[PATCH v4 0/2] Generic multi-page exclusion

[PATCH v4 0/2] Generic multi-page exclusion

[Petr Tesarik]

Kumagai-san,

this patch was inspired by this post of yours:

http://lists.infradead.org/pipermail/kexec/2013-November/010445.html

This is a preparatory series to add hugepage support without having
to care about the appropriate size of the cyclic buffer.

Changelog:
* v2:
  - Keep excluded regions per mem_map_data
  - Process excluded pages in chunks

* v3:
  - Keep excluded region extents in struct cycle

* v4:
  - Make sure that the excluded region extents are reset when
    a new ELF segment is processed.

Petr Tesarik (2):
  Generic handling of multi-page exclusions
  Get rid of overrun adjustments

 makedumpfile.c | 111 ++++++++++++++++++++-------------------------------------
 makedumpfile.h |   5 +++
 2 files changed, 43 insertions(+), 73 deletions(-)

-- 
1.8.4.5


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

[PATCH v4 1/2] Generic handling of multi-page exclusions

[Petr Tesarik]

When multiple pages are excluded from the dump, store the extents in
struct cycle and check if anything is still pending on the next invocation
of __exclude_unnecessary_pages. This assumes that:

  1. after __exclude_unnecessary_pages is called for a struct mem_map_data
     that extends beyond the current cycle, it is not called again during
     that cycle,
  2. in the next cycle, __exclude_unnecessary_pages is not called before
     this final struct mem_map_data.

Both assumptions are met if struct mem_map_data segments:

  1. do not overlap,
  2. are sorted by physical address in ascending order.

These two conditions are true for all supported memory models.

Note that the start PFN of the excluded extent is set to the end of the
current cycle (which is equal to the start of the next cycle, see
update_cycle), so only the part of the excluded region which falls beyond
current cycle buffer is valid. If the excluded region is completely
processed in the current cycle, the start PFN is bigger than the end PFN
and no work is done at the beginning of the next cycle.

After processing the leftover from last cycle, pfn_start and mem_map are
adjusted to skip the excluded pages. There is no check whether the
adjusted pfn_start is within the current cycle. Nothing bad happens if
it isn't, because pages outside the current cyclic region are ignored by
the subsequent loop, and the remainder is postponed to the next cycle by
exclude_range().

Signed-off-by: Petr Tesarik <ptesarik@suse.cz>
---
 makedumpfile.c | 52 ++++++++++++++++++++++++++++++++++++++--------------
 makedumpfile.h |  5 +++++
 2 files changed, 43 insertions(+), 14 deletions(-)

diff --git a/makedumpfile.c b/makedumpfile.c
index edfe30b..ab5b862 100644
--- a/makedumpfile.c
+++ b/makedumpfile.c
@@ -44,6 +44,9 @@ static void first_cycle(mdf_pfn_t start, mdf_pfn_t max, struct cycle *cycle)
 
 	if (cycle->end_pfn > max)
 		cycle->end_pfn = max;
+
+	cycle->exclude_pfn_start = 0;
+	cycle->exclude_pfn_end = 0;
 }
 
 static void update_cycle(mdf_pfn_t max, struct cycle *cycle)
@@ -4686,6 +4689,26 @@ initialize_2nd_bitmap_cyclic(struct cycle *cycle)
 	return TRUE;
 }
 
+static void
+exclude_range(mdf_pfn_t *counter, mdf_pfn_t pfn, mdf_pfn_t endpfn,
+    struct cycle *cycle)
+{
+	if (cycle) {
+		cycle->exclude_pfn_start = cycle->end_pfn;
+		cycle->exclude_pfn_end = endpfn;
+		cycle->exclude_pfn_counter = counter;
+
+		if (cycle->end_pfn < endpfn)
+			endpfn = cycle->end_pfn;
+	}
+
+	while (pfn < endpfn) {
+		if (clear_bit_on_2nd_bitmap_for_kernel(pfn, cycle))
+			(*counter)++;
+		++pfn;
+	}
+}
+
 int
 __exclude_unnecessary_pages(unsigned long mem_map,
     mdf_pfn_t pfn_start, mdf_pfn_t pfn_end, struct cycle *cycle)
@@ -4700,6 +4723,18 @@ __exclude_unnecessary_pages(unsigned long mem_map,
 	unsigned long flags, mapping, private = 0;
 
 	/*
+	 * If a multi-page exclusion is pending, do it first
+	 */
+	if (cycle && cycle->exclude_pfn_start < cycle->exclude_pfn_end) {
+		exclude_range(cycle->exclude_pfn_counter,
+			cycle->exclude_pfn_start, cycle->exclude_pfn_end,
+			cycle);
+
+		mem_map += (cycle->exclude_pfn_end - pfn_start) * SIZE(page);
+		pfn_start = cycle->exclude_pfn_end;
+	}
+
+	/*
 	 * Refresh the buffer of struct page, when changing mem_map.
 	 */
 	pfn_read_start = ULONGLONG_MAX;
@@ -4763,21 +4798,10 @@ __exclude_unnecessary_pages(unsigned long mem_map,
 		if ((info->dump_level & DL_EXCLUDE_FREE)
 		    && info->page_is_buddy
 		    && info->page_is_buddy(flags, _mapcount, private, _count)) {
-			int i, nr_pages = 1 << private;
+			int nr_pages = 1 << private;
+
+			exclude_range(&pfn_free, pfn, pfn + nr_pages, cycle);
 
