Merge: Update kernel-module support to v6.8

MR: https://gitlab.com/redhat/centos-stream/src/kernel/centos-stream-9/-/merge_requests/4061

# Merge Request Required Information

## Summary of Changes
RHIVOS is running into early mm init performance issues, and a long-term set of solutions is to improve the kernel linear map when kernel security is set to a max-level, a RHIVOS FuSa requirement, where all of memory is -not- read/writeable via the linear map (all of memory mapping from PAGE_OFFSET), but has strict execute-only, rodata, rw-data and no-execute pages.
Although RHEL9 and upstream can support the latter functionally, it is a significant performance issue as page-level mapping of the kernel linear map has to be employed from the default huge-page mappings that the various arch's support. The boot kernel itself is relatively easy to know how to map for optimal page-mappings and protection, because it is the first to load and ELF sections can be scanned for needed info; the same can't be said for all the loadable kernel modules, which is the impetus for page-splitting of the linear map (on x86) and the per-page-mapping on ARM64, where page-splitting of the linear map is not supported, but is the long-term optimal solution.
In order to make a step in this long-term effort, this patch series attempts to take the existing RHEL9 kernel module load support, which is barely 12 patches past the initial v5.14 base, and bring it up to a current, v6.8 version.
Of course, such an update brings a lot of other needed backports to apply cleanly, if the goal is to get close to upstream, maintain RHEL kmod support, and not regress.   Thus, this series results with major updates to dynamic-debug (since it involves modifying kernel module sections), kbuild, modpost, genksyms, and sprinkle an odd livepatch, fpatch, and BPF patch, although the latter were trimmed or dropped wherever possible.  
The split is approximately 150 kernel-module, 30 dyndbg, 80 modpost, 15 kbuild, 3 livepatch, 3 ftrace, 2 bpf (one being a fix for earlier kernel commit).
Note: modpost and related kbuild updates moved it to approximately v6.4.  A full update to 6.8 wasn't deemed necessary, and was an additional 30+ commits, and more kbuild modifications.  This effort was deemed sufficiently large and complete for the intended goal of making RHEL9 amenable to future updates to the kernel-module subsystem for posted patches on review now in linux-mm by Mike Rapaport. Those patches and expected follow-ons, will be backported to RHEL-9 when upstream settles on final updates in this area; these updates will make the kernel-load subsystem more common, and less arch-specific.
One patch from v6.9-rc1 was taken, modules: wait do_free_init correctly, to repair a race seen in the module-load path on a RHIVOS platform, which needed to sit on top of this series for ease of backporting.

v6: Evidently the rebase to -457 kept a merge conflict, which was a duplicate patch already taken in.  Latest series is now 299 patches vs 300.  No functional changes!

v5: rebased to latest kernel (-457) since gitlab punted due to claimed merge conflict; only conflict was relative source, due to other MRs pulled into cs9/9.5 ahead of this MR; no code changes, and (tkdiff+)diff-ing v4 patches to v5, showed no diffs to the author's naked eye.

v4: Just updated 3rd patch's revert to put Upstream status *after* Subject, so it shows correctly in a git-format output. No code changes from v3. (although CKI running a-muck after push'd update w/only a commit-log change).

v3: Rebase to -455 kernel since v2 was 8300 commits behind and had merge conflicts with JoeL's objtool update MR.

v2: Pulling out of Draft.
  : (Hopefully) fixed numerous nits (Jira: -> JIRA:; proper link so no more 404's, etc.)
  : add new/latest Fixes, some id'd by reviewers, some new to v6.9
  : Cleaned up/out bad merges that had introduced RHEL-only hunks
  : Significantly re-ordered the series to make it more bisectable; still breaks where the upstream maintainer tore code out of modpost.c and into a sed script, and then put the functionality back into modpost.c, and removed the sed script, which this series didn't backport since it was already large enough.
  : identified a failure with systemtap, that Will Cohen is repairing;  thus, this MR has to wait for a systemtap update before it will pass its check in the (brew? cki?) builds.
 
v1: Draft! 
This series has gone through some simple, preliminary testing, but it needs deep review by ftrace, BPF, livepatch, and rh-kabi support to ensure no regressions in these few, but corner kernel-modifying code paths.  rh-kabi tooling is a bit unknown, as it isn't in the kernel, but there are RHEL-only patches in the kernel for it.  
A patchreview run against the series was exed'd, and needed Fixes were added/included.
The list of self-documented omissions is listed below.  If new ones have popped in v6.9-rc<n>,
please forward them for addition.
Bisectability: The series is has known bisectability (patch-ordering) issues at the moment, but plan to re-shuffle the patches in v2 to improve if not make it completely bisectable.
Expected feedback will be incorporated in v2, and planned upgrade to full-MR/drop-Draft status.

