Re: preparing 2.6.10
#include <hallo.h>
* Andres Salomon [Wed, Jan 05 2005, 12:15:51AM]:
> Alright, enough people are bothering me for 2.6.10 that we should probably
> get it into sid. I've committed a bunch of bk backports to the svn 2.6.10
> k-s directory (small, obvious fixes for the most part); I'm going to aim
> for uploading 2.6.10 into NEW either Wed. or Thurs. unless anyone has
> objections.
Wishlist on patches (laptop/power management related):
http://people.ubuntu.com/~fabbione/misrouted-irq.dpatch (not enabled by
default, needs kernel options, solves the extremely nasty hard kernel
crash on resume from APM/ACPI suspend, thanks fabionne)
The IRQ-after-swsusp-awakening patch (attached). Needed to make
some hardware drivers work after resuming from swsusp (Source:
Message-ID: <20050103150505.GA4120@ip68-4-98-123.oc.oc.cox.net>)
The swsusp-speedup-patch (attached) - makes the swsusp in kernel a lot
faster on suspending (source:
Message-ID: <20041225175453.GA10222@elf.ucw.cz>)
Regards,
Eduard.
--
The past tempts us, the present confuses us, the future frightens us.
And our lives slip away, moment by moment, lost in the terrible
in-between. -- Centauri Emperor, "Babylon 5 - The Coming Of Shadows"
swsusp does not suspend and resume *all* devices, including system
devices. This has been the case since at least 2.6.9, if not earlier.
One effect of this is that resuming fails to properly reconfigure
interrupt routers. In 2.6.9 this was obscured by other kernel code,
but in 2.6.10 this often causes post-resume APIC errors and near-total
failure of some PCI devices (e.g. network, sound and USB controllers).
On at least one of my systems, without this patch I also have to "ifdown
eth0;ifup eth0" to get networking to function after resuming, even after
working around the interrupt routing problem mentioned above. With this
patch, networking simply works after a resume, and the ifdown/ifup is
no longer needed.
This patch is against 2.6.10-mm1, although it applies with an offset to
2.6.10-bk4 as well. I have tested it against 2.6.10-mm1 and 2.6.10-bk4,
with and without "noapic", with and without "acpi=off". However, I have
not tested it on a highmem system.
I believe this patch fixes a severe problem in swsusp; I would like to
see this patch (or at least *some* kind of fix for this problem) tested
more widely and committed to mainline before the 2.6.11 release.
Signed-off-by: Barry K. Nathan <barryn@pobox.com>
--- linux-2.6.10-mm1/kernel/power/swsusp.c 2005-01-03 02:16:15.175265255 -0800
+++ linux-2.6.10-mm1-bkn3/kernel/power/swsusp.c 2005-01-03 06:27:07.753344731 -0800
@@ -843,11 +843,22 @@
if ((error = arch_prepare_suspend()))
return error;
local_irq_disable();
+ /* At this point, device_suspend() has been called, but *not*
+ * device_power_down(). We *must* device_power_down() now.
+ * Otherwise, drivers for some devices (e.g. interrupt controllers)
+ * become desynchronized with the actual state of the hardware
+ * at resume time, and evil weirdness ensues.
+ */
+ if ((error = device_power_down(PM_SUSPEND_DISK))) {
+ local_irq_enable();
+ return error;
+ }
save_processor_state();
error = swsusp_arch_suspend();
/* Restore control flow magically appears here */
restore_processor_state();
restore_highmem();
+ device_power_up();
local_irq_enable();
return error;
}
-
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/
Hi!
Some machines are spending minutes of CPU time during suspend in
stupid O(n^2) algorithm. This patch replaces it with O(n) algorithm,
making swsusp usable to some people.
I'd like people to test this. It should probably spend few weeks
in -mm tree to get some beating. OTOH SUSE has variant of this patch
in its kernel.
