From: Thomas Schwinge <thomas@codesourcery.com>
To: Jakub Jelinek <jakub@redhat.com>
Cc: <gcc-patches@gcc.gnu.org>,
Tobias Burnus <tobias@codesourcery.com>,
"Marcel Vollweiler" <marcel@codesourcery.com>
Subject: Re: [committed] openmp: Conforming device numbers and omp_{initial,invalid}_device
Date: Wed, 15 Jun 2022 22:58:49 +0200 [thread overview]
Message-ID: <878rpx7ieu.fsf@euler.schwinge.homeip.net> (raw)
In-Reply-To: <YqmcZiX9pJDYPRb0@tucnak>
Hi Jakub!
On 2022-06-15T10:46:30+0200, Jakub Jelinek <jakub@redhat.com> wrote:
> On Tue, Jun 14, 2022 at 06:41:37PM +0200, Thomas Schwinge wrote:
>> On 2022-06-13T14:06:39+0200, Jakub Jelinek via Gcc-patches <gcc-patches@gcc.gnu.org> wrote:
>> > OpenMP 5.2 changed once more what device numbers are allowed.
>>
>> > libgomp/
>>
>> > * testsuite/libgomp.c-c++-common/target-is-accessible-1.c (main): Add
>> > test with omp_initial_device. Use -5 instead of -1 for negative value
>> > test.
>> > * testsuite/libgomp.fortran/target-is-accessible-1.f90 (main):
>> > Likewise. Reorder stop numbers.
>>
>> In an offloading configuration, I'm seeing:
>>
>> PASS: libgomp.fortran/get-mapped-ptr-1.f90 -O (test for excess errors)
>> [-PASS:-]{+FAIL:+} libgomp.fortran/get-mapped-ptr-1.f90 -O execution test
>>
>> Does that one need similar treatment?
>
> I assume not just that but libgomp.c-c++-common/get-mapped-ptr-1.c too?
No, that one I've not seen FAIL, tested in several different
configurations/systems/GPUs. Maybe we were just lucky -- or would there
be a more fundamental problem, when it was in fact "to be expected" that
it'd FAIL, but didn't?
> It both needs the same treatment, and in the get-mapped-ptr-1.c
> case there is even UB, while the Fortran version was using c_loc (q)
> as the host pointer, in C/C++ it was using q which was value of
> uninitialized pointer.
>
> Tested on x86_64-linux, committed to trunk.
ACK, thanks.
Grüße
Thomas
> 2022-06-15 Jakub Jelinek <jakub@redhat.com>
>
> * testsuite/libgomp.c-c++-common/get-mapped-ptr-1.c (main): Initialize
> q to ddress of an automatic variable. Use -5 instead of -1 in
> omp_get_mapped_ptr call. Add test with omp_initial_device.
> * testsuite/libgomp.fortran/get-mapped-ptr-1.f90 (main): Use -5 instead
> of -1 in omp_get_mapped_ptr call. Add test with omp_initial_device.
> Renumber stop arguments afterwards.
>
> --- libgomp/testsuite/libgomp.c-c++-common/get-mapped-ptr-1.c.jj 2022-05-09 09:09:20.963464303 +0200
> +++ libgomp/testsuite/libgomp.c-c++-common/get-mapped-ptr-1.c 2022-06-15 09:39:17.518746512 +0200
> @@ -6,7 +6,10 @@ main ()
> {
> int d = omp_get_default_device ();
> int id = omp_get_initial_device ();
> - void *p , *q;
> + int x;
> + void *p, *q;
> +
> + q = (void *) &x;
>
> if (d < 0 || d >= omp_get_num_devices ())
> d = id;
> @@ -18,7 +21,7 @@ main ()
> if (omp_target_associate_ptr (q, p, sizeof (int), 0, d) != 0)
> return 0;
>
> - if (omp_get_mapped_ptr (q, -1) != NULL)
> + if (omp_get_mapped_ptr (q, -5) != NULL)
> abort ();
>
> if (omp_get_mapped_ptr (q, omp_get_num_devices () + 1) != NULL)
> @@ -27,6 +30,9 @@ main ()
> if (omp_get_mapped_ptr (q, id) != q)
> abort ();
>
> + if (omp_get_mapped_ptr (q, omp_initial_device) != q)
> + abort ();
> +
> if (omp_get_mapped_ptr (q, d) != p)
> abort ();
>
> --- libgomp/testsuite/libgomp.fortran/get-mapped-ptr-1.f90.jj 2022-05-09 09:09:20.963464303 +0200
> +++ libgomp/testsuite/libgomp.fortran/get-mapped-ptr-1.f90 2022-06-15 09:43:11.632338293 +0200
> @@ -19,7 +19,7 @@ program main
> if (omp_target_associate_ptr (c_loc (q), p, c_sizeof (q), &
> 0_c_size_t, d) == 0) then
>
> - if(c_associated (omp_get_mapped_ptr (c_loc (q), -1))) &
> + if(c_associated (omp_get_mapped_ptr (c_loc (q), -5))) &
> stop 1
>
> if(c_associated (omp_get_mapped_ptr (c_loc (q), &
> @@ -29,14 +29,18 @@ program main
> if(.not. c_associated (omp_get_mapped_ptr (c_loc (q), id), c_loc (q))) &
> stop 3
>
> - if(.not. c_associated (omp_get_mapped_ptr (c_loc (q), d), p)) &
> + if(.not. c_associated (omp_get_mapped_ptr (c_loc (q), omp_initial_device), &
> + c_loc (q))) &
> stop 4
>
> - if (omp_target_disassociate_ptr (c_loc (q), d) /= 0) &
> + if(.not. c_associated (omp_get_mapped_ptr (c_loc (q), d), p)) &
> stop 5
>
> - if(c_associated (omp_get_mapped_ptr (c_loc (q), d))) &
> + if (omp_target_disassociate_ptr (c_loc (q), d) /= 0) &
> stop 6
> +
> + if(c_associated (omp_get_mapped_ptr (c_loc (q), d))) &
> + stop 7
> end if
>
> call omp_target_free (p, d)
>
>
> Jakub
-----------------
Siemens Electronic Design Automation GmbH; Anschrift: Arnulfstraße 201, 80634 München; Gesellschaft mit beschränkter Haftung; Geschäftsführer: Thomas Heurung, Frank Thürauf; Sitz der Gesellschaft: München; Registergericht München, HRB 106955
prev parent reply other threads:[~2022-06-15 20:58 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-06-13 12:06 Jakub Jelinek
2022-06-14 16:41 ` [committed] openmp: Conforming device numbers and omp_{initial, invalid}_device Thomas Schwinge
2022-06-15 8:46 ` [committed] openmp: Conforming device numbers and omp_{initial,invalid}_device Jakub Jelinek
2022-06-15 20:58 ` Thomas Schwinge [this message]
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=878rpx7ieu.fsf@euler.schwinge.homeip.net \
--to=thomas@codesourcery.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=jakub@redhat.com \
--cc=marcel@codesourcery.com \
--cc=tobias@codesourcery.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).