Thread (29 messages) 29 messages, 3 authors, 2024-12-11

Re: [RFC PATCH 13/16] arm64: mm: Reset pkey in __tlb_remove_table()

From: Kevin Brodsky <hidden>
Date: 2024-12-10 09:28:53
Also in: linux-hardening, lkml

On 09/12/2024 11:29, Peter Zijlstra wrote:
On Fri, Dec 06, 2024 at 10:11:07AM +0000, Kevin Brodsky wrote:
quoted
[...]
diff --git a/arch/arm64/include/asm/tlb.h b/arch/arm64/include/asm/tlb.h
index a947c6e784ed..d1611ffa6d91 100644
--- a/arch/arm64/include/asm/tlb.h
+++ b/arch/arm64/include/asm/tlb.h
@@ -10,10 +10,14 @@
 
 #include <linux/pagemap.h>
 #include <linux/swap.h>
+#include <linux/kpkeys.h>
 
 static inline void __tlb_remove_table(void *_table)
 {
-	free_page_and_swap_cache((struct page *)_table);
+	struct page *page = (struct page *)_table;
+
+	kpkeys_unprotect_pgtable_memory((unsigned long)page_address(page), 1);
+	free_page_and_swap_cache(page);
 }
Same as for the others, perhaps stick this in generic code instead of in
the arch code?
This should be doable, with some refactoring. __tlb_remove_table() is
currently called from two functions in mm/mmu_gather.c, I suppose I
could create a wrapper there that calls
kpkeys_unprotect_pgtable_memory() and then __tlb_remove_table(). Like in
the p4d case I do however wonder how robust this is, as
__tlb_remove_table() could end up being called from other places.

- Kevin
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help