This patch is based on Andi Kleen's work:
Implement autoprobing/loading of modules serving CPU
specific features (x86cpu autoloading).

And Kay Siever's work to get rid of sysdev cpu structures
and making use of struct device instead.

Before, the cpuid driver had to be loaded to get the x86cpu
autoloading feature. With this patch autoloading works through
the /sys/devices/system/cpu object

Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Dave Jones <davej@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Len Brown <lenb@kernel.org>
Acked-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NThomas Renninger <trenn@suse.de>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Don't try to describe the actual models for now.

v2: Fix typo: X86_VENDOR_ANY -> X86_FAMILY_ANY (trenn)
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NThomas Renninger <trenn@suse.de>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This marks all the x86 cpuinfo tables to the CPU specific device drivers,
to allow auto loading by udev. This should simplify the distribution
startup scripts for this greatly.

I didn't add MODULE_DEVICE_IDs to the centrino and p4-clockmod drivers,
because those probably shouldn't be auto loaded and the acpi driver
be used instead (not fully sure on that, would appreciate feedback)

The old nforce drivers autoload based on the PCI ID.

ACPI cpufreq is autoloaded in another patch.

v3: Autoload gx based on PCI IDs only. Remove cpu check (Dave Jones)
v4: Use newly introduce HW_PSTATE feature for powernow-k8 loading

Cc: Dave Jones <davej@redhat.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NThomas Renninger <trenn@suse.de>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

It is rather similar to CPB (boot capability) feature
and exists since fam10h (can be looked up in AMD's BKDG).

The feature is needed for powernow-k8 to cleanup init functions and to
provide proper autoloading matching with the new x86cpu modalias
feature.

Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Dave Jones <davej@redhat.com>
Cc: Borislav Petkov <bp@amd64.org>
Signed-off-by: NThomas Renninger <trenn@suse.de>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Use the new x86 cpuid autoprobe interface for the Intel coretemp
driver.

Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Jean Delvare <khali@linux-fr.org>
Cc: Guenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NThomas Renninger <trenn@suse.de>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Use the new x86 cpuid autoprobe interface.

Cc: Jean Delvare <khali@linux-fr.org>
Cc: Guenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NThomas Renninger <trenn@suse.de>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The only left over hole in automatic cpufreq driver loading was the loading
of ACPI cpufreq. This driver should be loaded when ACPI supports a _PDC
method and the CPU vendor wants to use acpi cpufreq.

Simply add a request module call to the acpi processor core driver
when this is true. This seems like the simplest solution for this.

Cc: Len Brown <lenb@kernel.org>
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NThomas Renninger <trenn@suse.de>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

With this it should be automatically loaded on suitable systems by
udev.

The old switch () is replaced with a table based approach, this
also cleans up the code.

Cc: Len Brown <lenb@kernel.org>
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NThomas Renninger <trenn@suse.de>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Add support for auto-loading of crypto drivers based on cpuid features.
This enables auto-loading of the VIA and Intel specific drivers
for AES, hashing and CRCs.

Requires the earlier infrastructure patch to add x86 modinfo.
I kept it all in a single patch for now.

I dropped the printks when the driver cpuid doesn't match (imho
drivers never should print anything in such a case)

One drawback is that udev doesn't know if the drivers are used or not,
so they will be unconditionally loaded at boot up. That's better
than not loading them at all, like it often happens.

Cc: Dave Jones <davej@redhat.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Jen Axboe <axboe@kernel.dk>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Huang Ying <ying.huang@intel.com>
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NThomas Renninger <trenn@suse.de>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

There's a growing number of drivers that support a specific x86 feature
or CPU.  Currently loading these drivers currently on a generic
distribution requires various driver specific hacks and it often
doesn't work.

This patch adds auto probing for drivers based on the x86 cpuid
information, in particular based on vendor/family/model number
and also based on CPUID feature bits.

For example a common issue is not loading the SSE 4.2 accelerated
CRC module: this can significantly lower the performance of BTRFS
which relies on fast CRC.

Another issue is loading the right CPUFREQ driver for the current CPU.
Currently distributions often try all all possible driver until
one sticks, which is not really a good way to do this.