Shout-out to Joe Lawrence who aided in debugging and providing fixes for well-hidden noarch build failures around Documentation generation, as well as warning cleanups for EXPORT'd init-tagged functions, which the update checks for now.  Joe was instrumental in finding key chunks of the modpost update that appears to have closed gaps in my original backport efforts.

Intentionally Omitted Fix: 0aa24a7 kbuild: do not try to parse *.cmd files for objects provided by compiler
   -- for parisc & sky arch's, not needed in RHEL9

Intentionally Omitted Fix: f5983da modpost: define more R_ARM_* for old distributions
  For old releases not having R_ARM_* in arch/arm/include/asm/elf.h, which RHEL9 has

Intentionally Omitted Fix: 08700ec linux/export: fix reference to exported functions for parisc64
 -- no parisc64 support in RHEL9

Intentionally Omitted Fix: 86495af media: dvb: symbol fixup for dvb_attach()
 -- not included in this backport due to partner request not to include until RHEL-10

Intentionally Omitted Fix: d81f0d7 Subject: kunit: add KUNIT_INIT_TABLE to init link
  -- will let KUNIT update bring in and enable as needed

## Approved Development Ticket
JIRA: https://issues.redhat.com/browse/RHEL-28063

Signed-off-by: Donald Dutile <ddutile@redhat.com>

Approved-by: Chris von Recklinghausen <crecklin@redhat.com>
Approved-by: David Arcari <darcari@redhat.com>
Approved-by: CKI KWF Bot <cki-ci-bot+kwf-gitlab-com@redhat.com>
Approved-by: Eric Chanudet <echanude@redhat.com>

Merged-by: Lucas Zampieri <lzampier@redhat.com>

Author: zampierilucas

Documentation/admin-guide/dynamic-debug-howto.rst

@@ -229,6 +229,7 @@ The flags are::
 
   p    enables the pr_debug() callsite.
   f    Include the function name in the printed message
+  s    Include the source file name in the printed message
   l    Include line number in the printed message
   m    Include module name in the printed message
   t    Include thread ID in messages not generated from interrupt context
@@ -240,17 +241,16 @@ have meaning, other flags ignored.
 For display, the flags are preceded by ``=``
 (mnemonic: what the flags are currently equal to).
 
-Note the regexp ``^[-+=][flmpt_]+$`` matches a flags specification.
-To clear all flags at once, use ``=_`` or ``-flmpt``.
+Note the regexp ``^[-+=][fslmpt_]+$`` matches a flags specification.
+To clear all flags at once, use ``=_`` or ``-fslmpt``.
 
 
 Debug messages during Boot Process
 ==================================
 
 To activate debug messages for core code and built-in modules during
 the boot process, even before userspace and debugfs exists, use
-``dyndbg="QUERY"``, ``module.dyndbg="QUERY"``, or ``ddebug_query="QUERY"``
-(``ddebug_query`` is obsoleted by ``dyndbg``, and deprecated).  QUERY follows
+``dyndbg="QUERY"`` or ``module.dyndbg="QUERY"``.  QUERY follows
 the syntax described above, but must not exceed 1023 characters.  Your
 bootloader may impose lower limits.
 
@@ -270,8 +270,7 @@ this boot parameter for debugging purposes.
 
 If ``foo`` module is not built-in, ``foo.dyndbg`` will still be processed at
 boot time, without effect, but will be reprocessed when module is
-loaded later. ``ddebug_query=`` and bare ``dyndbg=`` are only processed at
-boot.
+loaded later. Bare ``dyndbg=`` is only processed at boot.
 
 
 Debug Messages at Module Initialization Time
@@ -358,7 +357,7 @@ Examples
   // boot-args example, with newlines and comments for readability
   Kernel command line: ...
     // see what's going on in dyndbg=value processing
-    dynamic_debug.verbose=1
+    dynamic_debug.verbose=3
     // enable pr_debugs in 2 builtins, #cmt is stripped
     dyndbg="module params +p #cmt ; module sys +p"
     // enable pr_debugs in 2 functions in a module loaded later

Documentation/admin-guide/kernel-parameters.txt

@@ -918,11 +918,6 @@
 			Format: <port#>,<type>
 			See also Documentation/input/devices/joystick-parport.rst
 
-	ddebug_query=	[KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
-			time. See
-			Documentation/admin-guide/dynamic-debug-howto.rst for
-			details.  Deprecated, see dyndbg.
-
 	debug		[KNL] Enable kernel debugging (events log level).
 
 	debug_boot_weak_hash
@@ -1190,8 +1185,12 @@
 	nopku		[X86] Disable Memory Protection Keys CPU feature found
 			in some Intel CPUs.
 
