block: Refactor get_tmp_filename()
commit69fbfff95e849156985cf95e2010ffc8762e34e6
authorBin Meng <bin.meng@windriver.com>
Mon, 10 Oct 2022 04:04:31 +0000 (10 12:04 +0800)
committerKevin Wolf <kwolf@redhat.com>
Thu, 27 Oct 2022 18:13:32 +0000 (27 20:13 +0200)
tree1517329e87499effe15c69e372cff9b7efdb05f3
parent6b6471eee11dce4c995419b68441c6637be3d90a
block: Refactor get_tmp_filename()

At present there are two callers of get_tmp_filename() and they are
inconsistent.

One does:

    /* TODO: extra byte is a hack to ensure MAX_PATH space on Windows. */
    char *tmp_filename = g_malloc0(PATH_MAX + 1);
    ...
    ret = get_tmp_filename(tmp_filename, PATH_MAX + 1);

while the other does:

    s->qcow_filename = g_malloc(PATH_MAX);
    ret = get_tmp_filename(s->qcow_filename, PATH_MAX);

As we can see different 'size' arguments are passed. There are also
platform specific implementations inside the function, and the use
of snprintf is really undesirable.

The function name is also misleading. It creates a temporary file,
not just a filename.

Refactor this routine by changing its name and signature to:

    char *create_tmp_file(Error **errp)

and use g_get_tmp_dir() / g_mkstemp() for a consistent implementation.

While we are here, add some comments to mention that /var/tmp is
preferred over /tmp on non-win32 hosts.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Message-Id: <20221010040432.3380478-2-bin.meng@windriver.com>
[kwolf: Fixed incorrect errno negation and iotest 051]
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
block.c
block/vvfat.c
include/block/block_int-common.h
tests/qemu-iotests/051
tests/qemu-iotests/051.out
tests/qemu-iotests/051.pc.out