It works with existing udev without any changes. The code
exports the x86 information as a generic string in sysfs
that can be matched by udev's pattern matching.

This scheme does not support numeric ranges, so if you want to
handle e.g. ranges of model numbers they have to be encoded
in ASCII or simply all models or families listed. Fixing
that would require changing udev.

Another issue is that udev will happily load all drivers that match,
there is currently no nice way to stop a specific driver from
being loaded if it's not needed (e.g. if you don't need fast CRC)
But there are not that many cpu specific drivers around and they're
all not that bloated, so this isn't a particularly serious issue.

Originally this patch added the modalias to the normal cpu
sysdevs. However sysdevs don't have all the infrastructure
needed for udev, so it couldn't really autoload drivers.
This patch instead adds the CPU modaliases to the cpuid devices,
which are real devices with full support for udev. This implies
that the cpuid driver has to be loaded to use this.

This patch just adds infrastructure, some driver conversions
in followups.

Thanks to Kay for helping with some sysfs magic.

v2: Constifcation, some updates
v4: (trenn@suse.de):
    - Use kzalloc instead of kmalloc to terminate modalias buffer
    - Use uppercase hex values to match correctly against hex values containing
      letters

Cc: Dave Jones <davej@redhat.com>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Jen Axboe <axboe@kernel.dk>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Len Brown <lenb@kernel.org>
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Signed-off-by: NThomas Renninger <trenn@suse.de>
Acked-by: NH. Peter Anvin <hpa@zytor.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Cautious admins may want to restrict access to debugfs. Currently a
manual chown/chmod e.g. in an init script is needed to achieve that.
Distributions that want to make the mount options configurable need
to add extra config files. By allowing to set the root inode's uid,
gid and mode via mount options no such hacks are needed anymore.
Instead configuration becomes straight forward via fstab.
Signed-off-by: NLudwig Nussel <ludwig.nussel@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Now that there are no users of get_driver() or put_driver(), this
patch (as1513) removes those routines completely.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

As part of the removal of get_driver()/put_driver(), this patch
(as1512) gets rid of various useless and unnecessary calls in several
drivers.  In some cases it may be desirable to pin the driver by
calling try_module_get(), but that can be done later.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: "David S. Miller" <davem@davemloft.net>
CC: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
CC: Michael Buesch <m@bues.ch>
CC: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Remove useless {get,put}_driver - the caller of the functions
has to ensure valid driver pointers.
Signed-off-by: NSebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

As part of the removal of get_driver()/put_driver(), this patch
(as1511) changes all the places that add dynamic IDs for drivers.
Since these additions are done by writing to the drivers' sysfs
attribute files, and the attributes are removed when the drivers are
unregistered, there is no reason to take an extra reference to the
drivers.

The one exception is the pci-stub driver, which calls pci_add_dynid()
as part of its registration.  But again, there's no reason to take an
extra reference here, because the driver can't be unloaded while it is
being registered.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: Dmitry Torokhov <dmitry.torokhov@gmail.com>
CC: Jiri Kosina <jkosina@suse.cz>
CC: Jesse Barnes <jbarnes@virtuousgeek.org>
CC: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

As part of the removal of get_driver()/put_driver(), this patch
(as1510) changes driver_find(); it now drops the reference it acquires
before returning.  The patch also adjusts all the callers of
driver_find() to remove the now unnecessary calls to put_driver().

In addition, the patch adds a warning to driver_find(): Callers must
make sure the driver they are searching for does not get unloaded
while they are using it.  This has always been the case; driver_find()
has never prevented a driver from being unregistered or unloaded.
Hence the patch will not introduce any new bugs.  The existing callers
all seem to be okay in this respect, however I don't understand the
video drivers well enough to be certain about them.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: Dmitry Torokhov <dmitry.torokhov@gmail.com>
CC: Kyungmin Park <kyungmin.park@samsung.com>
CC: Andy Walls <awalls@md.metrocast.net>
CC: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Check if the sif is not NULL before de-referencing it
Signed-off-by: NJonghwan Choi <jhbird.choi@samsung.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

