Commit 7a166094 authored by Oleg Nesterov's avatar Oleg Nesterov Committed by Peter Zijlstra
Browse files

uprobes: kill xol_area->slot_count



Add the new helper, xol_get_slot_nr() which does
find_first_zero_bit() + test_and_set_bit().

xol_take_insn_slot() can wait for the "xol_get_slot_nr() < UINSNS_PER_PAGE"
event instead of "area->slot_count < UINSNS_PER_PAGE".

So we can kill area->slot_count and avoid atomic_inc() + atomic_dec(), this
simplifies the code and can slightly improve the performance.

Signed-off-by: default avatarOleg Nesterov <oleg@redhat.com>
Signed-off-by: default avatarPeter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: default avatarAndrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/r/20241001142458.GA13629@redhat.com
parent c16e2fdd
Loading
Loading
Loading
Loading
+15 −14
Original line number Diff line number Diff line
@@ -99,7 +99,6 @@ static LIST_HEAD(delayed_uprobe_list);
 */
struct xol_area {
	wait_queue_head_t 		wq;		/* if all slots are busy */
	atomic_t 			slot_count;	/* number of in-use slots */
	unsigned long 			*bitmap;	/* 0 = free slot */

	struct page			*page;
@@ -1556,7 +1555,6 @@ static struct xol_area *__create_xol_area(unsigned long vaddr)
	init_waitqueue_head(&area->wq);
	/* Reserve the 1st slot for get_trampoline_vaddr() */
	set_bit(0, area->bitmap);
	atomic_set(&area->slot_count, 1);
	insns = arch_uprobe_trampoline(&insns_size);
	arch_uprobe_copy_ixol(area->page, 0, insns, insns_size);

@@ -1629,24 +1627,28 @@ void uprobe_dup_mmap(struct mm_struct *oldmm, struct mm_struct *newmm)
	}
}

/*
 *  - search for a free slot.
 */
static unsigned long xol_take_insn_slot(struct xol_area *area)
static unsigned long xol_get_slot_nr(struct xol_area *area)
{
	unsigned int slot_nr;
	unsigned long slot_nr;

	for (;;) {
	slot_nr = find_first_zero_bit(area->bitmap, UINSNS_PER_PAGE);
	if (slot_nr < UINSNS_PER_PAGE) {
		if (!test_and_set_bit(slot_nr, area->bitmap))
				break;
			continue;
			return slot_nr;
	}
		wait_event(area->wq, (atomic_read(&area->slot_count) < UINSNS_PER_PAGE));

	return UINSNS_PER_PAGE;
}

	atomic_inc(&area->slot_count);
/*
 *  - search for a free slot.
 */
static unsigned long xol_take_insn_slot(struct xol_area *area)
{
	unsigned long slot_nr;

	wait_event(area->wq, (slot_nr = xol_get_slot_nr(area)) < UINSNS_PER_PAGE);

	return area->vaddr + slot_nr * UPROBE_XOL_SLOT_BYTES;
}

@@ -1682,7 +1684,6 @@ static void xol_free_insn_slot(struct uprobe_task *utask)

	slot_nr = offset / UPROBE_XOL_SLOT_BYTES;
	clear_bit(slot_nr, area->bitmap);
	atomic_dec(&area->slot_count);
	smp_mb__after_atomic(); /* pairs with prepare_to_wait() */
	if (waitqueue_active(&area->wq))
		wake_up(&area->wq);