From: "David Hildenbrand (Red Hat)" <david@kernel.org>
To: peng8420.li@gmail.com, linux-mm@kvack.org, akpm@linux-foundation.org
Cc: osalvador@suse.de, jgg@ziepe.ca, jhubbard@nvidia.com,
peterx@redhat.com, linux-kernel@vger.kernel.org,
dan.j.williams@intel.com
Subject: Re: [PATCH 1/2] Remove the "FOLL_TOUCH" test code from gup_test.c.
Date: Fri, 14 Nov 2025 10:06:51 +0100 [thread overview]
Message-ID: <e5c930a6-f10d-4715-aa80-64152b30d96e@kernel.org> (raw)
In-Reply-To: <20251114085038.191031-1-peng8420.li@gmail.com>
On 14.11.25 09:50, peng8420.li@gmail.com wrote:
> From: "peng8420.li" <peng8420.li@gmail.com>
>
> From submission 0f20bba1688bdf3b32df0162511a67d4eda15790, we know that "FOLL_TOUCH" is only for internal scope use;
That is not the proper way to cite commit IDs. Likely checkpatch.pl would complain here.
"Ever since commit 0f20bba1688b ("mm/gup: explicitly define and check
internal GUP flags, disallow FOLL_TOUCH") we marked FOLL_TOUCH as a GUP-internal flag."
>
> Therefore, remove the "FOLL_TOUCH" test code from gup_test.c;
>
> Otherwise, executing the test command "./gup_test -L -r 100 -z" will report the following warning log:
> TAP version 13
> 1..1
> ENCHMARK: Time: [ 39.363371] WARNING: CPU: 1 PID: 117 at mm/gup.c:2512 is_valid_gup_args+0x66/0x8c
> get:2818 put:46 [ 39.364043] Modules linked in:
> us# , truncated [ 39.364311] CPU: 1 UID: 0 PID: 117 Comm: gup_test Not tainted 6.18.0-rc5-00324-gd09eaf415c87 #29 NONE
> (size: 0)#
> [ 39.364434] Hardware name: riscv-virtio,qemu (DT)
> [ 39.364546] epc : is_valid_gup_args+0x66/0x8c
> [ 39.364596] ra : pin_user_pages+0x38/0x78
> [ 39.364630] epc : ffffffff802079e6 ra : ffffffff8020c214 sp : ff2000000041bd20
> [ 39.364650] gp : ffffffff81a26068 tp : ff60000080b3b000 t0 : ff2000000041bdf8
> [ 39.364678] t1 : 000000000000001e t2 : 0000000000000000 s0 : ff2000000041bd30
> [ 39.364697] s1 : ff60000081300000 a0 : ff60000081300000 a1 : 0000000000000000
> [ 39.364714] a2 : ff2000000041bd3c a3 : 0000000000080000 a4 : 0000000000000001
> [ 39.364731] a5 : 0000000000010101 a6 : 0000000000000001 a7 : 0000000000000000
> [ 39.364747] s2 : 00007fff7eeed000 s3 : 0000000000000001 s4 : 00007fff7eeee000
> [ 39.364761] s5 : 00007fff7eeeb838 s6 : 0000000000000000 s7 : 00007fff7eeed000
> [ 39.364825] s8 : ff60000081300000 s9 : 0000000000000000 s10: 0000000000000002
> [ 39.364842] s11: 000000092869bfdc t3 : 2152ffffffffffc0 t4 : 00000000001fffff
> [ 39.364855] t5 : ffffffffffffffff t6 : 0000000000000000
> [ 39.364867] status: 0000000200000120 badaddr: ffffffff802079e6 cause: 0000000000000003
> [ 39.364949] [<ffffffff802079e6>] is_valid_gup_args+0x66/0x8c
> [ 39.365036] [<ffffffff8020c214>] pin_user_pages+0x38/0x78
> [ 39.365049] [<ffffffff80278f78>] gup_test_ioctl+0x2b4/0xc08
> [ 39.365060] [<ffffffff80297da6>] __riscv_sys_ioctl+0xba/0xc4
> [ 39.365072] [<ffffffff80b93f96>] do_trap_ecall_u+0x296/0x370
> [ 39.365093] [<ffffffff80b9ef06>] handle_exception+0x146/0x152
> [ 39.365220] ---[ end trace 0000000000000000 ]---
> ok 1 ioctl status 0
Right, gup_tests simply forward any flags we pass in. I wonder if we would
want to filter them for ones we expect to test?
I guess we should add
Fixes: 0f20bba1688b ("mm/gup: explicitly define and check internal GUP flags, disallow FOLL_TOUCH")
Although it's a bit complicated with gup_tests.
>
> Signed-off-by: peng8420.li <peng8420.li@gmail.com>
> ---
> tools/testing/selftests/mm/gup_test.c | 22 ++++------------------
> 1 file changed, 4 insertions(+), 18 deletions(-)
>
> diff --git a/tools/testing/selftests/mm/gup_test.c b/tools/testing/selftests/mm/gup_test.c
> index 8900b840c17a..75f7134d529d 100644
> --- a/tools/testing/selftests/mm/gup_test.c
> +++ b/tools/testing/selftests/mm/gup_test.c
> @@ -19,7 +19,6 @@
>
> /* Just the flags we need, copied from mm.h: */
> #define FOLL_WRITE 0x01 /* check pte is writable */
> -#define FOLL_TOUCH 0x02 /* mark page accessed */
>
> #define GUP_TEST_FILE "/sys/kernel/debug/gup_test"
>
> @@ -93,7 +92,7 @@ int main(int argc, char **argv)
> {
> struct gup_test gup = { 0 };
> int filed, i, opt, nr_pages = 1, thp = -1, write = 1, nthreads = 1, ret;
> - int flags = MAP_PRIVATE, touch = 0;
> + int flags = MAP_PRIVATE;
> char *file = "/dev/zero";
> pthread_t *tid;
> char *p;
> @@ -170,10 +169,6 @@ int main(int argc, char **argv)
> case 'H':
> flags |= (MAP_HUGETLB | MAP_ANONYMOUS);
> break;
> - case 'z':
> - /* fault pages in gup, do not fault in userland */
> - touch = 1;
> - break;
AFAIKT, run_vmtests.sh does not set that flag, so dropping it should be fine.
--
Cheers
David
next prev parent reply other threads:[~2025-11-14 9:06 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-14 8:50 [PATCH 1/2] Remove the "FOLL_TOUCH" test code from gup_test.c peng8420.li
2025-11-14 8:50 ` [PATCH 2/2] Fix the name of the source file for the "FOLL_WRITE" macro copy peng8420.li
2025-11-14 8:57 ` David Hildenbrand (Red Hat)
2025-11-14 9:08 ` peng li
2025-11-14 9:06 ` David Hildenbrand (Red Hat) [this message]
2025-11-14 14:14 ` [PATCH 1/2] Remove the "FOLL_TOUCH" test code from gup_test.c peng li
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=e5c930a6-f10d-4715-aa80-64152b30d96e@kernel.org \
--to=david@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=dan.j.williams@intel.com \
--cc=jgg@ziepe.ca \
--cc=jhubbard@nvidia.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=osalvador@suse.de \
--cc=peng8420.li@gmail.com \
--cc=peterx@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox