Since commit 8b41fc44 ("kbuild: create modules.builtin without
Makefile.modbuiltin or tristate.conf"), MODULE_LICENSE declarations
are used to identify modules. As a consequence, uses of the macro
in non-modules will cause modprobe to misidentify their containing
object file as a module when it is not (false positives), and modprobe
might succeed rather than failing with a suitable error message.

So remove it in the files in this commit, none of which can be built as
modules.
Signed-off-by: Nick Alcock <nick.alcock@oracle.com>
Suggested-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: linux-modules@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: Hitomi Hasegawa <hasegawa-hitomi@fujitsu.com>
Cc: Talel Shenhar <talel@amazon.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Marc Zyngier <maz@kernel.org>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Since commit 8b41fc44 ("kbuild: create modules.builtin without
Makefile.modbuiltin or tristate.conf"), MODULE_LICENSE declarations
are used to identify modules. As a consequence, uses of the macro
in non-modules will cause modprobe to misidentify their containing
object file as a module when it is not (false positives), and modprobe
might succeed rather than failing with a suitable error message.

So remove it in the files in this commit, none of which can be built as
modules.
Signed-off-by: Nick Alcock <nick.alcock@oracle.com>
Suggested-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: linux-modules@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: Hitomi Hasegawa <hasegawa-hitomi@fujitsu.com>
Cc: Thor Thayer <thor.thayer@linux.intel.com>
Cc: Lee Jones <lee@kernel.org>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Since commit 8b41fc44 ("kbuild: create modules.builtin without
Makefile.modbuiltin or tristate.conf"), MODULE_LICENSE declarations
are used to identify modules. As a consequence, uses of the macro
in non-modules will cause modprobe to misidentify their containing
object file as a module when it is not (false positives), and modprobe
might succeed rather than failing with a suitable error message.

So remove it in the files in this commit, none of which can be built as
modules.
Signed-off-by: Nick Alcock <nick.alcock@oracle.com>
Suggested-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: linux-modules@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: Hitomi Hasegawa <hasegawa-hitomi@fujitsu.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Commit d6f819908f8aac ("module: fold usermode helper kmod into modules
directory") moves kmod helper implementation (kmod.c) to kernel/module/
directory but forgets to update its reference on kernel api doc, hence:

WARNING: kernel-doc './scripts/kernel-doc -rst -enable-lineno -sphinx-version 2.4.4 -export ./kernel/kmod.c' failed with return code 2

Update the reference.

Fixes: d6f819908f8aac ("module: fold usermode helper kmod into modules directory")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Link: https://lore.kernel.org/linux-next/20230324154413.19cc78be@canb.auug.org.au/Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
Reviewed-by: Liam Beguin <liambeguin@gmail.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

already_uses() is unnecessarily chatty.

`modprobe i915` yields 491 messages like:

  [   64.108744] i915 uses drm!

This is a normal situation, and isn't worth all the log entries.

NOTE: I've preserved the "does not use %s" messages, which happens
less often, but does happen.  Its not clear to me what it tells a
reader, or what info might improve the pr_debug's utility.

[ 6847.584999] main:already_uses:569: amdgpu does not use ttm!
[ 6847.585001] main:add_module_usage:584: Allocating new usage for amdgpu.
[ 6847.585014] main:already_uses:569: amdgpu does not use drm!
[ 6847.585016] main:add_module_usage:584: Allocating new usage for amdgpu.
[ 6847.585024] main:already_uses:569: amdgpu does not use drm_display_helper!
[ 6847.585025] main:add_module_usage:584: Allocating new usage for amdgpu.
[ 6847.585084] main:already_uses:569: amdgpu does not use drm_kms_helper!
[ 6847.585086] main:add_module_usage:584: Allocating new usage for amdgpu.
[ 6847.585175] main:already_uses:569: amdgpu does not use drm_buddy!
[ 6847.585176] main:add_module_usage:584: Allocating new usage for amdgpu.
[ 6847.585202] main:already_uses:569: amdgpu does not use i2c_algo_bit!
[ 6847.585204] main:add_module_usage:584: Allocating new usage for amdgpu.
[ 6847.585249] main:already_uses:569: amdgpu does not use gpu_sched!
[ 6847.585250] main:add_module_usage:584: Allocating new usage for amdgpu.
[ 6847.585314] main:already_uses:569: amdgpu does not use video!
[ 6847.585315] main:add_module_usage:584: Allocating new usage for amdgpu.
[ 6847.585409] main:already_uses:569: amdgpu does not use iommu_v2!
[ 6847.585410] main:add_module_usage:584: Allocating new usage for amdgpu.
[ 6847.585816] main:already_uses:569: amdgpu does not use drm_ttm_helper!
[ 6847.585818] main:add_module_usage:584: Allocating new usage for amdgpu.
[ 6848.762268] dyndbg: add-module: amdgpu.2533 sites

no functional changes.
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

move_module() pr_debug's "Final section addresses for $modname".
Add section addresses to the message, for anyone looking at these.

no functional changes.
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The pr_debug("Absolute symbol" ..) reports value, (which is usually
0), but not the name, which is more informative.  So add it.

no functional changes
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

layout_sections() and move_module() each issue ~50 messages for each
module loaded.  Add mod-name into their 2 header lines, to help the
reader find his module.

no functional changes.
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The kernel/kmod.c is already only built if we enabled modules, so
just stuff it under kernel/module/kmod.c and unify the MAINTAINERS
file for it.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The setup_load_info() was actually had ELF validation checks of its
own. To later cache useful variables as an secondary step just means
looping again over the ELF sections we just validated. We can simply
keep tabs of the key sections of interest as we validate the module
ELF section in one swoop, so do that and merge the two routines
together.

Expand a bit on the documentation / intent / goals.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The symbol and strings section validation currently happen in
setup_load_info() but since they are also doing validity checks
move this to elf_validity_check().
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The integrity of the struct module we load is important, and although
our ELF validator already checks that the module section must match
struct module, add a stop-gap check before we memcpy() the final minted
module. This also makes those inspecting the code what the goal is.

While at it, clarify the goal behind updating the sh_addr address.
The current comment is pretty misleading.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The ELF ".gnu.linkonce.this_module" section is special, it is what we
use to construct the struct module __this_module, which THIS_MODULE
points to. When userspace loads a module we always deal first with a
copy of the userspace buffer, and twiddle with the userspace copy's
version of the struct module. Eventually we allocate memory to do a
memcpy() of that struct module, under the assumption that the module
size is right. But we have no validity checks against the size or
the requirements for the section.

Add some validity checks for the special module section early and while
at it, cache the module section index early, so we don't have to do that
later.

While at it, just move over the assigment of the info->mod to make the
code clearer. The validity checker also adds an explicit size check to
ensure the module section size matches the kernel's run time size for
sizeof(struct module). This should prevent sloppy loads of modules
which are built today *without* actually increasing the size of
the struct module. A developer today can for example expand the size
of struct module, rebuild a directoroy 'make fs/xfs/' for example and
then try to insmode the driver there. That module would in effect have
an incorrect size. This new size check would put a stop gap against such
mistakes.

This also makes the entire goal of ".gnu.linkonce.this_module" pretty
clear. Before this patch verification of the goal / intent required some
Indian Jones whips, torches and cleaning up big old spider webs.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

This makes the routine easier to understand what the check its checking for.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Converge on a compromise: so long as we have a module hit our linked
list of modules we taint. That is, the module was about to become live.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Just move the signature taint into the helper:

  module_augment_kernel_taints()
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

It is silly to have taints spread out all over, we can just compromise
and add them if the module ever hit our linked list. Our sanity checkers
should just prevent crappy drivers / bogus ELF modules / etc and kconfig
options should be enough to let you *not* load things you don't want.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

check_modinfo() actually does two things:

 a) sanity checks, some of which are fatal, and so we
    prevent the user from completing trying to load a module
 b) taints the kernel

The taints are pretty heavy handed because we're tainting the kernel
*before* we ever even get to load the module into the modules linked
list. That is, it it can fail for other reasons later as we review the
module's structure.

