summaryrefslogtreecommitdiff
path: root/mm/percpu.c
diff options
context:
space:
mode:
authorDennis Zhou (Facebook) <dennisszhou@gmail.com>2017-07-15 22:23:09 -0400
committerTejun Heo <tj@kernel.org>2017-07-17 10:53:59 -0400
commit9c01516278ef872190fdda95aafaa8effdf6649a (patch)
treebb7ed3295fc4738d1b7860d1c827639cd4099504 /mm/percpu.c
parent6b9b6f39946c4a7a98fac31ee91bac43cd2b73a9 (diff)
percpu: update the header comment and pcpu_build_alloc_info comments
The header comment for percpu memory is a little hard to parse and is not super clear about how the first chunk is managed. This adds a little more clarity to the situation. There is also quite a bit of tricky logic in the pcpu_build_alloc_info. This adds a restructure of a comment to add a little more information. Unfortunately, you will still have to piece together a handful of other comments too, but should help direct you to the meaningful comments. Signed-off-by: Dennis Zhou <dennisszhou@gmail.com> Signed-off-by: Tejun Heo <tj@kernel.org>
Diffstat (limited to 'mm/percpu.c')
-rw-r--r--mm/percpu.c58
1 files changed, 32 insertions, 26 deletions
diff --git a/mm/percpu.c b/mm/percpu.c
index 9ec5fd498f1d..29244fb076ba 100644
--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -4,36 +4,35 @@
* Copyright (C) 2009 SUSE Linux Products GmbH
* Copyright (C) 2009 Tejun Heo <tj@kernel.org>
*
- * This file is released under the GPLv2.
+ * This file is released under the GPLv2 license.
*
- * This is percpu allocator which can handle both static and dynamic
- * areas. Percpu areas are allocated in chunks. Each chunk is
- * consisted of boot-time determined number of units and the first
- * chunk is used for static percpu variables in the kernel image
- * (special boot time alloc/init handling necessary as these areas
- * need to be brought up before allocation services are running).
- * Unit grows as necessary and all units grow or shrink in unison.
- * When a chunk is filled up, another chunk is allocated.
+ * The percpu allocator handles both static and dynamic areas. Percpu
+ * areas are allocated in chunks which are divided into units. There is
+ * a 1-to-1 mapping for units to possible cpus. These units are grouped
+ * based on NUMA properties of the machine.
*
* c0 c1 c2
* ------------------- ------------------- ------------
* | u0 | u1 | u2 | u3 | | u0 | u1 | u2 | u3 | | u0 | u1 | u
* ------------------- ...... ------------------- .... ------------
*
- * Allocation is done in offset-size areas of single unit space. Ie,
- * an area of 512 bytes at 6k in c1 occupies 512 bytes at 6k of c1:u0,
- * c1:u1, c1:u2 and c1:u3. On UMA, units corresponds directly to
- * cpus. On NUMA, the mapping can be non-linear and even sparse.
- * Percpu access can be done by configuring percpu base registers
- * according to cpu to unit mapping and pcpu_unit_size.
- *
- * There are usually many small percpu allocations many of them being
- * as small as 4 bytes. The allocator organizes chunks into lists
- * according to free size and tries to allocate from the fullest one.
- * Each chunk keeps the maximum contiguous area size hint which is
- * guaranteed to be equal to or larger than the maximum contiguous
- * area in the chunk. This helps the allocator not to iterate the
- * chunk maps unnecessarily.
+ * Allocation is done by offsets into a unit's address space. Ie., an
+ * area of 512 bytes at 6k in c1 occupies 512 bytes at 6k in c1:u0,
+ * c1:u1, c1:u2, etc. On NUMA machines, the mapping may be non-linear
+ * and even sparse. Access is handled by configuring percpu base
+ * registers according to the cpu to unit mappings and offsetting the
+ * base address using pcpu_unit_size.
+ *
+ * There is special consideration for the first chunk which must handle
+ * the static percpu variables in the kernel image as allocation services
+ * are not online yet. In short, the first chunk is structure like so:
+ *
+ * <Static | [Reserved] | Dynamic>
+ *
+ * The static data is copied from the original section managed by the
+ * linker. The reserved section, if non-zero, primarily manages static
+ * percpu variables from kernel modules. Finally, the dynamic section
+ * takes care of normal allocations.
*
* Allocation state in each chunk is kept using an array of integers
* on chunk->map. A positive value in the map represents a free
@@ -43,6 +42,12 @@
* Chunks can be determined from the address using the index field
* in the page struct. The index field contains a pointer to the chunk.
*
+ * These chunks are organized into lists according to free_size and
+ * tries to allocate from the fullest chunk first. Each chunk maintains
+ * a maximum contiguous area size hint which is guaranteed to be equal
+ * to or larger than the maximum contiguous area in the chunk. This
+ * helps prevent the allocator from iterating over chunks unnecessarily.
+ *
* To use this allocator, arch code should do the following:
*
* - define __addr_to_pcpu_ptr() and __pcpu_ptr_to_addr() to translate
@@ -1842,6 +1847,7 @@ static struct pcpu_alloc_info * __init pcpu_build_alloc_info(
*/
min_unit_size = max_t(size_t, size_sum, PCPU_MIN_UNIT_SIZE);
+ /* determine the maximum # of units that can fit in an allocation */
alloc_size = roundup(min_unit_size, atom_size);
upa = alloc_size / min_unit_size;
while (alloc_size % upa || (offset_in_page(alloc_size / upa)))
@@ -1868,9 +1874,9 @@ static struct pcpu_alloc_info * __init pcpu_build_alloc_info(
}
/*
- * Expand unit size until address space usage goes over 75%
- * and then as much as possible without using more address
- * space.
+ * Wasted space is caused by a ratio imbalance of upa to group_cnt.
+ * Expand the unit_size until we use >= 75% of the units allocated.
+ * Related to atom_size, which could be much larger than the unit_size.
*/
last_allocs = INT_MAX;
for (upa = max_upa; upa; upa--) {