On Thu, May 12, 2016 at 02:35:34PM +0100, Catalin Marinas wrote:

On Thu, May 12, 2016 at 03:20:00AM +0300, Yury Norov wrote:

quoted

I debugged preadv02 and pwritev02 failures and found very weird bug.
Test passes {iovec_base = 0xffffffff, iovec_len = 64} as one element
of vector, and kernel reports successful read/write.

There are 2 problems:
1. How kernel allows such address to be passed to fs subsystem;
2. How fs successes to read/write at non-mapped, and in fact non-user
address.

I don't know the answer on 2'nd question, and it might be something
generic. But I investigated first problem.

The problem is that compat_rw_copy_check_uvector() uses access_ok() to
validate user address, and on arm64 it ends up with checking buffer
end against current_thread_info()->addr_limit.

current_thread_info()->addr_limit for ilp32, and most probably for
aarch32 is equal to aarch64 one, and so adress_ok() doesn't fail.
It happens because on thread creation we call flush_old_exec() to set 
addr_limit, and completely ignore compat mode there.

I assume accesses beyond this address would fault anyway but I haven't
checked the code paths.

quoted

diff --git a/arch/arm64/include/asm/elf.h b/arch/arm64/include/asm/elf.h
index 7a39683..6ba4952 100644
--- a/arch/arm64/include/asm/elf.h
+++ b/arch/arm64/include/asm/elf.h

@@ -146,6 +146,7 @@ typedef struct user_fpsimd_state elf_fpregset_t;
 do {						\
 	clear_thread_flag(TIF_32BIT_AARCH64);	\
 	clear_thread_flag(TIF_32BIT);		\
+	set_fs(TASK_SIZE_64);			\
 } while (0)

See below.

quoted

diff --git a/arch/arm64/include/asm/uaccess.h b/arch/arm64/include/asm/uaccess.h
index 19cfdc5..3b0dd8d 100644
--- a/arch/arm64/include/asm/uaccess.h
+++ b/arch/arm64/include/asm/uaccess.h

@@ -51,7 +51,7 @@
 #define KERNEL_DS	(-1UL)
 #define get_ds()	(KERNEL_DS)
 
-#define USER_DS		TASK_SIZE_64
+#define USER_DS		TASK_SIZE

I agree with this.

quoted

 #define get_fs()	(current_thread_info()->addr_limit)
 
 static inline void set_fs(mm_segment_t fs)

diff --git a/arch/arm64/kernel/binfmt_elf32.c b/arch/arm64/kernel/binfmt_elf32.c
index 5487872..2e8d9f3 100644
--- a/arch/arm64/kernel/binfmt_elf32.c
+++ b/arch/arm64/kernel/binfmt_elf32.c

@@ -12,6 +12,7 @@
 do {						\
 	clear_thread_flag(TIF_32BIT_AARCH64);	\
 	set_thread_flag(TIF_32BIT);		\
+	set_fs(TASK_SIZE_32);			\
 } while (0)
 
 #define COMPAT_ARCH_DLINFO

diff --git a/arch/arm64/kernel/binfmt_ilp32.c b/arch/arm64/kernel/binfmt_ilp32.c
index a934fd4..a8599c6 100644
--- a/arch/arm64/kernel/binfmt_ilp32.c
+++ b/arch/arm64/kernel/binfmt_ilp32.c

@@ -59,6 +59,7 @@ static void cputime_to_compat_timeval(const cputime_t cputime,
 do {									\
 	set_thread_flag(TIF_32BIT_AARCH64);				\
 	clear_thread_flag(TIF_32BIT);					\
+	set_fs(TASK_SIZE_32);						\
 } while (0)

I don't think we need these two. AFAICT, flush_old_exec() takes care of
setting the USER_DS for the new thread.

That's true, but USER_DS depends on personality which is not set yet
for new thread, as I wrote above. In fact, I tried correct USER_DS
only, and it doesn't work

-- 
Catalin