-	<module>.async_probe [KNL]
-			Enable asynchronous probe on this module.
+	<module>.async_probe[=<bool>] [KNL]
+			If no <bool> value is specified or if the value
+			specified is not a valid <bool>, enable asynchronous
+			probe on this module.  Otherwise, enable/disable
+			asynchronous probe on this module as indicated by the
+			<bool> value. See also: module.async_probe
 
 	early_ioremap_debug [KNL]
 			Enable debug messages in early_ioremap support. This
@@ -3433,6 +3432,21 @@
 			For details see:
 			Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
 
+	module.async_probe=<bool>
+			[KNL] When set to true, modules will use async probing
+			by default. To enable/disable async probing for a
+			specific module, use the module specific control that
+			is documented under <module>.async_probe. When both
+			module.async_probe and <module>.async_probe are
+			specified, <module>.async_probe takes precedence for
+			the specific module.
+
+	module.enable_dups_trace
+			[KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set,
+			this means that duplicate request_module() calls will
+			trigger a WARN_ON() instead of a pr_warn(). Note that
+			if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s
+			will always be issued and this option does nothing.
 	module.sig_enforce
 			[KNL] When CONFIG_MODULE_SIG is set, this means that
 			modules without (valid) signatures will fail to load.

Documentation/core-api/kernel-api.rst

@@ -216,12 +216,30 @@ relay interface
 Module Support
 ==============
 
-Module Loading
---------------
+Kernel module auto-loading
+--------------------------
 
-.. kernel-doc:: kernel/kmod.c
+.. kernel-doc:: kernel/module/kmod.c
    :export:
 
+Module debugging
+----------------
+
+.. kernel-doc:: kernel/module/stats.c
+   :doc: module debugging statistics overview
+
+dup_failed_modules - tracks duplicate failed modules
+****************************************************
+
+.. kernel-doc:: kernel/module/stats.c
+   :doc: dup_failed_modules - tracks duplicate failed modules
+
+module statistics debugfs counters
+**********************************
+
+.. kernel-doc:: kernel/module/stats.c
+   :doc: module statistics debugfs counters
+
 Inter Module support
 --------------------
 

MAINTAINERS

@@ -10725,6 +10725,7 @@ F:	drivers/tty/serial/kgdboc.c
 F:	include/linux/kdb.h
 F:	include/linux/kgdb.h
 F:	kernel/debug/
+F:	kernel/module/kdb.c
 
 KHADAS MCU MFD DRIVER
 M:	Neil Armstrong <narmstrong@baylibre.com>
@@ -11182,6 +11183,7 @@ F:	arch/s390/include/asm/livepatch.h
 F:	arch/x86/include/asm/livepatch.h
 F:	include/linux/livepatch.h
 F:	kernel/livepatch/
+F:	kernel/module/livepatch.c
 F:	lib/livepatch/
 F:	samples/livepatch/
 F:	tools/testing/selftests/livepatch/
@@ -13050,7 +13052,7 @@ M:	Jessica Yu <jeyu@kernel.org>
 S:	Maintained
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/jeyu/linux.git modules-next
 F:	include/linux/module.h
-F:	kernel/module.c
+F:	kernel/module/
 
 MONOLITHIC POWER SYSTEM PMIC DRIVER
 M:	Saravanan Sekar <sravanhome@gmail.com>

arch/Kconfig

@@ -791,16 +791,6 @@ config CFI_CLANG
 
 	    https://clang.llvm.org/docs/ControlFlowIntegrity.html
 
-config CFI_CLANG_SHADOW
-	bool "Use CFI shadow to speed up cross-module checks"
-	default y
-	depends on CFI_CLANG && MODULES
-	help
-	  If you select this option, the kernel builds a fast look-up table of
-	  CFI check functions in loaded modules to reduce performance overhead.
-
-	  If unsure, say Y.
-
 config CFI_PERMISSIVE
 	bool "Use CFI in permissive mode"
 	depends on CFI_CLANG
@@ -946,6 +936,12 @@ config MODULES_USE_ELF_REL
 	  Modules only use ELF REL relocations.  Modules with ELF RELA
 	  relocations will give an error.
 
