Change 2:
split the poll_table structure into 2 parts:
One global part with just a function pointer and a priv variable.
The 2nd part is user specific:
for __pollwait: the poll table and the error variable
for ep_ptable_queue_proc: the pointer to the eventpoll data.
This change is a prerequisite for the next patch, that embedds a few
waitqueues in the __pollwait specific part of the poll structure.
--
Manfred
--------------010601020005090005040509
Content-Type: text/plain;
name="patch-poll-2-funcptr"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="patch-poll-2-funcptr"
// $Header$
// Kernel Version:
// VERSION = 2
// PATCHLEVEL = 5
// SUBLEVEL = 46
// EXTRAVERSION =
--- 2.5/include/linux/poll.h 2002-11-10 12:19:10.000000000 +0100
+++ build-2.5/include/linux/poll.h 2002-11-10 12:18:40.000000000 +0100
@@ -10,44 +10,28 @@
#include <linux/mm.h>
#include <asm/uaccess.h>
-struct poll_table_page;
+typedef struct poll_table_struct poll_table;
-typedef struct poll_table_struct {
+struct poll_table_struct {
+ void (*qproc)(void * priv, struct file * filp, wait_queue_head_t * wait_address);
void *priv;
- void (*qproc)(void *, wait_queue_head_t *);
+};
+
+struct poll_wrapper {
+ poll_table pt;
int error;
struct poll_table_page * table;
-} poll_table;
+};
-extern void __pollwait(struct file * filp, wait_queue_head_t * wait_address, poll_table *p);
+void poll_initwait(struct poll_wrapper* pt);
+void poll_freewait(struct poll_wrapper* pt);
static inline void poll_wait(struct file * filp, wait_queue_head_t * wait_address, poll_table *p)
{
if (p && wait_address)
- __pollwait(filp, wait_address, p);
+ p->qproc(p->priv, filp, wait_address);
}
-static inline void poll_initwait(poll_table* pt)
-{
- pt->qproc = NULL;
- pt->priv = NULL;
- pt->error = 0;
- pt->table = NULL;
-}
-
-static inline void poll_initwait_ex(poll_table* pt,
- void (*qproc)(void *, wait_queue_head_t *),
- void *priv)
-{
- pt->qproc = qproc;
- pt->priv = priv;
- pt->error = 0;
- pt->table = NULL;
-}
-
-extern void poll_freewait(poll_table* pt);
-
-
/*
* Scaleable version of the fd_set.
*/
--- 2.5/fs/select.c 2002-11-10 12:19:10.000000000 +0100
+++ build-2.5/fs/select.c 2002-11-10 12:18:40.000000000 +0100
@@ -54,7 +54,7 @@
* poll table.
*/
-void poll_freewait(poll_table* pt)
+void poll_freewait(struct poll_wrapper* pt)
{
struct poll_table_page * p = pt->table;
while (p) {
@@ -73,15 +73,11 @@
}
}
-void __pollwait(struct file * filp, wait_queue_head_t * wait_address, poll_table *p)
+static void __pollwait(void *priv, struct file * filp, wait_queue_head_t * wait_address)
{
+ struct poll_wrapper *p = priv;
struct poll_table_page *table = p->table;
- if (p->qproc) {
- p->qproc(p->priv, wait_address);
- return;
- }
-
if (!table || POLL_TABLE_FULL(table)) {
struct poll_table_page *new_table;
@@ -109,6 +105,14 @@
}
}
+void poll_initwait(struct poll_wrapper* pt)
+{
+ pt->pt.qproc = __pollwait;
+ pt->pt.priv = pt;
+ pt->error = 0;
+ pt->table = NULL;
+}
+
#define __IN(fds, n) (fds->in + n)
#define __OUT(fds, n) (fds->out + n)
#define __EX(fds, n) (fds->ex + n)
@@ -169,7 +173,8 @@
int do_select(int n, fd_set_bits *fds, long *timeout)
{
- poll_table table, *wait;
+ struct poll_wrapper table;
+ poll_table *wait;
int retval, i, off;
long __timeout = *timeout;
@@ -182,7 +187,7 @@
n = retval;
poll_initwait(&table);
- wait = &table;
+ wait = &table.pt;
if (!__timeout)
wait = NULL;
retval = 0;
@@ -389,10 +394,13 @@
}
static int do_poll(unsigned int nfds, unsigned int nchunks, unsigned int nleft,
- struct pollfd *fds[], poll_table *wait, long timeout)
+ struct pollfd *fds[], struct poll_wrapper *pw, long timeout)
{
int count;
- poll_table* pt = wait;
+ poll_table* pt = &pw->pt;
+
+ if (!timeout)
+ pt = NULL;
for (;;) {
unsigned int i;
@@ -406,7 +414,7 @@
pt = NULL;
if (count || !timeout || signal_pending(current))
break;
- count = wait->error;
+ count = pw->error;
