obj: fix data_store example transaction logic

pmem2: fix munmap when file mapping fails (posix)

Reservation region should be unmapped only when it wasn't provided
by the config, otherwise the job of cleaning up the reservation is
up to the invoking layer.

pmem2: implement vm_reservation API

Change pmem2_map to check if config structure reservation is set.
When the reservation is set, there is no need to create a placeholder
mapping to map a file.

test: fix cstyle issue in py testframework

test: fix usc permission check tool

pmem2: fix config test typo

basetest.py:265:13: E741 ambiguous variable name 'l'

test: fix symbol format in the scope tests

test: set libraries debug logs envs (py)

update valgrind suppressions for new glibc

Both __GI_mempcpy and __GI_memcpy can come here, the latter starting
with glibc 2.35.

fix build with -Og on some platforms

It's not obvious to the compiler that, as there's at least one extent,
these variables will always be initialized.

doc: fix too long line in the libpmem2.7 man

doc: align error section

doc: libpmem2.7 man update

test: update unmap to delete

pmem2: deregister valgrind mapping prior to unmap

pmem2: fix non-page-aligned deep flushes

test: document test framework 

... as well as very minor code cleanup

common: add header for FALLOC_FL_PUNCH_HOLE

This was raported by coverity as use-after-free, but I think
it's unlikely to be a real problem. But the fix is simple enough...

benchmark: fix memset scenario name