The genalloc code uses the bitmap API from include/linux/bitmap.h and
lib/bitmap.c, which is based on long values. Both bitmap_set from
lib/bitmap.c and bitmap_set_ll, which is the lockless version from
genalloc.c, use BITMAP_LAST_WORD_MASK to set the first bits in a long in
the bitmap.
That one uses (1 << bits) - 1, 0b111, if you are setting the first three
bits. This means that the API counts from the least significant bits
(LSB from now on) to the MSB. The LSB in the first long is bit 0, then.
The same works for the lookup functions.
The genalloc code uses longs for the bitmap, as it should. In
include/linux/genalloc.h, struct gen_pool_chunk has unsigned long
bits[0] as its last member. When allocating the struct, genalloc should
reserve enough space for the bitmap. This should be a proper number of
longs that can fit the amount of bits in the bitmap.
However, genalloc allocates an integer number of bytes that fit the
amount of bits, but may not be an integer amount of longs. 9 bytes, for
example, could be allocated for 70 bits.
This is a problem in itself if the Least Significat Bit in a long is in
the byte with the largest address, which happens in Big Endian machines.
This means genalloc is not allocating the byte in which it will try to
set or check for a bit.
This may end up in memory corruption, where genalloc will try to set the
bits it has not allocated. In fact, genalloc may not set these bits
because it may find them already set, because they were not zeroed since
they were not allocated. And that's what causes a BUG when
gen_pool_destroy is called and check for any set bits.
What really happens is that genalloc uses kmalloc_node with __GFP_ZERO
on gen_pool_add_virt. With SLAB and SLUB, this means the whole slab
will be cleared, not only the requested bytes. Since struct
gen_pool_chunk has a size that is a multiple of 8, and slab sizes are
multiples of 8, we get lucky and allocate and clear the right amount of
bytes.
Hower, this is not the case with SLOB or with older code that did memset
after allocating instead of using __GFP_ZERO.
So, a simple module as this (running 3.6.0), will cause a crash when
rmmod'ed.
[root@phantom-lp2 foo]# cat foo.c
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/genalloc.h>
MODULE_LICENSE("GPL");
MODULE_VERSION("0.1");
static struct gen_pool *foo_pool;
static __init int foo_init(void)
{
int ret;
foo_pool = gen_pool_create(10, -1);
if (!foo_pool)
return -ENOMEM;
ret = gen_pool_add(foo_pool, 0xa0000000, 32 << 10, -1);
if (ret) {
gen_pool_destroy(foo_pool);
return ret;
}
return 0;
}
static __exit void foo_exit(void)
{
gen_pool_destroy(foo_pool);
}
module_init(foo_init);
module_exit(foo_exit);
[root@phantom-lp2 foo]# zcat /proc/config.gz | grep SLOB
CONFIG_SLOB=y
[root@phantom-lp2 foo]# insmod ./foo.ko
[root@phantom-lp2 foo]# rmmod foo
------------[ cut here ]------------
kernel BUG at lib/genalloc.c:243!
cpu 0x4: Vector: 700 (Program Check) at [c0000000bb0e7960]
pc: c0000000003cb50c: .gen_pool_destroy+0xac/0x110
lr: c0000000003cb4fc: .gen_pool_destroy+0x9c/0x110
sp: c0000000bb0e7be0
msr: 8000000000029032
current = 0xc0000000bb0e0000
paca = 0xc000000006d30e00 softe: 0 irq_happened: 0x01
pid = 13044, comm = rmmod
kernel BUG at lib/genalloc.c:243!
[c0000000bb0e7ca0] d000000004b00020 .foo_exit+0x20/0x38 [foo]
[c0000000bb0e7d20] c0000000000dff98 .SyS_delete_module+0x1a8/0x290
[c0000000bb0e7e30] c0000000000097d4 syscall_exit+0x0/0x94
--- Exception: c00 (System Call) at 000000800753d1a0
SP (fffd0b0e640) is in userspace
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Benjamin Gaignard <benjamin.gaignard@stericsson.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Premit use of another algorithm than the default first-fit one. For
example a custom algorithm could be used to manage alignment requirements.
As I can't predict all the possible requirements/needs for all allocation
uses cases, I add a "free" field 'void *data' to pass any needed
information to the allocation function. For example 'data' could be used
to handle a structure where you store the alignment, the expected memory
bank, the requester device, or any information that could influence the
allocation algorithm.
An usage example may look like this:
struct my_pool_constraints {
int align;
int bank;
...
};
unsigned long my_custom_algo(unsigned long *map, unsigned long size,
unsigned long start, unsigned int nr, void *data)
{
struct my_pool_constraints *constraints = data;
...
deal with allocation contraints
...
return the index in bitmap where perform the allocation
}
void create_my_pool()
{
struct my_pool_constraints c;
struct gen_pool *pool = gen_pool_create(...);
gen_pool_add(pool, ...);
gen_pool_set_algo(pool, my_custom_algo, &c);
}
Add of best-fit algorithm function:
most of the time best-fit is slower then first-fit but memory fragmentation
is lower. The random buffer allocation/free tests don't show any arithmetic
relation between the allocation time and fragmentation but the
best-fit algorithm
is sometime able to perform the allocation when the first-fit can't.