Signed-off-by: Pavel Machek <pavel@suse.cz>
Pavel
--- clean-cvs/include/linux/page-flags.h 2004-08-24 20:26:54.000000000 +0200
+++ linux-cvs/include/linux/page-flags.h 2004-12-10 22:35:58.000000000 +0100
@@ -74,7 +74,7 @@
#define PG_swapcache 16 /* Swap page: swp_entry_t in private */
#define PG_mappedtodisk 17 /* Has blocks allocated on-disk */
#define PG_reclaim 18 /* To be reclaimed asap */
-
+#define PG_nosave_free 19 /* Page is free and should not be written */
/*
* Global page accounting. One instance per CPU. Only unsigned longs are
@@ -277,6 +277,10 @@
#define ClearPageNosave(page) clear_bit(PG_nosave, &(page)->flags)
#define TestClearPageNosave(page) test_and_clear_bit(PG_nosave, &(page)->flags)
+#define PageNosaveFree(page) test_bit(PG_nosave_free, &(page)->flags)
+#define SetPageNosaveFree(page) set_bit(PG_nosave_free, &(page)->flags)
+#define ClearPageNosaveFree(page) clear_bit(PG_nosave_free, &(page)->flags)
+
#define PageMappedToDisk(page) test_bit(PG_mappedtodisk, &(page)->flags)
#define SetPageMappedToDisk(page) set_bit(PG_mappedtodisk, &(page)->flags)
#define ClearPageMappedToDisk(page) clear_bit(PG_mappedtodisk, &(page)->flags)
--- clean-cvs/include/linux/suspend.h 2004-09-29 22:04:50.000000000 +0200
+++ linux-cvs/include/linux/suspend.h 2004-12-10 22:35:58.000000000 +0100
@@ -31,6 +31,7 @@
/* mm/page_alloc.c */
extern void drain_local_pages(void);
+extern void mark_free_pages(struct zone *zone);
/* kernel/power/swsusp.c */
extern int software_suspend(void);
--- clean-cvs/kernel/power/swsusp.c 2004-12-14 20:43:42.000000000 +0100
+++ linux-cvs/kernel/power/swsusp.c 2004-12-14 20:48:31.000000000 +0100
@@ -74,11 +74,9 @@
/* References to section boundaries */
extern char __nosave_begin, __nosave_end;
-extern int is_head_of_free_region(struct page *);
-
/* Variables to be preserved over suspend */
-int pagedir_order_check;
-int nr_copy_pages_check;
+static int pagedir_order_check;
+static int nr_copy_pages_check;
extern char resume_file[];
static dev_t resume_device;
@@ -426,12 +424,12 @@
static int save_highmem_zone(struct zone *zone)
{
unsigned long zone_pfn;
+ mark_free_pages(zone);
for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn) {
struct page *page;
struct highmem_page *save;
void *kaddr;
unsigned long pfn = zone_pfn + zone->zone_start_pfn;
- int chunk_size;
if (!(pfn%1000))
printk(".");
@@ -448,11 +446,9 @@
printk("highmem reserved page?!\n");
continue;
}
- if ((chunk_size = is_head_of_free_region(page))) {
- pfn += chunk_size - 1;
- zone_pfn += chunk_size - 1;
+ BUG_ON(PageNosave(page));
+ if (PageNosaveFree(page))
continue;
- }
save = kmalloc(sizeof(struct highmem_page), GFP_ATOMIC);
if (!save)
return -ENOMEM;
@@ -524,21 +520,16 @@
* We save a page if it's Reserved, and not in the range of pages
* statically defined as 'unsaveable', or if it isn't reserved, and
* isn't part of a free chunk of pages.
- * If it is part of a free chunk, we update @pfn to point to the last
- * page of the chunk.
*/
static int saveable(struct zone * zone, unsigned long * zone_pfn)
{
unsigned long pfn = *zone_pfn + zone->zone_start_pfn;
- unsigned long chunk_size;
struct page * page;
if (!pfn_valid(pfn))
return 0;
- if (!(pfn%1000))
- printk(".");
page = pfn_to_page(pfn);
BUG_ON(PageReserved(page) && PageNosave(page));
if (PageNosave(page))
@@ -547,10 +538,8 @@
pr_debug("[nosave pfn 0x%lx]", pfn);
return 0;
}
- if ((chunk_size = is_head_of_free_region(page))) {
- *zone_pfn += chunk_size - 1;
+ if (PageNosaveFree(page))
return 0;
- }
return 1;
}
@@ -563,10 +552,11 @@
nr_copy_pages = 0;
for_each_zone(zone) {
- if (!is_highmem(zone)) {
- for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn)
- nr_copy_pages += saveable(zone, &zone_pfn);
- }
+ if (is_highmem(zone))
+ continue;
+ mark_free_pages(zone);
+ for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn)
+ nr_copy_pages += saveable(zone, &zone_pfn);
}
}
@@ -576,52 +566,25 @@
struct zone *zone;
unsigned long zone_pfn;
struct pbe * pbe = pagedir_nosave;
+ int to_copy = nr_copy_pages;
for_each_zone(zone) {
- if (!is_highmem(zone))
- for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn) {
- if (saveable(zone, &zone_pfn)) {
- struct page * page;
- page = pfn_to_page(zone_pfn + zone->zone_start_pfn);
- pbe->orig_address = (long) page_address(page);
- /* copy_page is no usable for copying task structs. */