There is a misleading difference between /proc and /sys permissions, /proc is 0555 and /sys is 0755. But
as it is impossible to create or unlink something in /sys it would be nice to have same permissions.
Signed-off-by: NVitaly Kuznetsov <vitty@altlinux.ru>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

The current code unnecessarily limits the number of offers we handle.
Get rid of this limitation. As part of this cleanup, also get rid of an
unused define - MAX_MSG_TYPES.
Signed-off-by: NK. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: NHaiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Insert ddebug_exec_queries() in place of ddebug_exec_query().  It
splits the query string on [;\n], and calls ddebug_exec_query() on
each.  All queries are processed independent of errors, allowing a
query to fail, for example when a module is not installed.  Empty
lines and comments are skipped.  Errors are counted, and the last
error seen (negative) or the number of callsites found (0 or positive)
is returned.  Return code checks are altered accordingly.

With this, multiple queries can be given in ddebug_query, allowing
more selective enabling of callsites.  As a side effect, a set of
commands can be batched in:

	cat cmd-file > $DBGMT/dynamic_debug/control

We dont want a ddebug_query syntax error to kill the dynamic debug
facility, so dynamic_debug_init() zeros ddebug_exec_queries()'s return
code after logging the appropriate message, so that ddebug tables are
preserved and $DBGMT/dynamic_debug/control file is created.  This
would be appropriate even without accepting multiple queries.

This patch also alters ddebug_change() to return number of callsites
matched (which typically is the same as number of callsites changed).
ddebug_exec_query() also returns the number found, or a negative value
if theres a parse error on the query.

Splitting on [;\n] prevents their use in format-specs, but selecting
callsites on punctuation is brittle anyway, meaningful and selective
substrings are more typical.

Note: splitting queries on ';' before handling trailing #comments
means that a ';' also terminates a comment, and text after the ';' is
treated as another query.  This trailing query will almost certainly
result in a parse error and thus have no effect other than the error
message.  The double corner case with unexpected results is:

     ddebug_query="func foo +p # enable foo ; +p"
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Factor pr_info(query) out of ddebug_parse_query, into vpr_info_dq(),
for reuse later.  Also change the printed labels: file, func to agree
with the query-spec keywords accepted in the control file.  Pass ""
when string is null, to avoid "(null)" output from sprintf.  For
format print, use precision to skip last char, assuming its '\n', no
great harm if not, its a debug msg.
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

trim_prefix(path) skips past the absolute source path root, and
returns the pointer to the relative path from there.  It is used to
shorten the displayed path in $DBGMT/dynamic_debug/control via
ddebug_proc_show(), and in ddebug_change() to allow relative filenames
to be used in applied queries.  For example:

  ~# echo file kernel/freezer.c +p > $DBGMT/dynamic_debug/control

  kernel/freezer.c:128 [freezer]cancel_freezing p "  clean up: %s\012"

trim_prefix(path) insures common prefix before trimming it, so
out-of-tree module paths are shown as full absolute paths.
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Current query write buffer is 256 bytes, on stack.  In comparison, the
ddebug_query boot-arg is 1024.  Allocate the buffer off heap, and
enlarge it to 4096 bytes, big enough for ~100 queries (at 40 bytes
each), and error out if not.  This makes it play nicely with large
query sets (to be added later).  The buffer should be enough for most
uses, and others should probably be split into subsets.

[jbaron@redhat.com: changed USER_BUF_PAGE from 4095 -> 4096 ]
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

If a token begins with #, the remainder of query string is a comment,
so drop it.  Doing it here avoids '#' in quoted strings.
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

lineno:24 allows files with 4 million lines, an insane file-size, even
for never-to-get-in-tree machine generated code.  Reduce this to 18
bits, which still allows 256k lines.  This is still insanely big, but
its not raving mad.
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

If _ddebug table is empty (in a CONFIG_DYNAMIC_DEBUG build this
shouldn't happen), then warn (error?) and return early.  This skips
empty table scan and parsing of setup-string, including the pr_info
call noting the parse.  By inspection, copy return-code handling from
1st ddebug_add_module() callsite to 2nd.
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Issue error when a match-spec is given multiple times in a rule.
Previous code kept last one, but was silent about it.  Docs imply only
one is allowed by saying match-specs are ANDed together, given that
module M cannot match both A and B.  Also error when last_line < 1st_line.
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Change describe_flags() to emit '=[pmflt_]+' for current callsite
flags, or just '=_' when they're disabled.  Having '=' in output
allows a more selective grep expression; in contrast '-' may appear
in filenames, line-ranges, and format-strings.  '=' also has better
mnemonics, saying; "the current setting is equal to <flags>".

