Some \n are missing in linux mm/. They may make some difference as
messages are printed when they are full (to syslog), and -with this
messages- it can be too late...
Pavel
PS: I have system where swap is order of magnitude faster than root. I
tried to convince linux to put even code of executables into swap. I
replaced check for pte_dirty() with 1 in try_to_swap_out() - that
should make all pages 'dirty' and put them into swap (should I
actually set that bit for them?), then I killed 'replacing page-cached
entry on page ...' message from swap_state.c. It does not work. What
else should I do? Is this doable?
--- clean/mm/swap_state.c Sun Mar 8 12:50:08 1998
+++ linux/mm/swap_state.c Mon Mar 9 18:35:00 1998
@@ -64,14 +64,14 @@
#endif
if (PageTestandSetSwapCache(page)) {
printk("swap_cache: replacing non-empty entry %08lx "
- "on page %08lx",
+ "on page %08lx\n",
page->offset, page_address(page));
return 0;
}
if (page->inode) {
printk("swap_cache: replacing page-cached entry "
- "on page %08lx", page_address(page));
+ "on page %08lx\n", page_address(page));
return 0;
}
atomic_inc(&page->count);
page->inode = &swapper_inode;
@@ -138,18 +140,18 @@
{
if (!page->inode) {
printk ("VM: Removing swap cache page with zero inode hash "
- "on page %08lx", page_address(page));
+ "on page %08lx\n", page_address(page));
return;
}
if (page->inode != &swapper_inode) {
printk ("VM: Removing swap cache page with wrong inode hash "
- "on page %08lx", page_address(page));
+ "on page %08lx\n", page_address(page));
}
/*
* This will be a legal case once we have a more mature swap cache.
*/
if (atomic_read(&page->count) == 1) {
- printk ("VM: Removing page cache on unshared page %08lx",
+ printk ("VM: Removing page cache on unshared page %08lx\n",
page_address(page));
return;
}
-- I'm really pavel@atrey.karlin.mff.cuni.cz. Pavel Look at http://atrey.karlin.mff.cuni.cz/~pavel/ ;-).- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu