Le 03/05/2021 à 07:39, Jordan Niethe a écrit :
On Thu, Apr 29, 2021 at 3:04 PM Christophe Leroy
<christophe.le...@csgroup.eu> wrote:
Le 29/04/2021 à 05:15, Jordan Niethe a écrit :
If MODULES_{VADDR,END} are not defined set them to VMALLOC_START and
VMALLOC_END respectively. This reduces the need for special cases. For
example, powerpc's module_alloc() was previously predicated on
MODULES_VADDR being defined but now is unconditionally defined.
This will be useful reducing conditional code in other places that need
to allocate from the module region (i.e., kprobes).
Signed-off-by: Jordan Niethe <jniet...@gmail.com>
---
v10: New to series
v11: - Consider more places MODULES_VADDR was being used
---
arch/powerpc/include/asm/pgtable.h | 11 +++++++++++
arch/powerpc/kernel/module.c | 5 +----
arch/powerpc/mm/kasan/kasan_init_32.c | 10 +++++-----
arch/powerpc/mm/ptdump/ptdump.c | 4 ++--
4 files changed, 19 insertions(+), 11 deletions(-)
diff --git a/arch/powerpc/include/asm/pgtable.h
b/arch/powerpc/include/asm/pgtable.h
index c6a676714f04..882fda779648 100644
--- a/arch/powerpc/include/asm/pgtable.h
+++ b/arch/powerpc/include/asm/pgtable.h
@@ -39,6 +39,17 @@ struct mm_struct;
#define __S110 PAGE_SHARED_X
#define __S111 PAGE_SHARED_X
+#ifndef MODULES_VADDR
+#define MODULES_VADDR VMALLOC_START
+#define MODULES_END VMALLOC_END
+#endif
+
+#if defined(CONFIG_PPC_BOOK3S_32) && defined(CONFIG_STRICT_KERNEL_RWX)
No no.
TASK_SIZE > MODULES_VADDR is ALWAYS wrong, for any target, in any configuration.
Why is it a problem to leave the test as a BUILD_BUG_ON() in module_alloc() ?
On ppc64s, MODULES_VADDR is __vmalloc_start (a variable) and
TASK_SIZE depends on current.
Also for nohash like 44x, MODULES_VADDR is defined based on high_memory.
If I put it back in module_alloc() and wrap it with #ifdef
CONFIG_PPC_BOOK3S_32 will that be fine?
Thinking about it once more, I think the best approach is the one taken by Nick in
https://patchwork.ozlabs.org/project/linuxppc-dev/patch/20210502110050.324953-1-npig...@gmail.com/
Use MODULES_VADDR/MODULES_END when it exists, use VMALLOC_START/VMALLOC_END
otherwise.
I know I suggested to always define MODULES_VADDR, but maybe that's not the
best solution at the end.
For kprobes, is there a way to re-use functions from modules.c in
alloc_insn_page() ?
Christophe