-			for (i = 0; i < nr_pages; ++i) {
-				/*
-				 * According to combination of
-				 * MAX_ORDER and size of cyclic
-				 * buffer, this clearing bit operation
-				 * can overrun the cyclic buffer.
-				 *
-				 * See check_cyclic_buffer_overrun()
-				 * for the detail.
-				 */
-				if (clear_bit_on_2nd_bitmap_for_kernel((pfn + i), cycle))
-					pfn_free++;
-			}
 			pfn += nr_pages - 1;
 			mem_map += (nr_pages - 1) * SIZE(page);
 		}
diff --git a/makedumpfile.h b/makedumpfile.h
index 7acb23a..9402f05 100644
--- a/makedumpfile.h
+++ b/makedumpfile.h
@@ -1595,6 +1595,11 @@ int get_xen_info_ia64(void);
 struct cycle {
 	mdf_pfn_t start_pfn;
 	mdf_pfn_t end_pfn;
+
+	/* for excluding multi-page regions */
+	mdf_pfn_t exclude_pfn_start;
+	mdf_pfn_t exclude_pfn_end;
+	mdf_pfn_t *exclude_pfn_counter;
 };
 
 static inline int
-- 
1.8.4.5


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

[PATCH v4 2/2] Get rid of overrun adjustments

[Petr Tesarik]

Thanks to the previous commit, __exclude_unnecessary_pages does not
require any specific size of the cycle.

Signed-off-by: Petr Tesarik <ptesarik@suse.cz>
---
 makedumpfile.c | 59 ----------------------------------------------------------
 1 file changed, 59 deletions(-)

diff --git a/makedumpfile.c b/makedumpfile.c
index ab5b862..9f047f8 100644
--- a/makedumpfile.c
+++ b/makedumpfile.c
@@ -91,7 +91,6 @@ do { \
 		*ptr_long_table = value; \
 } while (0)
 
-static void check_cyclic_buffer_overrun(void);
 static void setup_page_is_buddy(void);
 
 void
@@ -3276,9 +3275,6 @@ out:
 		    !sadump_generate_elf_note_from_dumpfile())
 			return FALSE;
 
-		if (info->flag_cyclic && info->dump_level & DL_EXCLUDE_FREE)
-			check_cyclic_buffer_overrun();
-
 	} else {
 		if (!get_mem_map_without_mm())
 			return FALSE;
@@ -4302,61 +4298,6 @@ exclude_free_page(struct cycle *cycle)
 }
 
 /*
- * Let C be a cyclic buffer size and B a bitmap size used for
- * representing maximum block size managed by buddy allocator.
- *
- * For some combinations of C and B, clearing operation can overrun
- * the cyclic buffer. Let's consider three cases.
- *
- *   - If C == B, this is trivially safe.
- *
- *   - If B > C, overrun can easily happen.
- *
- *   - In case of C > B, if C mod B != 0, then there exist n > m > 0,
- *     B > b > 0 such that n x C = m x B + b. This means that clearing
- *     operation overruns cyclic buffer (B - b)-bytes in the
- *     combination of n-th cycle and m-th block.
- *
- *     Note that C mod B != 0 iff (m x C) mod B != 0 for some m.
- *
- * If C == B, C mod B == 0 always holds. Again, if B > C, C mod B != 0
- * always holds. Hence, it's always sufficient to check the condition
- * C mod B != 0 in order to determine whether overrun can happen or
- * not.
- *
- * The bitmap size used for maximum block size B is calculated from
- * MAX_ORDER as:
- *
- *   B := DIVIDE_UP((1 << (MAX_ORDER - 1)), BITS_PER_BYTE)
- *
- * Normally, MAX_ORDER is 11 at default. This is configurable through
- * CONFIG_FORCE_MAX_ZONEORDER.
- */
-static void
-check_cyclic_buffer_overrun(void)
-{
-	int max_order = ARRAY_LENGTH(zone.free_area);
-	int max_order_nr_pages = 1 << (max_order - 1);
-	unsigned long max_block_size = divideup(max_order_nr_pages, BITPERBYTE);
-
-	if (info->bufsize_cyclic % max_block_size) {
-		unsigned long bufsize;
-
-		if (max_block_size > info->bufsize_cyclic) {
-			MSG("WARNING: some free pages are not filtered.\n");
-			return;
-		}
-
-		bufsize = info->bufsize_cyclic;
-		info->bufsize_cyclic = round(bufsize, max_block_size);
-		info->pfn_cyclic = info->bufsize_cyclic * BITPERBYTE;
-
-		MSG("cyclic buffer size has been changed: %lu => %lu\n",
-		    bufsize, info->bufsize_cyclic);
-	}
-}
-
-/*
  * For the kernel versions from v2.6.17 to v2.6.37.
  */
 static int
-- 
1.8.4.5


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec

RE: [PATCH v4 0/2] Generic multi-page exclusion

[Atsushi Kumagai]

Hello Petr,

>Kumagai-san,
>
>this patch was inspired by this post of yours:
>
>http://lists.infradead.org/pipermail/kexec/2013-November/010445.html
>
>This is a preparatory series to add hugepage support without having
>to care about the appropriate size of the cyclic buffer.
>
>Changelog:
>* v2:
>  - Keep excluded regions per mem_map_data
>  - Process excluded pages in chunks
>
>* v3:
>  - Keep excluded region extents in struct cycle
>
>* v4:
>  - Make sure that the excluded region extents are reset when
>    a new ELF segment is processed.

This version looks good to me, I'll merge it into v1.5.7.
Thanks for your great work!


Atsushi Kumagai

>Petr Tesarik (2):
>  Generic handling of multi-page exclusions
>  Get rid of overrun adjustments
>
> makedumpfile.c | 111 ++++++++++++++++++++-------------------------------------
> makedumpfile.h |   5 +++
> 2 files changed, 43 insertions(+), 73 deletions(-)
>
>--
>1.8.4.5


_______________________________________________
kexec mailing list
kexec@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/kexec