@lplewa It fixes #5597 and not #5598

Reviewed 2 of 2 files at r1, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @lplewa)

-- commits line 4 at r1:
#5597

Code quote:

#5598

Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @lplewa)

src/libpmem2/map_posix.c line 2 at r1 (raw file):

// SPDX-License-Identifier: BSD-3-Clause
/* Copyright 2019-2021, Intel Corporation */

Suggestion:

2023

src/test/unittest/ut_file.c line 2 at r1 (raw file):

// SPDX-License-Identifier: BSD-3-Clause
/* Copyright 2014-2020, Intel Corporation */

Suggestion:

2014-2023

Reviewed 1 of 2 files at r1.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @lplewa)

Reviewed 1 of 2 files at r1, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @lplewa)

Reviewed 2 of 2 files at r2, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @lplewa)

@lplewa please rebase to stable-1.13

Reviewed 2 of 2 files at r2, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @lplewa)

Reviewed 2 of 2 files at r2, all commit messages.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @lplewa)

a discussion (no related file):
Maybe it is just me, but after applying these changes I still have one remaining problem:

$ ./RUNTESTS.py pmem2_integration -u 39 -g byte --force-enable pmemcheck -b debug
pmem2_integration/TEST39: SETUP (medium/debug/pmemcheck/byte)
Last 16 lines of /home/jmichal/work/pmdk/src/test/pmem2_integration/pmemcheck39.log below (whole file has 16 lines):
==1998050== pmemcheck-1.0, a simple persistent store checker
==1998050== Copyright (c) 2014-2020, Intel Corporation
==1998050== Using Valgrind-3.20.0 and LibVEX; rerun with -h for copyright info
==1998050== Command: /home/jmichal/work/pmdk/src/test/pmem2_integration/pmem2_integration test_map_from_existing /tmp/pmem2_integration_39/testfile
==1998050== Parent PID: 1998038
==1998050== 
==1998050== 
==1998050== Number of stores not made persistent: 1
==1998050== Stores not made persistent properly:
==1998050== [0]    at 0x401481A: _dl_fixup (dl-machine.h:239)
==1998050==    by 0x4001C3D: _dl_runtime_resolve_xsave (dl-trampoline.h:126)
==1998050==    by 0x402883: TEST_CASE_PROCESS (unittest.h:700)
==1998050==    by 0x405A7A: main (pmem2_integration.c:1030)
==1998050==     Address: 0x60f0c8       size: 8 state: DIRTY
==1998050== Total memory not made persistent: 8
==1998050== ERROR SUMMARY: 1 errors

src/libpmem2/map_posix.c line 593 at r2 (raw file):

		return ret;

	VALGRIND_REMOVE_PMEM_MAPPING(map_addr, map_len);

What if pmem2_map_from_existing() is called with src->type != PMEM2_SOURCE_FD so, in effect, the new mapping is not registered to Valgrind? So, I guess, it boils down to whether it is possible that the mapping is not registered (and I think it is possible)? And how Valgrind will react in this case?

src/test/unittest/ut_file.c line 375 at r2 (raw file):

	addr = ut_mmap(file, line, func, NULL, size, PROT_READ | PROT_WRITE,
		MAP_SHARED, fd, 0);
	return addr;

addr here is the most ignored variable I have seen for quite some time. xD
What is the purpose of existence? You may ask.

Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @lplewa)

Reviewable status: 0 of 2 files reviewed, 2 unresolved discussions (waiting on @grom72, @janekmi, and @ldorau)

src/libpmem2/map_posix.c line 593 at r2 (raw file):

Done.

src/test/unittest/ut_file.c line 375 at r2 (raw file):

Done

Reviewed 2 of 2 files at r3, all commit messages.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @janekmi and @lplewa)

Reviewed 2 of 2 files at r3, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @janekmi and @lplewa)

src/libpmem2/map_posix.c line 595 at r3 (raw file):

#ifndef _WIN32
	if (map->source.type == PMEM2_SOURCE_FD) {
		 VALGRIND_REMOVE_PMEM_MAPPING(map_addr, map_len);

Remove space before VALGRIND_REMOVE_PMEM_MAPPING

Reviewed 1 of 2 files at r3, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @lplewa)

Reviewed 1 of 2 files at r3.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @lplewa)

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @lplewa)

a discussion (no related file):

A funny thing. @grom72 confirmed this fix works for him. So, the score is 2:1 for this commit being the fix.
I guess I am doing something atypical which causes this uncommon issue.

I think we can ignore it for the discussion around this commit.

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @lplewa)

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @lplewa)

a discussion (no related file):

The score is 3:1 ;-) - both:

$ ./RUNTESTS.py pmem2_integration -u 39 --force-enable pmemcheck
$ ./RUNTESTS.py pmem2_integration -u 41 --force-enable pmemcheck

pass for me.