But this commit makes no functional changes, it just makes the intent
clearer and splits the code up where needed to make that happen.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The work to taint the kernel due to a module should be split
up eventually. To aid with this, split up the tainting on
check_modinfo_livepatch().

This let's us bring more early checks together which do return
a value, and makes changes easier to read later where we stuff
all the work to do the taints in one single routine.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The set_license() routine would seem to a reader to do some sort of
setting, but it does not. It just adds a taint if the license is
not set or proprietary.

This makes what the code is doing clearer, so much we can remove
the comment about it.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

This moves check_modinfo() to early_mod_check(). This
doesn't make any functional changes either, as check_modinfo()
was the first call on layout_and_allocate(), so we're just
moving it back one routine and at the end.

This let's us keep separate the checkers from the allocator.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Move early sanity checkers for the module into a helper.
This let's us make it clear when we are working with the
local copy of the module prior to allocation.

This produces no functional changes, it just makes subsequent
changes easier to read.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Add a for_each_modinfo_entry() to make it easier to read and use.
This produces no functional changes but makes this code easiert
to read as we are used to with loops in the kernel and trims more
lines of code.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

This makes it clearer what it is doing. While at it,
make it available to other code other than main.c.
This will be used in the subsequent patch and make
the changes easier to read.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Instead of forward declaring routines for get_modinfo() just move
everything up. This makes no functional changes.
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Use kunmap_local() to unmap pages locally mapped with kmap_local_page().

kunmap_local() must be called on the kernel virtual address returned by
kmap_local_page(), differently from how we use kunmap() which instead
expects the mapped page as its argument.

In module_zstd_decompress() we currently map with kmap_local_page() and
unmap with kunmap(). This breaks the code and so it should be fixed.

Cc: Piotr Gorski <piotrgorski@cachyos.org>
Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Stephen Boyd <swboyd@chromium.org>
Cc: Ira Weiny <ira.weiny@intel.com>
Fixes: 169a58ad ("module/decompress: Support zstd in-kernel decompression")
Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Piotr Gorski <piotrgorski@cachyos.org>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The parameter 'struct module *' in the hook function associated with
{module_}kallsyms_on_each_symbol() is no longer used. Delete it.
Suggested-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Reviewed-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Currently, in dynamic_debug.h we only provide
DEFINE_DYNAMIC_DEBUG_METADATA() and DYNAMIC_DEBUG_BRANCH()
definitions if CONFIG_DYNAMIC_CORE is enabled. Thus, drivers
such as infiniband srp (see: drivers/infiniband/ulp/srp/ib_srp.c)
must provide their own definitions for !CONFIG_DYNAMIC_CORE.

Thus, let's move this !CONFIG_DYNAMIC_CORE case into dynamic_debug.h.
However, the dynamic debug interfaces should really only be defined
if CONFIG_DYNAMIC_DEBUG is set or CONFIG_DYNAMIC_CORE is set along
with DYNAMIC_DEBUG_MODULE, (see:
Documentation/admin-guide/dynamic-debug-howto.rst). Thus, the
undefined case becomes: !((CONFIG_DYNAMIC_DEBUG ||
(CONFIG_DYNAMIC_CORE && DYNAMIC_DEBUG_MODULE)).
With those changes in place, we can remove the !CONFIG_DYNAMIC_CORE
case from ib_srp.c

This change was prompted by a build breakeage in ib_srp.c stemming
from the inclusion of dynamic_debug.h unconditionally in module.h, due
to commit 7deabd67 ("dyndbg: use the module notifier callbacks").
In that case, if we have CONFIG_DYNAMIC_CORE=y and
CONFIG_DYNAMIC_DEBUG=n then the definitions for
DEFINE_DYNAMIC_DEBUG_METADATA() and DYNAMIC_DEBUG_BRANCH() are defined
once in ib_srp.c and then again in the dynamic_debug.h. This had been
working prior to the above referenced commit because dynamic_debug.h
was only pulled into ib_srp.c conditinally via printk.h if
CONFIG_DYNAMIC_DEBUG was set.

