execmem: add support for cache of large ROX pages

jira LE-4694
Rebuild_History Non-Buildable kernel-6.12.0-55.43.1.el10_0
commit-author Mike Rapoport (Microsoft) <[email protected]>
commit 2e45474

Using large pages to map text areas reduces iTLB pressure and improves
performance.

Extend execmem_alloc() with an ability to use huge pages with ROX
permissions as a cache for smaller allocations.

To populate the cache, a writable large page is allocated from vmalloc
with VM_ALLOW_HUGE_VMAP, filled with invalid instructions and then
remapped as ROX.

The direct map alias of that large page is exculded from the direct map.

Portions of that large page are handed out to execmem_alloc() callers
without any changes to the permissions.

When the memory is freed with execmem_free() it is invalidated again so
that it won't contain stale instructions.

An architecture has to implement execmem_fill_trapping_insns() callback
and select ARCH_HAS_EXECMEM_ROX configuration option to be able to use the
ROX cache.

The cache is enabled on per-range basis when an architecture sets
EXECMEM_ROX_CACHE flag in definition of an execmem_range.

Link: https://lkml.kernel.org/r/[email protected]
	Signed-off-by: Mike Rapoport (Microsoft) <[email protected]>
	Reviewed-by: Luis Chamberlain <[email protected]>
	Tested-by: kdevops <[email protected]>
	Cc: Andreas Larsson <[email protected]>
	Cc: Andy Lutomirski <[email protected]>
	Cc: Ard Biesheuvel <[email protected]>
	Cc: Arnd Bergmann <[email protected]>
	Cc: Borislav Petkov (AMD) <[email protected]>
	Cc: Brian Cain <[email protected]>
	Cc: Catalin Marinas <[email protected]>
	Cc: Christophe Leroy <[email protected]>
	Cc: Christoph Hellwig <[email protected]>
	Cc: Dave Hansen <[email protected]>
	Cc: Dinh Nguyen <[email protected]>
	Cc: Geert Uytterhoeven <[email protected]>
	Cc: Guo Ren <[email protected]>
	Cc: Helge Deller <[email protected]>
	Cc: Huacai Chen <[email protected]>
	Cc: Ingo Molnar <[email protected]>
	Cc: Johannes Berg <[email protected]>
	Cc: John Paul Adrian Glaubitz <[email protected]>
	Cc: Kent Overstreet <[email protected]>
	Cc: Liam R. Howlett <[email protected]>
	Cc: Mark Rutland <[email protected]>
	Cc: Masami Hiramatsu (Google) <[email protected]>
	Cc: Matt Turner <[email protected]>
	Cc: Max Filippov <[email protected]>
	Cc: Michael Ellerman <[email protected]>
	Cc: Michal Simek <[email protected]>
	Cc: Oleg Nesterov <[email protected]>
	Cc: Palmer Dabbelt <[email protected]>
	Cc: Peter Zijlstra <[email protected]>
	Cc: Richard Weinberger <[email protected]>
	Cc: Russell King <[email protected]>
	Cc: Song Liu <[email protected]>
	Cc: Stafford Horne <[email protected]>
	Cc: Steven Rostedt (Google) <[email protected]>
	Cc: Suren Baghdasaryan <[email protected]>
	Cc: Thomas Bogendoerfer <[email protected]>
	Cc: Thomas Gleixner <[email protected]>
	Cc: Uladzislau Rezki (Sony) <[email protected]>
	Cc: Vineet Gupta <[email protected]>
	Cc: Will Deacon <[email protected]>
	Signed-off-by: Andrew Morton <[email protected]>
(cherry picked from commit 2e45474)
	Signed-off-by: Jonathan Maple <[email protected]>

Author: PlaidCat

arch/Kconfig

@@ -1024,6 +1024,14 @@ config ARCH_WANTS_EXECMEM_LATE
 	  enough entropy for module space randomization, for instance
 	  arm64.
 
+config ARCH_HAS_EXECMEM_ROX
+	bool
+	depends on MMU && !HIGHMEM
+	help
+	  For architectures that support allocations of executable memory
+	  with read-only execute permissions. Architecture must implement
+	  execmem_fill_trapping_insns() callback to enable this.
+
 config HAVE_IRQ_EXIT_ON_IRQ_STACK
 	bool
 	help

include/linux/execmem.h

@@ -53,6 +53,20 @@ enum execmem_range_flags {
 	EXECMEM_ROX_CACHE	= (1 << 1),
 };
 
+#ifdef CONFIG_ARCH_HAS_EXECMEM_ROX
+/**
+ * execmem_fill_trapping_insns - set memory to contain instructions that
+ *				 will trap
+ * @ptr:	pointer to memory to fill
+ * @size:	size of the range to fill
+ * @writable:	is the memory poited by @ptr is writable or ROX
+ *
+ * A hook for architecures to fill execmem ranges with invalid instructions.
+ * Architectures that use EXECMEM_ROX_CACHE must implement this.
+ */
+void execmem_fill_trapping_insns(void *ptr, size_t size, bool writable);
+#endif
+
 /**
  * struct execmem_range - definition of an address space suitable for code and
  *			  related data allocations