License cleanup: add SPDX GPL-2.0 license identifier to files with no license
Many source files in the tree are missing licensing information, which
makes it harder for compliance tools to determine the correct license.
By default all files without license information are under the default
license of the kernel, which is GPL version 2.
Update the files which contain no license information with the 'GPL-2.0'
SPDX license identifier. The SPDX identifier is a legally binding
shorthand, which can be used instead of the full boiler plate text.
This patch is based on work done by Thomas Gleixner and Kate Stewart and
Philippe Ombredanne.
How this work was done:
Patches were generated and checked against linux-4.14-rc6 for a subset of
the use cases:
- file had no licensing information it it.
- file was a */uapi/* one with no licensing information in it,
- file was a */uapi/* one with existing licensing information,
Further patches will be generated in subsequent months to fix up cases
where non-standard license headers were used, and references to license
had to be inferred by heuristics based on keywords.
The analysis to determine which SPDX License Identifier to be applied to
a file was done in a spreadsheet of side by side results from of the
output of two independent scanners (ScanCode & Windriver) producing SPDX
tag:value files created by Philippe Ombredanne. Philippe prepared the
base worksheet, and did an initial spot review of a few 1000 files.
The 4.13 kernel was the starting point of the analysis with 60,537 files
assessed. Kate Stewart did a file by file comparison of the scanner
results in the spreadsheet to determine which SPDX license identifier(s)
to be applied to the file. She confirmed any determination that was not
immediately clear with lawyers working with the Linux Foundation.
Criteria used to select files for SPDX license identifier tagging was:
- Files considered eligible had to be source code files.
- Make and config files were included as candidates if they contained >5
lines of source
- File already had some variant of a license header in it (even if <5
lines).
All documentation files were explicitly excluded.
The following heuristics were used to determine which SPDX license
identifiers to apply.
- when both scanners couldn't find any license traces, file was
considered to have no license information in it, and the top level
COPYING file license applied.
For non */uapi/* files that summary was:
SPDX license identifier # files
---------------------------------------------------|-------
GPL-2.0 11139
and resulted in the first patch in this series.
If that file was a */uapi/* path one, it was "GPL-2.0 WITH
Linux-syscall-note" otherwise it was "GPL-2.0". Results of that was:
SPDX license identifier # files
---------------------------------------------------|-------
GPL-2.0 WITH Linux-syscall-note 930
and resulted in the second patch in this series.
- if a file had some form of licensing information in it, and was one
of the */uapi/* ones, it was denoted with the Linux-syscall-note if
any GPL family license was found in the file or had no licensing in
it (per prior point). Results summary:
SPDX license identifier # files
---------------------------------------------------|------
GPL-2.0 WITH Linux-syscall-note 270
GPL-2.0+ WITH Linux-syscall-note 169
((GPL-2.0 WITH Linux-syscall-note) OR BSD-2-Clause) 21
((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) 17
LGPL-2.1+ WITH Linux-syscall-note 15
GPL-1.0+ WITH Linux-syscall-note 14
((GPL-2.0+ WITH Linux-syscall-note) OR BSD-3-Clause) 5
LGPL-2.0+ WITH Linux-syscall-note 4
LGPL-2.1 WITH Linux-syscall-note 3
((GPL-2.0 WITH Linux-syscall-note) OR MIT) 3
((GPL-2.0 WITH Linux-syscall-note) AND MIT) 1
and that resulted in the third patch in this series.
- when the two scanners agreed on the detected license(s), that became
the concluded license(s).
- when there was disagreement between the two scanners (one detected a
license but the other didn't, or they both detected different
licenses) a manual inspection of the file occurred.
- In most cases a manual inspection of the information in the file
resulted in a clear resolution of the license that should apply (and
which scanner probably needed to revisit its heuristics).
- When it was not immediately clear, the license identifier was
confirmed with lawyers working with the Linux Foundation.
- If there was any question as to the appropriate license identifier,
the file was flagged for further research and to be revisited later
in time.
In total, over 70 hours of logged manual review was done on the
spreadsheet to determine the SPDX license identifiers to apply to the
source files by Kate, Philippe, Thomas and, in some cases, confirmation
by lawyers working with the Linux Foundation.
Kate also obtained a third independent scan of the 4.13 code base from
FOSSology, and compared selected files where the other two scanners
disagreed against that SPDX file, to see if there was new insights. The
Windriver scanner is based on an older version of FOSSology in part, so
they are related.
Thomas did random spot checks in about 500 files from the spreadsheets
for the uapi headers and agreed with SPDX license identifier in the
files he inspected. For the non-uapi files Thomas did random spot checks
in about 15000 files.
In initial set of patches against 4.14-rc6, 3 files were found to have
copy/paste license identifier errors, and have been fixed to reflect the
correct identifier.
Additionally Philippe spent 10 hours this week doing a detailed manual
inspection and review of the 12,461 patched files from the initial patch
version early this week with:
- a full scancode scan run, collecting the matched texts, detected
license ids and scores
- reviewing anything where there was a license detected (about 500+
files) to ensure that the applied SPDX license was correct
- reviewing anything where there was no detection but the patch license
was not GPL-2.0 WITH Linux-syscall-note to ensure that the applied
SPDX license was correct
This produced a worksheet with 20 files needing minor correction. This
worksheet was then exported into 3 different .csv files for the
different types of files to be modified.
These .csv files were then reviewed by Greg. Thomas wrote a script to
parse the csv files and add the proper SPDX tag to the file, in the
format that the file expected. This script was further refined by Greg
based on the output to detect more types of files automatically and to
distinguish between header and source .c files (which need different
comment types.) Finally Greg ran the script using the .csv files to
generate the patches.
Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org>
Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2017-11-01 14:07:57 +00:00
|
|
|
// SPDX-License-Identifier: GPL-2.0
|
2008-05-12 19:20:56 +00:00
|
|
|
/* Support for MMIO probes.
|
|
|
|
* Benfit many code from kprobes
|
|
|
|
* (C) 2002 Louis Zhuang <louis.zhuang@intel.com>.
|
|
|
|
* 2007 Alexander Eichner
|
|
|
|
* 2008 Pekka Paalanen <pq@iki.fi>
|
|
|
|
*/
|
|
|
|
|
2009-12-09 18:45:36 +00:00
|
|
|
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
|
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
#include <linux/list.h>
|
2008-05-19 11:35:24 +00:00
|
|
|
#include <linux/rculist.h>
|
2008-05-12 19:20:56 +00:00
|
|
|
#include <linux/spinlock.h>
|
|
|
|
#include <linux/hash.h>
|
2016-07-14 00:18:55 +00:00
|
|
|
#include <linux/export.h>
|
2008-05-12 19:20:56 +00:00
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/uaccess.h>
|
|
|
|
#include <linux/ptrace.h>
|
|
|
|
#include <linux/preempt.h>
|
2008-05-12 19:20:57 +00:00
|
|
|
#include <linux/percpu.h>
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
#include <linux/kdebug.h>
|
2008-05-12 19:20:57 +00:00
|
|
|
#include <linux/mutex.h>
|
2008-05-12 19:21:03 +00:00
|
|
|
#include <linux/io.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
|
|
|
#include <linux/slab.h>
|
2008-05-12 19:20:56 +00:00
|
|
|
#include <asm/cacheflush.h>
|
|
|
|
#include <asm/tlbflush.h>
|
2008-05-12 19:21:03 +00:00
|
|
|
#include <linux/errno.h>
|
2008-05-12 19:20:58 +00:00
|
|
|
#include <asm/debugreg.h>
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
#include <linux/mmiotrace.h>
|
2008-05-12 19:20:56 +00:00
|
|
|
|
|
|
|
#define KMMIO_PAGE_HASH_BITS 4
|
|
|
|
#define KMMIO_PAGE_TABLE_SIZE (1 << KMMIO_PAGE_HASH_BITS)
|
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
struct kmmio_fault_page {
|
|
|
|
struct list_head list;
|
|
|
|
struct kmmio_fault_page *release_next;
|
2016-03-03 01:03:11 +00:00
|
|
|
unsigned long addr; /* the requested address */
|
2009-04-28 19:17:50 +00:00
|
|
|
pteval_t old_presence; /* page presence prior to arming */
|
2009-03-01 14:11:58 +00:00
|
|
|
bool armed;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Number of times this page has been registered as a part
|
|
|
|
* of a probe. If zero, page is disarmed and this may be freed.
|
2009-02-24 19:44:15 +00:00
|
|
|
* Used only by writers (RCU) and post_kmmio_handler().
|
|
|
|
* Protected by kmmio_lock, when linked into kmmio_page_table.
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
*/
|
|
|
|
int count;
|
2010-06-13 21:56:54 +00:00
|
|
|
|
|
|
|
bool scheduled_for_release;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
struct kmmio_delayed_release {
|
|
|
|
struct rcu_head rcu;
|
|
|
|
struct kmmio_fault_page *release_list;
|
|
|
|
};
|
|
|
|
|
2008-05-12 19:20:56 +00:00
|
|
|
struct kmmio_context {
|
|
|
|
struct kmmio_fault_page *fpage;
|
|
|
|
struct kmmio_probe *probe;
|
|
|
|
unsigned long saved_flags;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
unsigned long addr;
|
2008-05-12 19:20:56 +00:00
|
|
|
int active;
|
|
|
|
};
|
|
|
|
|
|
|
|
static DEFINE_SPINLOCK(kmmio_lock);
|
|
|
|
|
2008-05-12 19:20:58 +00:00
|
|
|
/* Protected by kmmio_lock */
|
2008-05-12 19:20:56 +00:00
|
|
|
unsigned int kmmio_count;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
|
|
|
|
/* Read-protected by RCU, write-protected by kmmio_lock. */
|
2008-05-12 19:20:56 +00:00
|
|
|
static struct list_head kmmio_page_table[KMMIO_PAGE_TABLE_SIZE];
|
|
|
|
static LIST_HEAD(kmmio_probes);
|
|
|
|
|
2016-03-03 01:03:11 +00:00
|
|
|
static struct list_head *kmmio_page_list(unsigned long addr)
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
{
|
2016-03-03 01:03:11 +00:00
|
|
|
unsigned int l;
|
|
|
|
pte_t *pte = lookup_address(addr, &l);
|
|
|
|
|
|
|
|
if (!pte)
|
|
|
|
return NULL;
|
|
|
|
addr &= page_level_mask(l);
|
|
|
|
|
|
|
|
return &kmmio_page_table[hash_long(addr, KMMIO_PAGE_HASH_BITS)];
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
}
|
|
|
|
|
2008-05-12 19:20:57 +00:00
|
|
|
/* Accessed per-cpu */
|
|
|
|
static DEFINE_PER_CPU(struct kmmio_context, kmmio_ctx);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* this is basically a dynamic stabbing problem:
|
|
|
|
* Could use the existing prio tree code or
|
|
|
|
* Possible better implementations:
|
|
|
|
* The Interval Skip List: A Data Structure for Finding All Intervals That
|
|
|
|
* Overlap a Point (might be simple)
|
|
|
|
* Space Efficient Dynamic Stabbing with Fast Queries - Mikkel Thorup
|
|
|
|
*/
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
/* Get the kmmio at this addr (if any). You must be holding RCU read lock. */
|
2008-05-12 19:20:56 +00:00
|
|
|
static struct kmmio_probe *get_kmmio_probe(unsigned long addr)
|
|
|
|
{
|
|
|
|
struct kmmio_probe *p;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
list_for_each_entry_rcu(p, &kmmio_probes, list) {
|
2009-04-28 19:17:48 +00:00
|
|
|
if (addr >= p->addr && addr < (p->addr + p->len))
|
2008-05-12 19:20:56 +00:00
|
|
|
return p;
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
/* You must be holding RCU read lock. */
|
2016-03-03 01:03:11 +00:00
|
|
|
static struct kmmio_fault_page *get_kmmio_fault_page(unsigned long addr)
|
2008-05-12 19:20:56 +00:00
|
|
|
{
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
struct list_head *head;
|
2009-04-28 19:17:49 +00:00
|
|
|
struct kmmio_fault_page *f;
|
2016-03-03 01:03:11 +00:00
|
|
|
unsigned int l;
|
|
|
|
pte_t *pte = lookup_address(addr, &l);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2016-03-03 01:03:11 +00:00
|
|
|
if (!pte)
|
|
|
|
return NULL;
|
|
|
|
addr &= page_level_mask(l);
|
|
|
|
head = kmmio_page_list(addr);
|
2009-04-28 19:17:49 +00:00
|
|
|
list_for_each_entry_rcu(f, head, list) {
|
2016-03-03 01:03:11 +00:00
|
|
|
if (f->addr == addr)
|
2009-04-28 19:17:49 +00:00
|
|
|
return f;
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2009-04-28 19:17:50 +00:00
|
|
|
static void clear_pmd_presence(pmd_t *pmd, bool clear, pmdval_t *old)
|
2009-03-01 14:12:48 +00:00
|
|
|
{
|
|
|
|
pmdval_t v = pmd_val(*pmd);
|
2009-04-28 19:17:50 +00:00
|
|
|
if (clear) {
|
|
|
|
*old = v & _PAGE_PRESENT;
|
|
|
|
v &= ~_PAGE_PRESENT;
|
|
|
|
} else /* presume this has been called with clear==true previously */
|
|
|
|
v |= *old;
|
2009-03-01 14:12:48 +00:00
|
|
|
set_pmd(pmd, __pmd(v));
|
|
|
|
}
|
|
|
|
|
2009-04-28 19:17:50 +00:00
|
|
|
static void clear_pte_presence(pte_t *pte, bool clear, pteval_t *old)
|
2009-03-01 14:12:48 +00:00
|
|
|
{
|
|
|
|
pteval_t v = pte_val(*pte);
|
2009-04-28 19:17:50 +00:00
|
|
|
if (clear) {
|
|
|
|
*old = v & _PAGE_PRESENT;
|
|
|
|
v &= ~_PAGE_PRESENT;
|
|
|
|
} else /* presume this has been called with clear==true previously */
|
|
|
|
v |= *old;
|
2009-03-01 14:12:48 +00:00
|
|
|
set_pte_atomic(pte, __pte(v));
|
|
|
|
}
|
|
|
|
|
2009-04-28 19:17:50 +00:00
|
|
|
static int clear_page_presence(struct kmmio_fault_page *f, bool clear)
|
2008-05-12 19:20:56 +00:00
|
|
|
{
|
2008-05-12 19:21:14 +00:00
|
|
|
unsigned int level;
|
2016-03-03 01:03:11 +00:00
|
|
|
pte_t *pte = lookup_address(f->addr, &level);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2008-05-12 19:20:56 +00:00
|
|
|
if (!pte) {
|
2016-03-03 01:03:11 +00:00
|
|
|
pr_err("no pte for addr 0x%08lx\n", f->addr);
|
2009-01-30 17:38:59 +00:00
|
|
|
return -1;
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
|
|
|
|
2008-05-12 19:20:58 +00:00
|
|
|
switch (level) {
|
|
|
|
case PG_LEVEL_2M:
|
2009-04-28 19:17:50 +00:00
|
|
|
clear_pmd_presence((pmd_t *)pte, clear, &f->old_presence);
|
2008-05-12 19:20:58 +00:00
|
|
|
break;
|
|
|
|
case PG_LEVEL_4K:
|
2009-04-28 19:17:50 +00:00
|
|
|
clear_pte_presence(pte, clear, &f->old_presence);
|
2008-05-12 19:20:58 +00:00
|
|
|
break;
|
|
|
|
default:
|
2009-12-09 18:45:36 +00:00
|
|
|
pr_err("unexpected page level 0x%x.\n", level);
|
2009-01-30 17:38:59 +00:00
|
|
|
return -1;
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
|
|
|
|
2016-03-03 01:03:11 +00:00
|
|
|
__flush_tlb_one(f->addr);
|
2009-01-30 17:38:59 +00:00
|
|
|
return 0;
|
2008-05-12 19:20:58 +00:00
|
|
|
}
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2009-03-01 14:11:58 +00:00
|
|
|
/*
|
|
|
|
* Mark the given page as not present. Access to it will trigger a fault.
|
|
|
|
*
|
|
|
|
* Struct kmmio_fault_page is protected by RCU and kmmio_lock, but the
|
|
|
|
* protection is ignored here. RCU read lock is assumed held, so the struct
|
|
|
|
* will not disappear unexpectedly. Furthermore, the caller must guarantee,
|
|
|
|
* that double arming the same virtual address (page) cannot occur.
|
|
|
|
*
|
|
|
|
* Double disarming on the other hand is allowed, and may occur when a fault
|
|
|
|
* and mmiotrace shutdown happen simultaneously.
|
|
|
|
*/
|
|
|
|
static int arm_kmmio_fault_page(struct kmmio_fault_page *f)
|
2008-05-12 19:20:58 +00:00
|
|
|
{
|
2009-03-01 14:11:58 +00:00
|
|
|
int ret;
|
2009-12-09 18:45:36 +00:00
|
|
|
WARN_ONCE(f->armed, KERN_ERR pr_fmt("kmmio page already armed.\n"));
|
2009-03-01 14:11:58 +00:00
|
|
|
if (f->armed) {
|
2016-03-03 01:03:11 +00:00
|
|
|
pr_warning("double-arm: addr 0x%08lx, ref %d, old %d\n",
|
|
|
|
f->addr, f->count, !!f->old_presence);
|
2009-03-01 14:11:58 +00:00
|
|
|
}
|
2009-04-28 19:17:50 +00:00
|
|
|
ret = clear_page_presence(f, true);
|
2016-03-03 01:03:11 +00:00
|
|
|
WARN_ONCE(ret < 0, KERN_ERR pr_fmt("arming at 0x%08lx failed.\n"),
|
|
|
|
f->addr);
|
2009-03-01 14:11:58 +00:00
|
|
|
f->armed = true;
|
2009-01-30 17:38:59 +00:00
|
|
|
return ret;
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
|
|
|
|
2009-03-01 14:11:58 +00:00
|
|
|
/** Restore the given page to saved presence state. */
|
|
|
|
static void disarm_kmmio_fault_page(struct kmmio_fault_page *f)
|
2008-05-12 19:20:56 +00:00
|
|
|
{
|
2009-04-28 19:17:50 +00:00
|
|
|
int ret = clear_page_presence(f, false);
|
2009-03-01 14:11:58 +00:00
|
|
|
WARN_ONCE(ret < 0,
|
2016-03-03 01:03:11 +00:00
|
|
|
KERN_ERR "kmmio disarming at 0x%08lx failed.\n", f->addr);
|
2009-03-01 14:11:58 +00:00
|
|
|
f->armed = false;
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
/*
|
|
|
|
* This is being called from do_page_fault().
|
|
|
|
*
|
|
|
|
* We may be in an interrupt or a critical section. Also prefecthing may
|
|
|
|
* trigger a page fault. We may be in the middle of process switch.
|
|
|
|
* We cannot take any locks, because we could be executing especially
|
|
|
|
* within a kmmio critical section.
|
|
|
|
*
|
|
|
|
* Local interrupts are disabled, so preemption cannot happen.
|
|
|
|
* Do not enable interrupts, do not sleep, and watch out for other CPUs.
|
|
|
|
*/
|
2008-05-12 19:20:56 +00:00
|
|
|
/*
|
|
|
|
* Interrupts are disabled on entry as trap3 is an interrupt gate
|
tree-wide: fix assorted typos all over the place
That is "success", "unknown", "through", "performance", "[re|un]mapping"
, "access", "default", "reasonable", "[con]currently", "temperature"
, "channel", "[un]used", "application", "example","hierarchy", "therefore"
, "[over|under]flow", "contiguous", "threshold", "enough" and others.
Signed-off-by: André Goddard Rosa <andre.goddard@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2009-11-14 15:09:05 +00:00
|
|
|
* and they remain disabled throughout this function.
|
2008-05-12 19:20:56 +00:00
|
|
|
*/
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
int kmmio_handler(struct pt_regs *regs, unsigned long addr)
|
2008-05-12 19:20:56 +00:00
|
|
|
{
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
struct kmmio_context *ctx;
|
|
|
|
struct kmmio_fault_page *faultpage;
|
2008-05-12 19:20:58 +00:00
|
|
|
int ret = 0; /* default to fault not handled */
|
2016-03-03 01:03:11 +00:00
|
|
|
unsigned long page_base = addr;
|
|
|
|
unsigned int l;
|
|
|
|
pte_t *pte = lookup_address(addr, &l);
|
|
|
|
if (!pte)
|
|
|
|
return -EINVAL;
|
|
|
|
page_base &= page_level_mask(l);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Preemption is now disabled to prevent process switch during
|
|
|
|
* single stepping. We can only handle one active kmmio trace
|
|
|
|
* per cpu, so ensure that we finish it before something else
|
2008-05-12 19:20:57 +00:00
|
|
|
* gets to run. We also hold the RCU read lock over single
|
|
|
|
* stepping to avoid looking up the probe and kmmio_fault_page
|
|
|
|
* again.
|
2008-05-12 19:20:56 +00:00
|
|
|
*/
|
|
|
|
preempt_disable();
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
rcu_read_lock();
|
2008-05-12 19:20:57 +00:00
|
|
|
|
2016-03-03 01:03:11 +00:00
|
|
|
faultpage = get_kmmio_fault_page(page_base);
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
if (!faultpage) {
|
|
|
|
/*
|
|
|
|
* Either this page fault is not caused by kmmio, or
|
|
|
|
* another CPU just pulled the kmmio probe from under
|
2008-05-12 19:20:58 +00:00
|
|
|
* our feet. The latter case should not be possible.
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
*/
|
|
|
|
goto no_kmmio;
|
|
|
|
}
|
|
|
|
|
|
|
|
ctx = &get_cpu_var(kmmio_ctx);
|
2008-05-12 19:20:56 +00:00
|
|
|
if (ctx->active) {
|
2016-03-03 01:03:11 +00:00
|
|
|
if (page_base == ctx->addr) {
|
2008-05-12 19:20:58 +00:00
|
|
|
/*
|
2009-02-05 11:02:02 +00:00
|
|
|
* A second fault on the same page means some other
|
|
|
|
* condition needs handling by do_page_fault(), the
|
|
|
|
* page really not being present is the most common.
|
2008-05-12 19:20:58 +00:00
|
|
|
*/
|
2009-12-09 18:45:36 +00:00
|
|
|
pr_debug("secondary hit for 0x%08lx CPU %d.\n",
|
|
|
|
addr, smp_processor_id());
|
2009-02-05 11:02:02 +00:00
|
|
|
|
|
|
|
if (!faultpage->old_presence)
|
2009-12-09 18:45:36 +00:00
|
|
|
pr_info("unexpected secondary hit for address 0x%08lx on CPU %d.\n",
|
|
|
|
addr, smp_processor_id());
|
2009-02-05 11:02:02 +00:00
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* Prevent overwriting already in-flight context.