Also, the exported functions in lib/dynamic_debug.c itself may
not have a prototype if CONFIG_DYNAMIC_DEBUG=n and
CONFIG_DYNAMIC_CORE=y. This would trigger the -Wmissing-prototypes
warning.

The exported functions are behind (include/linux/dynamic_debug.h):

if defined(CONFIG_DYNAMIC_DEBUG) || \
 (defined(CONFIG_DYNAMIC_DEBUG_CORE) && defined(DYNAMIC_DEBUG_MODULE))

Thus, by adding -DDYNAMIC_CONFIG_MODULE to the lib/Makefile we
can ensure that the exported functions have a prototype in all cases,
since lib/dynamic_debug.c is built whenever
CONFIG_DYNAMIC_DEBUG_CORE=y.

Fixes: 7deabd67 ("dyndbg: use the module notifier callbacks")
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/oe-kbuild-all/202303071444.sIbZTDCy-lkp@intel.com/Signed-off-by: Jason Baron <jbaron@akamai.com>
[mcgrof: adjust commit log, and remove urldefense from URL]
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

There is quite a bit of tribal knowledge around proper use of try_module_get()
and requiring *somehow* the module to still exist to use this call in a way
that is safe. Document this bit of tribal knowledge. To be clear, you should
only use try_module_get() *iff* you are 100% sure the module already does
exist and is not on its way out.

You can be sure the module still exists and is alive through:

1) Direct protection with its refcount: you know some earlier caller called
   __module_get() safely
2) Implied protection: there is an implied protection against module removal

Having an idea of when you are sure __module_get() might be called earlier is
easy to understand however the implied protection requires an example. We use
sysfs an an example for implied protection without a direct module reference
count bump. kernfs / sysfs uses its own internal reference counting for files
being actively used, when such file are active they completely prevent
the module from being removed. kernfs protects this with its kernfs_active().
Effort has been put into verifying the kernfs implied protection works by
using a currently out-of-tree test_sysfs selftest test #32 [0]:

./tools/testing/selftests/sysfs/sysfs.sh -t 0032

Without kernfs / sysfs preventing module removal through its active reference
count (kernfs_active()) the write would fail or worse, a crash would happen in
this test and it does not.

Similar safeguards are required for other users of try_module_get() *iff*
they are not ensuring the above rule 1) is followed.

[0] https://lore.kernel.org/lkml/20211029184500.2821444-4-mcgrof@kernel.org/Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

After a change to linux/module.h, dyndbg.h is now included
indirectly from the decompressor for lz4 support, which in turn
causes a build failure on 32-bit Arm:

In file included from include/linux/module.h:30,
                 from arch/arm/boot/compressed/../../../../lib/lz4/lz4_decompress.c:39,
                 from arch/arm/boot/compressed/../../../../lib/decompress_unlz4.c:10,
                 from arch/arm/boot/compressed/decompress.c:59:
include/linux/dynamic_debug.h: In function 'ddebug_dyndbg_module_param_cb':
include/linux/dynamic_debug.h:307:14: error: implicit declaration of function 'strcmp' [-Werror=implicit-function-declaration]
  307 |         if (!strcmp(param, "dyndbg")) {
      |              ^~~~~~
include/linux/dynamic_debug.h:1:1: note: 'strcmp' is defined in header '<string.h>'; did you forget to '#include <string.h>'?
  +++ |+#include <string.h>

The decompressor has its own replacement for the linux/string.h contents,
so the normal declaration is not visible here. Since the function is
not actually called, it is sufficient to add a declaration, and this
is in fact the correct one as it matches the definition in
arch/arm/boot/compressed/string.c.

Fixes: 7deabd67 ("dyndbg: use the module notifier callbacks")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Jason Baron <jbaron@akamai.com>
Reviewed-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Bring dynamic debug in line with other subsystems by using the module
notifier callbacks. This results in a net decrease in core module
code.

Additionally, Jim Cromie has a new dynamic debug classmap feature,
which requires that jump labels be initialized prior to dynamic debug.
Specifically, the new feature toggles a jump label from the existing
dynamic_debug_setup() function. However, this does not currently work
properly, because jump labels are initialized via the
'module_notify_list' notifier chain, which is invoked after the
current call to dynamic_debug_setup(). Thus, this patch ensures that
jump labels are initialized prior to dynamic debug by setting the
dynamic debug notifier priority to 0, while jump labels have the
higher priority of 1.

Tested by Jim using his new test case, and I've verfied the correct
printing via: # modprobe test_dynamic_debug dyndbg.

Link: https://lore.kernel.org/lkml/20230113193016.749791-21-jim.cromie@gmail.com/Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/oe-kbuild-all/202302190427.9iIK2NfJ-lkp@intel.com/Tested-by: Jim Cromie <jim.cromie@gmail.com>
Reviewed-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
CC: Jim Cromie <jim.cromie@gmail.com>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Jason Baron <jbaron@akamai.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The 'base' parameter to __ddebug_add_module() is no longer in use
after: Commit b7b4eebd ("dyndbg: gather __dyndbg[] state into
struct _ddebug_info").

Cc: Jim Cromie <jim.cromie@gmail.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tested-by: Jim Cromie <jim.cromie@gmail.com>
Reviewed-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
Signed-off-by: Jason Baron <jbaron@akamai.com>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Since commit ee6d3dd4 ("driver core: make kobj_type constant.")
the driver core allows the usage of const struct kobj_type.

Take advantage of this to constify the structure definition to prevent
modification at runtime.
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Correct the struct's field/member name from mod_mem to mem.

Fixes this build error:
../arch/mips/kernel/vpe.c: In function 'vpe_elfload':
../arch/mips/kernel/vpe.c:643:41: error: 'struct module' has no member named 'mod_mem'
  643 |         v->load_addr = alloc_progmem(mod.mod_mem[MOD_TEXT].size);

Fixes: 2ece476a2346 ("module: replace module_layout with module_memory")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Song Liu <song@kernel.org>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: linux-mips@vger.kernel.org
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

The function within is defined in the main.c file, but not called
elsewhere, so remove this unused function.

This routine became no longer used after commit ("module: replace
module_layout with module_memory").

kernel/module/main.c:3007:19: warning: unused function 'within'.
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=4035Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
[mcgrof: adjust commit log to explain why this change is needed]
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

module_layout manages different types of memory (text, data, rodata, etc.)
in one allocation, which is problematic for some reasons:

1. It is hard to enable CONFIG_STRICT_MODULE_RWX.
2. It is hard to use huge pages in modules (and not break strict rwx).
3. Many archs uses module_layout for arch-specific data, but it is not
   obvious how these data are used (are they RO, RX, or RW?)

Improve the scenario by replacing 2 (or 3) module_layout per module with
up to 7 module_memory per module:

        MOD_TEXT,
        MOD_DATA,
        MOD_RODATA,
        MOD_RO_AFTER_INIT,
        MOD_INIT_TEXT,
        MOD_INIT_DATA,
        MOD_INIT_RODATA,

and allocating them separately. This adds slightly more entries to
mod_tree (from up to 3 entries per module, to up to 7 entries per
module). However, this at most adds a small constant overhead to
__module_address(), which is expected to be fast.

