On Thursday 10 March 2005 09:12, Thomas Gleixner wrote:
On Wed, 2005-03-09 at 20:52 +0100, Blaisorblade wrote:
quoted
quoted
Are you sure this is really the best option in this instance?
Sometimes, static data initialisation is more efficient than
code-based manual initialisation, especially when the memory
is written to anyway.
Agreed, theoretically, but this was done for multiple reasons globally,
for instance as a preparation to Ingo Molnar's preemption patches. There
was mention of this on lwn.net about this:
http://lwn.net/Articles/108719/
Those patches did only the conversion of
static spinlock_t lock = SPIN_LOCK_UNLOCKED;
lock = SPIN_LOCK_UNLOCKED;
to
static DEFINE_SPINLOCK(lock);
spin_lock_init(lock);
First: I didn't write the patch, only forwarded it, so I just guessed why it
was done.
The latter is spin_lock_init(&lock); (since someone got confused about this).
However, this is a .lock = SPIN_LOCK_UNLOCKED ->
spin_lock_init(&struct.lock),
so I don't understand what changes... the structure is initialized inside a
function, so there's no change.
If you want to do static initialization inside of structures, then you
have to define a seperate MACRO similar to the static initialization of
list_head's inside of structures:
static struct sysfs_dirent sysfs_root = {
.s_sibling = LIST_HEAD_INIT(sysfs_root.s_sibling),
I don't see the need here... and the initialization is not in static code; it
changes this code snippet:
void *port_data(int port_num)
{
//...
*port = ((struct port_list)
{ .list = LIST_HEAD_INIT(port->list),
.has_connection = 0,
.sem = __SEMAPHORE_INITIALIZER(port->sem,
0),
.lock = SPIN_LOCK_UNLOCKED,
.port = port_num,
.fd = fd,
.pending = LIST_HEAD_INIT(port->pending),
.connections =
LIST_HEAD_INIT(port->connections) });
So you are all doing some confusion (in fact I guess Andrew realized this when
he merged this anyway).
--
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729
http://www.user-mode-linux.org/~blaisorblade