This allows grep "=_" <dbgfs>/dynamic_debug/control to see disabled
callsites while avoiding the many occurrences of " = " seen in format
strings.

Enlarge flagsbufs to handle additional flag char, and alter
ddebug_parse_flags() to allow flags=0, so that user can turn off all
debug flags via:

  ~# echo =_ > <dbgfs>/dynamic_debug/control
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Convert 'if (x !=NULL)' checks into 'if (x)'.
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Issue keyword/parsing errors even w/o verbose set;
uncover otherwize mysterious non-functionality.
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Replace strcpy with strlcpy, and add define for the size constant.

[jbaron@redhat.com: Use DDEBUG_STRING_SIZE for overflow check]
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Allow changing dynamic_debug verbosity at run-time, to ease debugging
of ddebug queries as you add them, improving usability.

at boot time: dynamic_debug.verbose=1
at runtime:
root@voyage:~# echo 1 > /sys/module/dynamic_debug/parameters/verbose
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

If CONFIG_DYNAMIC_DEBUG is defined, honor it over DEBUG, so that
pr_debug()s are controllable, instead of always-on.  When DEBUG is
also defined, change _DPRINTK_FLAGS_DEFAULT to enable printing by
default.

Also adding _DPRINTK_FLAGS_INCL_MODNAME would be nice, but there are
numerous cases of pr_debug(NAME ": ...), which would result in double
printing of module-name.  So defer this until things settle.

Cc: David Miller <davem@davemloft.net>
Cc: Joe Perches <joe@perches.com>
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Currently any enabled dynamic-debug flag on a pr_debug callsite will
enable printing, even if _DPRINTK_FLAGS_PRINT is off.  Checking print
flag directly allows "-p" to disable callsites without fussing with
other flags, so the following disables everything, without altering
flags user may have set:

	echo -p > $DBGFS/dynamic_debug/control
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Style cleanups.
Signed-off-by: NJim Cromie <jim.cromie@gmail.com>
Signed-off-by: NJason Baron <jbaron@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Tracking the number of subdirectories requires an extra field that increases
the size of sysfs_dirent.  nlinks are not particularly interesting for sysfs
and the nlink counts are wrong when network namespaces are involved so stop
counting them, and always return nlink == 1.  Userspace already knows that
directories with nlink == 1 have an nlink count they can't use to count
subdirectories.

This reduces the size of sysfs_dirent by 8 bytes on 64bit platforms.
Signed-off-by: NEric W. Biederman <ebiederm@xmission.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Store the sysfs inode number in an unsided int because
ida inode allocator can return at most a 31 bit number,
reducing the size of struct sysfs_dirent by 8 bytes
on 64bit platforms.
Signed-off-by: NEric W. Biederman <ebiederm@xmission.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

On 32bit this reduces sizeof(struct sysfs_dirent) by 2 bytes.
Signed-off-by: NEric W. Biederman <ebiederm@xmission.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Compute a 31 bit hash of directory entries (that can fit in a signed
32bit off_t) and index the sysfs directory entries by that hash,
replacing the per directory indexes by name and by inode.  Because we
now only use a single rbtree this reduces the size of sysfs_dirent by 2
pointers.  Because we have fewer cases to deal with the code is now
simpler.

For now I use the simple hash that the dcache uses as that is easy to
use and seems simple enough.

In addition to makeing the code simpler using a hash for the file
position in readdir brings sysfs in line with other filesystems that
have non-trivial directory structures.
Signed-off-by: NEric W. Biederman <ebiederm@xmission.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>