+config ARCH_WANTS_MODULES_DATA_IN_VMALLOC
+	bool
+	help
+	  For architectures like powerpc/32 which have constraints on module
+	  allocation and need to allocate module data outside of module area.
+
 config HAVE_IRQ_EXIT_ON_IRQ_STACK
 	bool
 	help

arch/arc/kernel/unwind.c

@@ -374,6 +374,8 @@ void *unwind_add_table(struct module *module, const void *table_start,
 		       unsigned long table_size)
 {
 	struct unwind_table *table;
+	struct module_memory *core_text;
+	struct module_memory *init_text;
 
 	if (table_size <= 0)
 		return NULL;
@@ -382,11 +384,11 @@ void *unwind_add_table(struct module *module, const void *table_start,
 	if (!table)
 		return NULL;
 
-	init_unwind_table(table, module->name,
-			  module->core_layout.base, module->core_layout.size,
-			  module->init_layout.base, module->init_layout.size,
-			  table_start, table_size,
-			  NULL, 0);
+	core_text = &module->mem[MOD_TEXT];
+	init_text = &module->mem[MOD_INIT_TEXT];
+
+	init_unwind_table(table, module->name, core_text->base, core_text->size,
+			  init_text->base, init_text->size, table_start, table_size, NULL, 0);
 
 	init_unwind_hdr(table, unw_hdr_alloc);
 

arch/arm/Kconfig

@@ -88,6 +88,7 @@ config ARM
 	select HAVE_EBPF_JIT if !CPU_ENDIAN_BE32
 	select HAVE_CONTEXT_TRACKING_USER
 	select HAVE_C_RECORDMCOUNT
+	select HAVE_BUILDTIME_MCOUNT_SORT
 	select HAVE_DEBUG_KMEMLEAK if !XIP_KERNEL
 	select HAVE_DMA_CONTIGUOUS if MMU
 	select HAVE_DYNAMIC_FTRACE if !XIP_KERNEL && !CPU_ENDIAN_BE32 && MMU

arch/arm/boot/compressed/decompress.c

@@ -31,6 +31,7 @@
 /* Not needed, but used in some headers pulled in by decompressors */
 extern char * strstr(const char * s1, const char *s2);
 extern size_t strlen(const char *s);
+extern int strcmp(const char *cs, const char *ct);
 extern int memcmp(const void *cs, const void *ct, size_t count);
 extern char * strchrnul(const char *, int);
 

arch/arm/kernel/module-plts.c

@@ -28,11 +28,6 @@ static const u32 fixed_plts[] = {
 #endif
 };
 
-static bool in_init(const struct module *mod, unsigned long loc)
-{
-	return loc - (u32)mod->init_layout.base < mod->init_layout.size;
-}
-
 static void prealloc_fixed(struct mod_plt_sec *pltsec, struct plt_entries *plt)
 {
 	int i;
@@ -50,8 +45,8 @@ static void prealloc_fixed(struct mod_plt_sec *pltsec, struct plt_entries *plt)
 
 u32 get_module_plt(struct module *mod, unsigned long loc, Elf32_Addr val)
 {
-	struct mod_plt_sec *pltsec = !in_init(mod, loc) ? &mod->arch.core :
-							  &mod->arch.init;
+	struct mod_plt_sec *pltsec = !within_module_init(loc, mod) ?
+						&mod->arch.core : &mod->arch.init;
 	struct plt_entries *plt;
 	int idx;
 

arch/arm64/kernel/module-plts.c

@@ -66,17 +66,12 @@ static bool plt_entries_equal(const struct plt_entry *a,
 	       (q + aarch64_insn_adrp_get_offset(le32_to_cpu(b->adrp)));
 }
 
-static bool in_init(const struct module *mod, void *loc)
-{
-	return (u64)loc - (u64)mod->init_layout.base < mod->init_layout.size;
-}
-
 u64 module_emit_plt_entry(struct module *mod, Elf64_Shdr *sechdrs,
 			  void *loc, const Elf64_Rela *rela,
 			  Elf64_Sym *sym)
 {
-	struct mod_plt_sec *pltsec = !in_init(mod, loc) ? &mod->arch.core :
-							  &mod->arch.init;
+	struct mod_plt_sec *pltsec = !within_module_init((unsigned long)loc, mod) ?
+						&mod->arch.core : &mod->arch.init;
 	struct plt_entry *plt = (struct plt_entry *)sechdrs[pltsec->plt_shndx].sh_addr;
 	int i = pltsec->plt_num_entries;
 	int j = i - 1;
@@ -106,8 +101,8 @@ u64 module_emit_plt_entry(struct module *mod, Elf64_Shdr *sechdrs,
 u64 module_emit_veneer_for_adrp(struct module *mod, Elf64_Shdr *sechdrs,
 				void *loc, u64 val)
 {
-	struct mod_plt_sec *pltsec = !in_init(mod, loc) ? &mod->arch.core :
-							  &mod->arch.init;
+	struct mod_plt_sec *pltsec = !within_module_init((unsigned long)loc, mod) ?
+						&mod->arch.core : &mod->arch.init;
 	struct plt_entry *plt = (struct plt_entry *)sechdrs[pltsec->plt_shndx].sh_addr;
 	int i = pltsec->plt_num_entries++;
 	u32 br;