summaryrefslogtreecommitdiffstats
path: root/include/asm-generic
diff options
context:
space:
mode:
authorArnd Bergmann <arnd@arndb.de>2023-05-31 16:33:56 +0200
committerArnd Bergmann <arnd@arndb.de>2023-05-31 16:33:56 +0200
commit3b1ddbb62e7e7d2433781cd2f59d77c2f9aee934 (patch)
tree3eb6d3c791428d993976f45d2ecffcc0c6ee183e /include/asm-generic
parent7877cb91f1081754a1487c144d85dc0d2e2e7fc4 (diff)
parentef7d0f5d03b9e65c9daa8dfe8b405b10566055eb (diff)
downloadlinux-stable-3b1ddbb62e7e7d2433781cd2f59d77c2f9aee934.tar.gz
linux-stable-3b1ddbb62e7e7d2433781cd2f59d77c2f9aee934.tar.bz2
linux-stable-3b1ddbb62e7e7d2433781cd2f59d77c2f9aee934.zip
Merge tag 'virt-to-pfn-for-arch-v6.5-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-integrator into asm-generic
This is an attempt to harden the typing on virt_to_pfn() and pfn_to_virt(). Making virt_to_pfn() a static inline taking a strongly typed (const void *) makes the contract of a passing a pointer of that type to the function explicit and exposes any misuse of the macro virt_to_pfn() acting polymorphic and accepting many types such as (void *), (unitptr_t) or (unsigned long) as arguments without warnings. For symmetry, we do the same with pfn_to_virt(). The problem with this inconsistent typing was pointed out by Russell King: https://lore.kernel.org/linux-arm-kernel/YoJDKJXc0MJ2QZTb@shell.armlinux.org.uk/ And confirmed by Andrew Morton: https://lore.kernel.org/linux-mm/20220701160004.2ffff4e5ab59a55499f4c736@linux-foundation.org/ So the recognition of the problem is widespread. These platforms have been chosen as initial conversion targets: - ARM - ARM64/Aarch64 - asm-generic (including for example x86) - m68k The idea is that if this goes in, it will block further misuse of the function signatures due to the large compile coverage, and then I can go in and fix the remaining architectures on a one-by-one basis. Some of the patches have been circulated before but were not picked up by subsystem maintainers, so now the arch tree is target for this series. It has passed zeroday builds after a lot of iterations in my personal tree, but there could be some randconfig outliers. New added or deeply hidden problems appear all the time so some minor fallout can be expected. * tag 'virt-to-pfn-for-arch-v6.5-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-integrator: m68k/mm: Make pfn accessors static inlines arm64: memory: Make virt_to_pfn() a static inline ARM: mm: Make virt_to_pfn() a static inline asm-generic/page.h: Make pfn accessors static inlines xen/netback: Pass (void *) to virt_to_page() netfs: Pass a pointer to virt_to_page() cifs: Pass a pointer to virt_to_page() in cifsglob cifs: Pass a pointer to virt_to_page() riscv: mm: init: Pass a pointer to virt_to_page() ARC: init: Pass a pointer to virt_to_pfn() in init m68k: Pass a pointer to virt_to_pfn() virt_to_page() fs/proc/kcore.c: Pass a pointer to virt_addr_valid()
Diffstat (limited to 'include/asm-generic')
-rw-r--r--include/asm-generic/page.h12
1 files changed, 10 insertions, 2 deletions
diff --git a/include/asm-generic/page.h b/include/asm-generic/page.h
index c0be2edeb484..9773582fd96e 100644
--- a/include/asm-generic/page.h
+++ b/include/asm-generic/page.h
@@ -74,8 +74,16 @@ extern unsigned long memory_end;
#define __va(x) ((void *)((unsigned long) (x)))
#define __pa(x) ((unsigned long) (x))
-#define virt_to_pfn(kaddr) (__pa(kaddr) >> PAGE_SHIFT)
-#define pfn_to_virt(pfn) __va((pfn) << PAGE_SHIFT)
+static inline unsigned long virt_to_pfn(const void *kaddr)
+{
+ return __pa(kaddr) >> PAGE_SHIFT;
+}
+#define virt_to_pfn virt_to_pfn
+static inline void *pfn_to_virt(unsigned long pfn)
+{
+ return __va(pfn) << PAGE_SHIFT;
+}
+#define pfn_to_virt pfn_to_virt
#define virt_to_page(addr) pfn_to_page(virt_to_pfn(addr))
#define page_to_virt(page) pfn_to_virt(page_to_pfn(page))