|
|
|
|
* This should not happen, let's hope disarming at
|
|
|
|
* least prevents a panic.
|
|
|
|
*/
|
2009-12-09 18:45:36 +00:00
|
|
|
pr_emerg("recursive probe hit on CPU %d, for address 0x%08lx. Ignoring.\n",
|
|
|
|
smp_processor_id(), addr);
|
|
|
|
pr_emerg("previous hit was at 0x%08lx.\n", ctx->addr);
|
2009-02-05 11:02:02 +00:00
|
|
|
disarm_kmmio_fault_page(faultpage);
|
|
|
|
}
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
goto no_kmmio_ctx;
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
|
|
|
ctx->active++;
|
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
ctx->fpage = faultpage;
|
2016-03-03 01:03:11 +00:00
|
|
|
ctx->probe = get_kmmio_probe(page_base);
|
2008-05-12 19:20:58 +00:00
|
|
|
ctx->saved_flags = (regs->flags & (X86_EFLAGS_TF | X86_EFLAGS_IF));
|
2016-03-03 01:03:11 +00:00
|
|
|
ctx->addr = page_base;
|
2008-05-12 19:20:56 +00:00
|
|
|
|
|
|
|
if (ctx->probe && ctx->probe->pre_handler)
|
|
|
|
ctx->probe->pre_handler(ctx->probe, regs, addr);
|
|
|
|
|
2008-05-12 19:20:57 +00:00
|
|
|
/*
|
|
|
|
* Enable single-stepping and disable interrupts for the faulting
|
|
|
|
* context. Local interrupts must not get enabled during stepping.
|
|
|
|
*/
|
2008-05-12 19:20:58 +00:00
|
|
|
regs->flags |= X86_EFLAGS_TF;
|
|
|
|
regs->flags &= ~X86_EFLAGS_IF;
|
2008-05-12 19:20:56 +00:00
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
/* Now we set present bit in PTE and single step. */
|
2009-03-01 14:11:58 +00:00
|
|
|
disarm_kmmio_fault_page(ctx->fpage);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2008-05-12 19:20:57 +00:00
|
|
|
/*
|
|
|
|
* If another cpu accesses the same page while we are stepping,
|
|
|
|
* the access will not be caught. It will simply succeed and the
|
|
|
|
* only downside is we lose the event. If this becomes a problem,
|
|
|
|
* the user should drop to single cpu before tracing.
|
|
|
|
*/
|
|
|
|
|
2008-05-12 19:20:57 +00:00
|
|
|
put_cpu_var(kmmio_ctx);
|
2008-05-12 19:20:58 +00:00
|
|
|
return 1; /* fault handled */
|
2008-05-12 19:20:56 +00:00
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
no_kmmio_ctx:
|
|
|
|
put_cpu_var(kmmio_ctx);
|
2008-05-12 19:20:56 +00:00
|
|
|
no_kmmio:
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
rcu_read_unlock();
|
2008-05-12 19:20:56 +00:00
|
|
|
preempt_enable_no_resched();
|
2008-05-12 19:20:58 +00:00
|
|
|
return ret;
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Interrupts are disabled on entry as trap1 is an interrupt gate
|
tree-wide: fix assorted typos all over the place
That is "success", "unknown", "through", "performance", "[re|un]mapping"
, "access", "default", "reasonable", "[con]currently", "temperature"
, "channel", "[un]used", "application", "example","hierarchy", "therefore"
, "[over|under]flow", "contiguous", "threshold", "enough" and others.
Signed-off-by: André Goddard Rosa <andre.goddard@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2009-11-14 15:09:05 +00:00
|
|
|
* and they remain disabled throughout this function.
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
* This must always get called as the pair to kmmio_handler().
|
2008-05-12 19:20:56 +00:00
|
|
|
*/
|
|
|
|
static int post_kmmio_handler(unsigned long condition, struct pt_regs *regs)
|
|
|
|
{
|
2008-05-12 19:20:57 +00:00
|
|
|
int ret = 0;
|
|
|
|
struct kmmio_context *ctx = &get_cpu_var(kmmio_ctx);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2008-05-12 19:20:58 +00:00
|
|
|
if (!ctx->active) {
|
2009-04-28 19:17:51 +00:00
|
|
|
/*
|
|
|
|
* debug traps without an active context are due to either
|
|
|
|
* something external causing them (f.e. using a debugger while
|
|
|
|
* mmio tracing enabled), or erroneous behaviour
|
|
|
|
*/
|
2009-12-09 18:45:36 +00:00
|
|
|
pr_warning("unexpected debug trap on CPU %d.\n",
|
|
|
|
smp_processor_id());
|
2008-05-12 19:20:57 +00:00
|
|
|
goto out;
|
2008-05-12 19:20:58 +00:00
|
|
|
}
|
2008-05-12 19:20:56 +00:00
|
|
|
|
|
|
|
if (ctx->probe && ctx->probe->post_handler)
|
|
|
|
ctx->probe->post_handler(ctx->probe, condition, regs);
|
|
|
|
|
2009-02-24 19:44:15 +00:00
|
|
|
/* Prevent racing against release_kmmio_fault_page(). */
|
|
|
|
spin_lock(&kmmio_lock);
|
|
|
|
if (ctx->fpage->count)
|
|
|
|
arm_kmmio_fault_page(ctx->fpage);
|
|
|
|
spin_unlock(&kmmio_lock);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2008-05-12 19:20:58 +00:00
|
|
|
regs->flags &= ~X86_EFLAGS_TF;
|
2008-05-12 19:20:56 +00:00
|
|
|
regs->flags |= ctx->saved_flags;
|
|
|
|
|
|
|
|
/* These were acquired in kmmio_handler(). */
|
|
|
|
ctx->active--;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
BUG_ON(ctx->active);
|
2008-05-12 19:20:57 +00:00
|
|
|
rcu_read_unlock();
|
2008-05-12 19:20:56 +00:00
|
|
|
preempt_enable_no_resched();
|
|
|
|
|
|
|
|
/*
|
|
|
|
* if somebody else is singlestepping across a probe point, flags
|
|
|
|
* will have TF set, in which case, continue the remaining processing
|
|
|
|
* of do_debug, as if this is not a probe hit.
|
|
|
|
*/
|
2008-05-12 19:20:58 +00:00
|
|
|
if (!(regs->flags & X86_EFLAGS_TF))
|
2008-05-12 19:20:57 +00:00
|
|
|
ret = 1;
|
|
|
|
out:
|
|
|
|
put_cpu_var(kmmio_ctx);
|
|
|
|
return ret;
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
/* You must be holding kmmio_lock. */
|
2016-03-03 01:03:11 +00:00
|
|
|
static int add_kmmio_fault_page(unsigned long addr)
|
2008-05-12 19:20:56 +00:00
|
|
|
{
|
|
|
|
struct kmmio_fault_page *f;
|
|
|
|
|
2016-03-03 01:03:11 +00:00
|
|
|
f = get_kmmio_fault_page(addr);
|
2008-05-12 19:20:56 +00:00
|
|
|
if (f) {
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
if (!f->count)
|
2009-03-01 14:11:58 +00:00
|
|
|
arm_kmmio_fault_page(f);
|
2008-05-12 19:20:56 +00:00
|
|
|
f->count++;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-03-01 14:11:58 +00:00
|
|
|
f = kzalloc(sizeof(*f), GFP_ATOMIC);
|
2008-05-12 19:20:56 +00:00
|
|
|
if (!f)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
f->count = 1;
|
2016-03-03 01:03:11 +00:00
|
|
|
f->addr = addr;
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2009-03-01 14:11:58 +00:00
|
|
|
if (arm_kmmio_fault_page(f)) {
|
2009-01-30 17:38:59 +00:00
|
|
|
kfree(f);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2016-03-03 01:03:11 +00:00
|
|
|
list_add_rcu(&f->list, kmmio_page_list(f->addr));
|
2008-05-12 19:20:56 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
/* You must be holding kmmio_lock. */
|
2016-03-03 01:03:11 +00:00
|
|
|
static void release_kmmio_fault_page(unsigned long addr,
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
struct kmmio_fault_page **release_list)
|
2008-05-12 19:20:56 +00:00
|
|
|
{
|
|
|
|
struct kmmio_fault_page *f;
|
|
|
|
|
2016-03-03 01:03:11 +00:00
|
|
|
f = get_kmmio_fault_page(addr);
|
2008-05-12 19:20:56 +00:00
|
|
|
if (!f)
|
|
|
|
return;
|
|
|
|
|
|
|
|
f->count--;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
BUG_ON(f->count < 0);
|
2008-05-12 19:20:56 +00:00
|
|
|
if (!f->count) {
|
2009-03-01 14:11:58 +00:00
|
|
|
disarm_kmmio_fault_page(f);
|
2010-06-13 21:56:54 +00:00
|
|
|
if (!f->scheduled_for_release) {
|
|
|
|
f->release_next = *release_list;
|
|
|
|
*release_list = f;
|
|
|
|
f->scheduled_for_release = true;
|
|
|
|
}
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-05-12 19:21:03 +00:00
|
|
|
/*
|
|
|
|
* With page-unaligned ioremaps, one or two armed pages may contain
|
|
|
|
* addresses from outside the intended mapping. Events for these addresses
|
|
|
|
* are currently silently dropped. The events may result only from programming
|
|
|
|
* mistakes by accessing addresses before the beginning or past the end of a
|
|
|
|
* mapping.
|
|
|
|
*/
|
2008-05-12 19:20:56 +00:00
|
|
|
int register_kmmio_probe(struct kmmio_probe *p)
|
|
|
|
{
|
2008-05-12 19:20:57 +00:00
|
|
|
unsigned long flags;
|
2008-05-12 19:20:56 +00:00
|
|
|
int ret = 0;
|
|
|
|
unsigned long size = 0;
|
2017-11-27 07:51:39 +00:00
|
|
|
unsigned long addr = p->addr & PAGE_MASK;
|
2008-05-12 19:21:03 +00:00
|
|
|
const unsigned long size_lim = p->len + (p->addr & ~PAGE_MASK);
|
2016-03-03 01:03:11 +00:00
|
|
|
unsigned int l;
|
|
|
|
pte_t *pte;
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2008-05-12 19:20:57 +00:00
|
|
|
spin_lock_irqsave(&kmmio_lock, flags);
|
2017-11-27 07:51:39 +00:00
|
|
|
if (get_kmmio_probe(addr)) {
|
2008-05-12 19:20:56 +00:00
|
|
|
ret = -EEXIST;
|
|
|
|
goto out;
|
|
|
|
}
|
2016-03-03 01:03:11 +00:00
|
|
|
|
2017-11-27 07:51:39 +00:00
|
|
|
pte = lookup_address(addr, &l);
|
2016-03-03 01:03:11 +00:00
|
|
|
if (!pte) {
|
|
|
|
ret = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2008-05-12 19:20:57 +00:00
|
|
|
kmmio_count++;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
list_add_rcu(&p->list, &kmmio_probes);
|
2008-05-12 19:21:03 +00:00
|
|
|
while (size < size_lim) {
|
2017-11-27 07:51:39 +00:00
|
|
|
if (add_kmmio_fault_page(addr + size))
|
2009-12-09 18:45:36 +00:00
|
|
|
pr_err("Unable to set page fault.\n");
|
2016-03-03 01:03:11 +00:00
|
|
|
size += page_level_size(l);
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
|
|
|
out:
|
2008-05-12 19:20:57 +00:00
|
|
|
spin_unlock_irqrestore(&kmmio_lock, flags);
|
2008-05-12 19:20:56 +00:00
|
|
|
/*
|
|
|
|
* XXX: What should I do here?
|
|
|
|
* Here was a call to global_flush_tlb(), but it does not exist
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
* anymore. It seems it's not needed after all.
|
2008-05-12 19:20:56 +00:00
|
|
|
*/
|
|
|
|
return ret;
|
|
|
|
}
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
EXPORT_SYMBOL(register_kmmio_probe);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
static void rcu_free_kmmio_fault_pages(struct rcu_head *head)
|
|
|
|
{
|
|
|
|
struct kmmio_delayed_release *dr = container_of(
|
|
|
|
head,
|
|
|
|
struct kmmio_delayed_release,
|
|
|
|
rcu);
|
2009-04-28 19:17:49 +00:00
|
|
|
struct kmmio_fault_page *f = dr->release_list;
|
|
|
|
while (f) {
|
|
|
|
struct kmmio_fault_page *next = f->release_next;
|
|
|
|
BUG_ON(f->count);
|
|
|
|
kfree(f);
|
|
|
|
f = next;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
}
|
|
|
|
kfree(dr);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void remove_kmmio_fault_pages(struct rcu_head *head)
|
|
|
|
{
|
2009-03-08 18:21:35 +00:00
|
|
|
struct kmmio_delayed_release *dr =
|
|
|
|
container_of(head, struct kmmio_delayed_release, rcu);
|
2009-04-28 19:17:49 +00:00
|
|
|
struct kmmio_fault_page *f = dr->release_list;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
struct kmmio_fault_page **prevp = &dr->release_list;
|
|
|
|
unsigned long flags;
|
2009-03-08 18:21:35 +00:00
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
spin_lock_irqsave(&kmmio_lock, flags);
|
2009-04-28 19:17:49 +00:00
|
|
|
while (f) {
|
|
|
|
if (!f->count) {
|
|
|
|
list_del_rcu(&f->list);
|
|
|
|
prevp = &f->release_next;
|
2009-03-08 18:21:35 +00:00
|
|
|
} else {
|
2009-04-28 19:17:49 +00:00
|
|
|
*prevp = f->release_next;
|
2010-06-13 21:56:54 +00:00
|
|
|
f->release_next = NULL;
|
|
|
|
f->scheduled_for_release = false;
|
2009-03-08 18:21:35 +00:00
|
|
|
}
|
2010-06-13 21:56:54 +00:00
|
|
|
f = *prevp;
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
}
|
|
|
|
spin_unlock_irqrestore(&kmmio_lock, flags);
|
2009-03-08 18:21:35 +00:00
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
/* This is the real RCU destroy call. */
|
|
|
|
call_rcu(&dr->rcu, rcu_free_kmmio_fault_pages);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Remove a kmmio probe. You have to synchronize_rcu() before you can be
|
2008-05-12 19:20:57 +00:00
|
|
|
* sure that the callbacks will not be called anymore. Only after that
|
|
|
|
* you may actually release your struct kmmio_probe.
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
*
|
|
|
|
* Unregistering a kmmio fault page has three steps:
|
|
|
|
* 1. release_kmmio_fault_page()
|
|
|
|
* Disarm the page, wait a grace period to let all faults finish.
|
|
|
|
* 2. remove_kmmio_fault_pages()
|
|
|
|
* Remove the pages from kmmio_page_table.
|
|
|
|
* 3. rcu_free_kmmio_fault_pages()
|
2009-12-05 15:41:34 +00:00
|
|
|
* Actually free the kmmio_fault_page structs as with RCU.
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
*/
|
2008-05-12 19:20:56 +00:00
|
|
|
void unregister_kmmio_probe(struct kmmio_probe *p)
|
|
|
|
{
|
2008-05-12 19:20:57 +00:00
|
|
|
unsigned long flags;
|
2008-05-12 19:20:56 +00:00
|
|
|
unsigned long size = 0;
|
2017-11-27 07:51:39 +00:00
|
|
|
unsigned long addr = p->addr & PAGE_MASK;
|
2008-05-12 19:21:03 +00:00
|
|
|
const unsigned long size_lim = p->len + (p->addr & ~PAGE_MASK);
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
struct kmmio_fault_page *release_list = NULL;
|
|
|
|
struct kmmio_delayed_release *drelease;
|
2016-03-03 01:03:11 +00:00
|
|
|
unsigned int l;
|
|
|
|
pte_t *pte;
|
|
|
|
|
2017-11-27 07:51:39 +00:00
|
|
|
pte = lookup_address(addr, &l);
|
2016-03-03 01:03:11 +00:00
|
|
|
if (!pte)
|
|
|
|
return;
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2008-05-12 19:20:57 +00:00
|
|
|
spin_lock_irqsave(&kmmio_lock, flags);
|
2008-05-12 19:21:03 +00:00
|
|
|
while (size < size_lim) {
|
2017-11-27 07:51:39 +00:00
|
|
|
release_kmmio_fault_page(addr + size, &release_list);
|
2016-03-03 01:03:11 +00:00
|
|
|
size += page_level_size(l);
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
list_del_rcu(&p->list);
|
2008-05-12 19:20:56 +00:00
|
|
|
kmmio_count--;
|
2008-05-12 19:20:57 +00:00
|
|
|
spin_unlock_irqrestore(&kmmio_lock, flags);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2010-06-13 21:56:54 +00:00
|
|
|
if (!release_list)
|
|
|
|
return;
|
|
|
|
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
drelease = kmalloc(sizeof(*drelease), GFP_ATOMIC);
|
|
|
|
if (!drelease) {
|
2009-12-09 18:45:36 +00:00
|
|
|
pr_crit("leaking kmmio_fault_page objects.\n");
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
drelease->release_list = release_list;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is not really RCU here. We have just disarmed a set of
|
|
|
|
* pages so that they cannot trigger page faults anymore. However,
|
|
|
|
* we cannot remove the pages from kmmio_page_table,
|
|
|
|
* because a probe hit might be in flight on another CPU. The
|
|
|
|
* pages are collected into a list, and they will be removed from
|
|
|
|
* kmmio_page_table when it is certain that no probe hit related to
|
|
|
|
* these pages can be in flight. RCU grace period sounds like a
|
|
|
|
* good choice.
|
|
|
|
*
|
|
|
|
* If we removed the pages too early, kmmio page fault handler might
|
|
|
|
* not find the respective kmmio_fault_page and determine it's not
|
|
|
|
* a kmmio fault, when it actually is. This would lead to madness.
|
|
|
|
*/
|
|
|
|
call_rcu(&drelease->rcu, remove_kmmio_fault_pages);
|
2008-05-12 19:20:56 +00:00
|
|
|
}
|
x86: mmiotrace full patch, preview 1
kmmio.c handles the list of mmio probes with callbacks, list of traced
pages, and attaching into the page fault handler and die notifier. It
arms, traps and disarms the given pages, this is the core of mmiotrace.
mmio-mod.c is a user interface, hooking into ioremap functions and
registering the mmio probes. It also decodes the required information
from trapped mmio accesses via the pre and post callbacks in each probe.
Currently, hooking into ioremap functions works by redefining the symbols
of the target (binary) kernel module, so that it calls the traced
versions of the functions.
The most notable changes done since the last discussion are:
- kmmio.c is a built-in, not part of the module
- direct call from fault.c to kmmio.c, removing all dynamic hooks
- prepare for unregistering probes at any time
- make kmmio re-initializable and accessible to more than one user
- rewrite kmmio locking to remove all spinlocks from page fault path
Can I abuse call_rcu() like I do in kmmio.c:unregister_kmmio_probe()
or is there a better way?
The function called via call_rcu() itself calls call_rcu() again,
will this work or break? There I need a second grace period for RCU
after the first grace period for page faults.
Mmiotrace itself (mmio-mod.c) is still a module, I am going to attack
that next. At some point I will start looking into how to make mmiotrace
a tracer component of ftrace (thanks for the hint, Ingo). Ftrace should
make the user space part of mmiotracing as simple as
'cat /debug/trace/mmio > dump.txt'.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 19:20:57 +00:00
|
|
|
EXPORT_SYMBOL(unregister_kmmio_probe);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2009-04-28 19:17:51 +00:00
|
|
|
static int
|
|
|
|
kmmio_die_notifier(struct notifier_block *nb, unsigned long val, void *args)
|
2008-05-12 19:20:56 +00:00
|
|
|
{
|
|
|
|
struct die_args *arg = args;
|
2010-01-16 09:39:30 +00:00
|
|
|
unsigned long* dr6_p = (unsigned long *)ERR_PTR(arg->err);
|
2008-05-12 19:20:56 +00:00
|
|
|
|
2010-01-16 09:39:30 +00:00
|
|
|
if (val == DIE_DEBUG && (*dr6_p & DR_STEP))
|
|
|
|
if (post_kmmio_handler(*dr6_p, arg->regs) == 1) {
|
2009-06-01 18:17:06 +00:00
|
|
|
/*
|
|
|
|
* Reset the BS bit in dr6 (pointed by args->err) to
|
|
|
|
* denote completion of processing
|
|
|
|
*/
|
2010-01-16 09:39:30 +00:00
|
|
|
*dr6_p &= ~DR_STEP;
|
2008-05-12 19:20:56 +00:00
|
|
|
return NOTIFY_STOP;
|
2009-06-01 18:17:06 +00:00
|
|
|
}
|
2008-05-12 19:20:56 +00:00
|
|
|
|
|
|
|
return NOTIFY_DONE;
|
|
|
|
}
|
2008-05-12 19:20:58 +00:00
|
|
|
|
|
|
|
static struct notifier_block nb_die = {
|
|
|
|
.notifier_call = kmmio_die_notifier
|
|
|
|
};
|
|
|
|
|
2009-04-28 19:17:51 +00:00
|
|
|
int kmmio_init(void)
|
2008-05-12 19:20:58 +00:00
|
|
|
{
|
|
|
|
int i;
|
2009-04-28 19:17:51 +00:00
|
|
|
|
2008-05-12 19:20:58 +00:00
|
|
|
for (i = 0; i < KMMIO_PAGE_TABLE_SIZE; i++)
|
|
|
|
INIT_LIST_HEAD(&kmmio_page_table[i]);
|
2009-04-28 19:17:51 +00:00
|
|
|
|
2008-05-12 19:20:58 +00:00
|
|
|
return register_die_notifier(&nb_die);
|
|
|
|
}
|
2009-04-28 19:17:51 +00:00
|
|
|
|
|
|
|
void kmmio_cleanup(void)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
unregister_die_notifier(&nb_die);
|
|
|
|
for (i = 0; i < KMMIO_PAGE_TABLE_SIZE; i++) {
|
|
|
|
WARN_ONCE(!list_empty(&kmmio_page_table[i]),
|
|
|
|
KERN_ERR "kmmio_page_table not empty at cleanup, any further tracing will leak memory.\n");
|
|
|
|
}
|
|
|
|
}
|