- memcpy((void *)pbe->address, (void *)pbe->orig_address, PAGE_SIZE);
- pbe++;
- }
- }
- }
-}
-
-
-static void free_suspend_pagedir_zone(struct zone *zone, unsigned long pagedir)
-{
- unsigned long zone_pfn, pagedir_end, pagedir_pfn, pagedir_end_pfn;
- pagedir_end = pagedir + (PAGE_SIZE << pagedir_order);
- pagedir_pfn = __pa(pagedir) >> PAGE_SHIFT;
- pagedir_end_pfn = __pa(pagedir_end) >> PAGE_SHIFT;
- for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn) {
- struct page *page;
- unsigned long pfn = zone_pfn + zone->zone_start_pfn;
- if (!pfn_valid(pfn))
- continue;
- page = pfn_to_page(pfn);
- if (!TestClearPageNosave(page))
- continue;
- else if (pfn >= pagedir_pfn && pfn < pagedir_end_pfn)
+ if (is_highmem(zone))
continue;
- __free_page(page);
- }
-}
-
-void swsusp_free(void)
-{
- unsigned long p = (unsigned long)pagedir_save;
- struct zone *zone;
- for_each_zone(zone) {
- if (!is_highmem(zone))
- free_suspend_pagedir_zone(zone, p);
+ mark_free_pages(zone);
+ for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn) {
+ if (saveable(zone, &zone_pfn)) {
+ struct page * page;
+ page = pfn_to_page(zone_pfn + zone->zone_start_pfn);
+ pbe->orig_address = (long) page_address(page);
+ /* copy_page is not usable for copying task structs. */
+ memcpy((void *)pbe->address, (void *)pbe->orig_address, PAGE_SIZE);
+ pbe++;
+ to_copy--;
+ }
+ }
}
- free_pages(p, pagedir_order);
+ BUG_ON(to_copy);
}
@@ -687,6 +650,24 @@
return 0;
}
+/**
+ * free_image_pages - Free pages allocated for snapshot
+ */
+
+static void free_image_pages(void)
+{
+ struct pbe * p;
+ int i;
+
+ p = pagedir_save;
+ for (i = 0, p = pagedir_save; i < nr_copy_pages; i++, p++) {
+ if (p->address) {
+ ClearPageNosave(virt_to_page(p->address));
+ free_page(p->address);
+ p->address = 0;
+ }
+ }
+}
/**
* alloc_image_pages - Allocate pages for the snapshot.
@@ -700,18 +681,19 @@
for (i = 0, p = pagedir_save; i < nr_copy_pages; i++, p++) {
p->address = get_zeroed_page(GFP_ATOMIC | __GFP_COLD);
- if(!p->address)
- goto Error;
+ if (!p->address)
+ return -ENOMEM;
SetPageNosave(virt_to_page(p->address));
}
return 0;
- Error:
- do {
- if (p->address)
- free_page(p->address);
- p->address = 0;
- } while (p-- > pagedir_save);
- return -ENOMEM;
+}
+
+void swsusp_free(void)
+{
+ BUG_ON(PageNosave(virt_to_page(pagedir_save)));
+ BUG_ON(PageNosaveFree(virt_to_page(pagedir_save)));
+ free_image_pages();
+ free_pages((unsigned long) pagedir_save, pagedir_order);
}
@@ -1113,6 +1087,7 @@
return -EPERM;
}
nr_copy_pages = swsusp_info.image_pages;
+ pagedir_order = get_bitmask_order(SUSPEND_PD_PAGES(nr_copy_pages));
return error;
}
@@ -1179,9 +1154,7 @@
int i, n = swsusp_info.pagedir_pages;
int error = 0;
- pagedir_order = get_bitmask_order(n);
-
- addr =__get_free_pages(GFP_ATOMIC, pagedir_order);
+ addr = __get_free_pages(GFP_ATOMIC, pagedir_order);
if (!addr)
return -ENOMEM;
pagedir_nosave = (struct pbe *)addr;
--- clean-cvs/mm/page_alloc.c 2004-11-19 12:54:54.000000000 +0100
+++ linux-cvs/mm/page_alloc.c 2004-12-10 22:35:59.000000000 +0100
@@ -437,26 +437,30 @@
#endif /* CONFIG_PM || CONFIG_HOTPLUG_CPU */
#ifdef CONFIG_PM
-int is_head_of_free_region(struct page *page)
+
+void mark_free_pages(struct zone *zone)
{
- struct zone *zone = page_zone(page);
- unsigned long flags;
+ unsigned long zone_pfn, flags;
int order;
struct list_head *curr;
- /*
- * Should not matter as we need quiescent system for
- * suspend anyway, but...
- */
+ if (!zone->spanned_pages)
+ return;
+
spin_lock_irqsave(&zone->lock, flags);
+ for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn)
+ ClearPageNosaveFree(pfn_to_page(zone_pfn + zone->zone_start_pfn));
+
for (order = MAX_ORDER - 1; order >= 0; --order)
- list_for_each(curr, &zone->free_area[order].free_list)
- if (page == list_entry(curr, struct page, lru)) {
- spin_unlock_irqrestore(&zone->lock, flags);
- return 1 << order;
- }
+ list_for_each(curr, &zone->free_area[order].free_list) {
+ unsigned long start_pfn, i;
+
+ start_pfn = page_to_pfn(list_entry(curr, struct page, lru));
+
+ for (i=0; i < (1<<order); i++)
+ SetPageNosaveFree(pfn_to_page(start_pfn+i));
+ }
spin_unlock_irqrestore(&zone->lock, flags);
- return 0;
}
/*
--
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!
-
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/
Reply to: