for the first sections.
+.hlm 0
+.na
+.\" The name should be formatted this way for apropos to work
+.SH NAME
+sleep_on \- synchronization using a condition variable
+.\"
+.\"
+.SH SYNOPSIS
+.B #include <linux/sched.h>
+.HP
+.BI "void sleep_on(struct\ wait_queue**" condition ")"
+.\"
+.\" Turn hyphenation and adjustment back on.
+.hlm 1
+.ad
+.SH DESCRIPTION
+The
+.B sleep_on
+function provides a means for synchronizing between processes and
+with interrupt handlers. The
+.I condition
+parameter is a pointer to a pointer to the opaque wait_queue
+type. Initialize the condition variable to zero, then pass a pointer
+to it to the
+.B sleep_on
+function. The basic idea is as follows:
+.PP
+.na
+struct wait_queue*con = 0;
+.br
+	[...]
+.br
+sleep_on(&con);
+.ad
+.PP
+While a process is sleeping, it is fully interruptible, no matter what
+the cpu state when entering the function. The cpu state is restored on
+being awakened.
+.PP
+If a null(0) is passed to
+.BR sleep_on ,
+it returns immediately, without sleeping. This is a no-op.
+.SH "RETURN VALUE"
+The
+.B sleep_on
+function only returns when explicitly awakened.
+.SH AVAILABILITY
+Linux 1+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR wake_up "(9)"
+.ad
+.hlm 1
+.PP
+.I "/usr/src/linux/kernel/sched.c"
+.SH AUTHOR
+Stephen Williams <steve@icarus.com>
+.SH BUGS
+A call to
+.I "sleep_on(0)"
+seems like an interesting way to test for and momentarily allow
+interrupts, but that is not what happens.
+.PP
+The
+.B sleep_on
+function cannot be called by interrupt handlers.
+.PP
+The function is not atomic with the condition tests that the driver
+writer might include, so the code executing the
+.B sleep_on
+function is protected from interrupts. Failure to do so generally
+leads to race conditions.
\ No newline at end of file
diff --git a/upstream/mageia-cauldron/man9/wake_up.9 b/upstream/mageia-cauldron/man9/wake_up.9
new file mode 100644
index 00000000..6ca4c2ef
--- /dev/null
+++ b/upstream/mageia-cauldron/man9/wake_up.9
@@ -0,0 +1,81 @@
+.\" -*- nroff -*-
+.\"
+.\" copyright (C) 1997 Stephen Williams <steve@icarus.com>
+.\"
+.\" Permission is granted to make and distribute verbatim copies of this
+.\" manual provided the copyright notice and this permission notice are
+.\" preserved on all copies.
+.\"
+.\" Permission is granted to copy and distribute modified versions of this
+.\" manual under the conditions for verbatim copying, provided that the
+.\" entire resulting derived work is distributed under the terms of a
+.\" permission notice identical to this one
+.\" 
+.\" Since the Linux kernel and libraries are constantly changing, this
+.\" manual page may be incorrect or out-of-date.  The author(s) assume no
+.\" responsibility for errors or omissions, or for damages resulting from
+.\" the use of the information contained herein.  The author(s) may not
+.\" have taken the same level of care in the production of this manual,
+.\" which is licensed free of charge, as they might when working
+.\" professionally.
+.\" 
+.\" Formatted or processed versions of this manual, if unaccompanied by
+.\" the source, must acknowledge the copyright and authors of this work.
+.\"
+.TH wake_up 9 "$Date:$" "Linux DDI" "Kernel Functions"
+.\"
+.\" Turn off hyphenation and right justification for the first sections.
+.hlm 0
+.na
+.\" The name should be formatted this way for apropos to work
+.SH NAME
+wake_up \- wake up sleeping processes
+.\"
+.\"
+.SH SYNOPSIS
+.B #include <linux/sched.h>
+.HP
+.BI "void wake_up(struct\ wait_queue**" condition ")"
+.\"
+.\" Turn hyphenation and adjustment back on.
+.hlm 1
+.ad
+.SH DESCRIPTION
+The
+.B wake_up
+function is the opposite of the
+.BR sleep_on "(9)"
+function in that it awakens processes that have gone to sleep using
+the same condition variable. All the processes sleeping on the given
+condition are awakened. If there are no processes sleeping on the
+condition, then none are affected.
+.PP
+Unlike the
+.BR sleep_on "(9)"
+function,
+.B wake_up
+does not block and may be called by interrupt handlers. It is in fact
+the principle means of synchronizing with interrupt events.
+.PP
+If the
+.I condition
+parameter is NULL, or there are no processes sleeping on
+.I condition,
+the call to
+.B wake_up
+is a no-op.
+.SH "RETURN VALUE"
+None.
+.SH AVAILABILITY
+Linux 1+
+.SH "SEE ALSO"
+.hlm 0
+.na
+.BR sleep_on "(9)"
+.ad
+.hlm 1
+.PP
+.I "/usr/src/linux/kernel/sched.c"
+.SH AUTHOR
+Stephen Williams <steve@icarus.com>
+.SH BUGS