if (count)
break;
timeout = schedule_timeout(timeout);
@@ -419,7 +427,7 @@
{
int i, j, fdcount, err;
struct pollfd **fds;
- poll_table table, *wait;
+ struct poll_wrapper table;
int nchunks, nleft;
/* Do a sanity check on nfds ... */
@@ -435,9 +443,6 @@
}
poll_initwait(&table);
- wait = &table;
- if (!timeout)
- wait = NULL;
err = -ENOMEM;
fds = NULL;
@@ -474,7 +479,7 @@
goto out_fds1;
}
- fdcount = do_poll(nfds, nchunks, nleft, fds, wait, timeout);
+ fdcount = do_poll(nfds, nchunks, nleft, fds, &table, timeout);
/* OK, now copy the revents fields back to user space. */
for(i=0; i < nchunks; i++)
--- 2.5/fs/eventpoll.c 2002-11-10 12:19:10.000000000 +0100
+++ build-2.5/fs/eventpoll.c 2002-11-10 12:18:25.000000000 +0100
@@ -199,7 +199,7 @@
static struct epitem *ep_find(struct eventpoll *ep, struct file *file);
static void ep_use_epitem(struct epitem *dpi);
static void ep_release_epitem(struct epitem *dpi);
-static void ep_ptable_queue_proc(void *priv, wait_queue_head_t *whead);
+static void ep_ptable_queue_proc(void * priv, struct file * filp, wait_queue_head_t * whead);
static int ep_insert(struct eventpoll *ep, struct pollfd *pfd, struct file *tfile);
static unsigned int ep_get_file_events(struct file *file);
static int ep_modify(struct eventpoll *ep, struct epitem *dpi, unsigned int events);
@@ -784,14 +784,19 @@
DPI_MEM_FREE(dpi);
}
+struct epoll_wrapper {
+ poll_table pt;
+ struct epitem *dpi;
+};
/*
* This is the callback that is used to add our wait queue to the
* target file wakeup lists.
*/
-static void ep_ptable_queue_proc(void *priv, wait_queue_head_t *whead)
+static void ep_ptable_queue_proc(void * priv, struct file * filp, wait_queue_head_t * whead)
{
- struct epitem *dpi = priv;
+ struct epoll_wrapper *ew = priv;
+ struct epitem *dpi = ew->dpi;
/* No more than EP_MAX_POLL_QUEUE wait queue are supported */
if (dpi->nwait < EP_MAX_POLL_QUEUE) {
@@ -801,13 +806,20 @@
}
}
+static inline void epoll_table_init(struct epoll_wrapper* ew,
+ struct epitem *dpi)
+{
+ ew->pt.qproc = ep_ptable_queue_proc;
+ ew->pt.priv = ew;
+ ew->dpi = dpi;
+}
static int ep_insert(struct eventpoll *ep, struct pollfd *pfd, struct file *tfile)
{
int error, i, revents;
unsigned long flags;
struct epitem *dpi;
- poll_table pt;
+ struct epoll_wrapper ew;
error = -ENOMEM;
if (!(dpi = DPI_MEM_ALLOC()))
@@ -828,9 +840,8 @@
}
/* Attach the item to the poll hooks */
- poll_initwait_ex(&pt, ep_ptable_queue_proc, dpi);
- revents = tfile->f_op->poll(tfile, &pt);
- poll_freewait(&pt);
+ epoll_table_init(&ew, dpi);
+ revents = tfile->f_op->poll(tfile, &ew.pt);
/* We have to drop the new item inside our item list to keep track of it */
write_lock_irqsave(&ep->lock, flags);
@@ -854,7 +865,8 @@
/*
- * Returns the current events of the given file.
+ * Returns the current events of the given file.
+ * poll_table set to NULL to avoid any poll queue insertion.
*/
static unsigned int ep_get_file_events(struct file *file)
{
--- 2.5/arch/sparc64/solaris/timod.c 2002-11-10 12:19:10.000000000 +0100
+++ build-2.5/arch/sparc64/solaris/timod.c 2002-11-10 12:15:16.000000000 +0100
@@ -651,10 +651,11 @@
SOLD("LISTEN done");
}
if (!(filp->f_flags & O_NONBLOCK)) {
- poll_table wait_table, *wait;
+ struct poll_wrapper wait_table;
+ poll_table *wait;
poll_initwait(&wait_table);
- wait = &wait_table;
+ wait = &wait_table.pt;
for(;;) {
SOLD("loop");
set_current_state(TASK_INTERRUPTIBLE);
--- 2.5/kernel/ksyms.c 2002-11-10 12:19:10.000000000 +0100
+++ build-2.5/kernel/ksyms.c 2002-11-10 12:15:16.000000000 +0100
@@ -276,7 +276,7 @@
EXPORT_SYMBOL(generic_file_llseek);
EXPORT_SYMBOL(remote_llseek);
EXPORT_SYMBOL(no_llseek);
-EXPORT_SYMBOL(__pollwait);
+EXPORT_SYMBOL(poll_initwait);
EXPORT_SYMBOL(poll_freewait);
EXPORT_SYMBOL(ROOT_DEV);
EXPORT_SYMBOL(find_get_page);
--------------010601020005090005040509--
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/