Various archs use module_layout for different data. These data are put
into different module_memory based on their location in module_layout.
IOW, data that used to go with text is allocated with MOD_MEM_TYPE_TEXT;
data that used to go with data is allocated with MOD_MEM_TYPE_DATA, etc.

module_memory simplifies quite some of the module code. For example,
ARCH_WANTS_MODULES_DATA_IN_VMALLOC is a lot cleaner, as it just uses a
different allocator for the data. kernel/module/strict_rwx.c is also
much cleaner with module_memory.
Signed-off-by: Song Liu <song@kernel.org>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Guenter Roeck <linux@roeck-us.net>
Cc: Christophe Leroy <christophe.leroy@csgroup.eu>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Reviewed-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>

Commit aa47a7c2 ("lib/cpumask: deprecate nr_cpumask_bits") resulted
in the cpumask operations potentially becoming hugely less efficient,
because suddenly the cpumask was always considered to be variable-sized.

The optimization was then later added back in a limited form by commit
6f9c07be ("lib/cpumask: add FORCE_NR_CPUS config option"), but that
FORCE_NR_CPUS option is not useful in a generic kernel and more of a
special case for embedded situations with fixed hardware.

Instead, just re-introduce the optimization, with some changes.

Instead of depending on CPUMASK_OFFSTACK being false, and then always
using the full constant cpumask width, this introduces three different
cpumask "sizes":

 - the exact size (nr_cpumask_bits) remains identical to nr_cpu_ids.

   This is used for situations where we should use the exact size.

 - the "small" size (small_cpumask_bits) is the NR_CPUS constant if it
   fits in a single word and the bitmap operations thus end up able
   to trigger the "small_const_nbits()" optimizations.

   This is used for the operations that have optimized single-word
   cases that get inlined, notably the bit find and scanning functions.

 - the "large" size (large_cpumask_bits) is the NR_CPUS constant if it
   is an sufficiently small constant that makes simple "copy" and
   "clear" operations more efficient.

   This is arbitrarily set at four words or less.

As a an example of this situation, without this fixed size optimization,
cpumask_clear() will generate code like

        movl    nr_cpu_ids(%rip), %edx
        addq    $63, %rdx
        shrq    $3, %rdx
        andl    $-8, %edx
        callq   memset@PLT

on x86-64, because it would calculate the "exact" number of longwords
that need to be cleared.

In contrast, with this patch, using a MAX_CPU of 64 (which is quite a
reasonable value to use), the above becomes a single

	movq $0,cpumask

instruction instead, because instead of caring to figure out exactly how
many CPU's the system has, it just knows that the cpumask will be a
single word and can just clear it all.

Note that this does end up tightening the rules a bit from the original
version in another way: operations that set bits in the cpumask are now
limited to the actual nr_cpu_ids limit, whereas we used to do the
nr_cpumask_bits thing almost everywhere in the cpumask code.

But if you just clear bits, or scan for bits, we can use the simpler
compile-time constants.

In the process, remove 'cpumask_complement()' and 'for_each_cpu_not()'
which were not useful, and which fundamentally have to be limited to
'nr_cpu_ids'.  Better remove them now than have somebody introduce use
of them later.

Of course, on x86-64 with MAXSMP there is no sane small compile-time
constant for the cpumask sizes, and we end up using the actual CPU bits,
and will generate the above kind of horrors regardless.  Please don't
use MAXSMP unless you really expect to have machines with thousands of
cores.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull crypto fix from Herbert Xu:
 "Fix a regression in the caam driver"

* tag 'v6.3-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: caam - Fix edesc/iv ordering mixup

Pull x86 updates from Thomas Gleixner:
 "A small set of updates for x86:

   - Return -EIO instead of success when the certificate buffer for SEV
     guests is not large enough

   - Allow STIPB to be enabled with legacy IBSR. Legacy IBRS is cleared
     on return to userspace for performance reasons, but the leaves user
     space vulnerable to cross-thread attacks which STIBP prevents.
     Update the documentation accordingly"

* tag 'x86-urgent-2023-03-05' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  virt/sev-guest: Return -EIO if certificate buffer is not large enough
  Documentation/hw-vuln: Document the interaction between IBRS and STIBP
  x86/speculation: Allow enabling STIBP with legacy IBRS