Re: [PATCH v17 15/15] selftests, arm64: add a selftest for passing tagged pointers to kernel

From: Szabolcs Nagy
Date: Wed Jun 12 2019 - 08:36:01 EST


On 12/06/2019 12:43, Andrey Konovalov wrote:
> --- /dev/null
> +++ b/tools/testing/selftests/arm64/tags_lib.c
> @@ -0,0 +1,62 @@
> +// SPDX-License-Identifier: GPL-2.0
> +
> +#include <stdlib.h>
> +#include <sys/prctl.h>
> +
> +#define TAG_SHIFT (56)
> +#define TAG_MASK (0xffUL << TAG_SHIFT)
> +
> +#define PR_SET_TAGGED_ADDR_CTRL 55
> +#define PR_GET_TAGGED_ADDR_CTRL 56
> +#define PR_TAGGED_ADDR_ENABLE (1UL << 0)
> +
> +void *__libc_malloc(size_t size);
> +void __libc_free(void *ptr);
> +void *__libc_realloc(void *ptr, size_t size);
> +void *__libc_calloc(size_t nmemb, size_t size);

this does not work on at least musl.

the most robust solution would be to implement
the malloc apis with mmap/munmap/mremap, if that's
too cumbersome then use dlsym RTLD_NEXT (although
that has the slight wart that in glibc it may call
calloc so wrapping calloc that way is tricky).

in simple linux tests i'd just use static or
stack allocations or mmap.

if a generic preloadable lib solution is needed
then do it properly with pthread_once to avoid
races etc.

> +
> +static void *tag_ptr(void *ptr)
> +{
> + static int tagged_addr_err = 1;
> + unsigned long tag = 0;
> +
> + /*
> + * Note that this code is racy. We only use it as a part of a single
> + * threaded test application. Beware of using in multithreaded ones.
> + */
> + if (tagged_addr_err == 1)
> + tagged_addr_err = prctl(PR_SET_TAGGED_ADDR_CTRL,
> + PR_TAGGED_ADDR_ENABLE, 0, 0, 0);
> +
> + if (!ptr)
> + return ptr;
> + if (!tagged_addr_err)
> + tag = rand() & 0xff;
> +
> + return (void *)((unsigned long)ptr | (tag << TAG_SHIFT));
> +}
> +
> +static void *untag_ptr(void *ptr)
> +{
> + return (void *)((unsigned long)ptr & ~TAG_MASK);
> +}
> +
> +void *malloc(size_t size)
> +{
> + return tag_ptr(__libc_malloc(size));
> +}
...