This new algorithm help to remove static allocations on ESRAM, a small but
fast on-chip RAM of few KB, used for high-performance uses cases like DMA
linked lists, graphic accelerators, encoders/decoders. On the Ux500
(in the ARM tree) we have define 5 ESRAM banks of 128 KB each and use of
static allocations becomes unmaintainable:
cd arch/arm/mach-ux500 && grep -r ESRAM .
./include/mach/db8500-regs.h:/* Base address and bank offsets for ESRAM */
./include/mach/db8500-regs.h:#define U8500_ESRAM_BASE 0x40000000
./include/mach/db8500-regs.h:#define U8500_ESRAM_BANK_SIZE 0x00020000
./include/mach/db8500-regs.h:#define U8500_ESRAM_BANK0 U8500_ESRAM_BASE
./include/mach/db8500-regs.h:#define U8500_ESRAM_BANK1 (U8500_ESRAM_BASE + U8500_ESRAM_BANK_SIZE)
./include/mach/db8500-regs.h:#define U8500_ESRAM_BANK2 (U8500_ESRAM_BANK1 + U8500_ESRAM_BANK_SIZE)
./include/mach/db8500-regs.h:#define U8500_ESRAM_BANK3 (U8500_ESRAM_BANK2 + U8500_ESRAM_BANK_SIZE)
./include/mach/db8500-regs.h:#define U8500_ESRAM_BANK4 (U8500_ESRAM_BANK3 + U8500_ESRAM_BANK_SIZE)
./include/mach/db8500-regs.h:#define U8500_ESRAM_DMA_LCPA_OFFSET 0x10000
./include/mach/db8500-regs.h:#define U8500_DMA_LCPA_BASE
(U8500_ESRAM_BANK0 + U8500_ESRAM_DMA_LCPA_OFFSET)
./include/mach/db8500-regs.h:#define U8500_DMA_LCLA_BASE U8500_ESRAM_BANK4
I want to use genalloc to do dynamic allocations but I need to be able to
fine tune the allocation algorithm. I my case best-fit algorithm give
better results than first-fit, but it will not be true for every use case.
Signed-off-by: Benjamin Gaignard <benjamin.gaignard@stericsson.com>
Cc: Huang Ying <ying.huang@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
For files only using THIS_MODULE and/or EXPORT_SYMBOL, map
them onto including export.h -- or if the file isn't even
using those, then just delete the include. Fix up any implicit
include dependencies that were being masked by module.h along
the way.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
This version of the gen_pool memory allocator supports lockless
operation.
This makes it safe to use in NMI handlers and other special
unblockable contexts that could otherwise deadlock on locks. This is
implemented by using atomic operations and retries on any conflicts.
The disadvantage is that there may be livelocks in extreme cases. For
better scalability, one gen_pool allocator can be used for each CPU.
The lockless operation only works if there is enough memory available.
If new memory is added to the pool a lock has to be still taken. So
any user relying on locklessness has to ensure that sufficient memory
is preallocated.
The basic atomic operation of this allocator is cmpxchg on long. On
architectures that don't have NMI-safe cmpxchg implementation, the
allocator can NOT be used in NMI handler. So code uses the allocator
in NMI handler should depend on CONFIG_ARCH_HAVE_NMI_SAFE_CMPXCHG.
Signed-off-by: Huang Ying <ying.huang@intel.com>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Len Brown <len.brown@intel.com>
So we can specify the virtual address as the base of the pool chunk and
then get physical addresses for hardware IP.
For example on at91 we will use this on spi, uart or macb
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Cc: Nicolas Ferre <nicolas.ferre@atmel.com>
Cc: Patrice VILCHEZ <patrice.vilchez@atmel.com>
Cc: Jes Sorensen <jes@wildopensource.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
bitmap_find_next_zero_area requires the size of the bitmap, we instead
passed the last suitable position. This made it impossible to allocate
from the end of the pool.
Fixes a regression introduced by 243797f59b
("genalloc: use bitmap_find_next_zero_area").
Signed-off-by: Imre Deak <imre.deak@nokia.com>
Cc: Zygo Blaxell <zygo.blaxell@xandros.com>
Cc: Tejun Heo <tj@kernel.org>
Acked-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
There is a call to write_lock() in gen_pool_destroy which is not balanced
by any corresponding write_unlock(). This causes problems with preemption
because the preemption-disable counter is incremented in the write_lock()
call, but never decremented by any call to write_unlock(). This bug is
gen_pool_destroy, and one of them is non-x86 arch-specific code.
Signed-off-by: Zygo Blaxell <zygo.blaxell@xandros.com>
Cc: Jiri Kosina <trivial@kernel.org>
Cc: Steve Wise <swise@opengridcomputing.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kmalloc_node() and kmem_cache_alloc_node() were not available in a zeroing
variant in the past. But with __GFP_ZERO it is possible now to do zeroing
while allocating.
Use __GFP_ZERO to remove the explicit clearing of memory via memset whereever
we can.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
lib/genalloc.c: In function 'gen_pool_alloc':
lib/genalloc.c:151: warning: passing argument 2 of '__set_bit' from incompatible pointer type
lib/genalloc.c: In function 'gen_pool_free':
lib/genalloc.c:190: warning: passing argument 2 of '__clear_bit' from incompatible pointer type
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The exported kernel interfaces of genpool allocator need to adhere to
the requirements of kernel-doc.
Signed-off-by: Dean Nelson <dcn@sgi.com>
Cc: Steve Wise <swise@opengridcomputing.com>
Acked-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Modules using the genpool allocator need to be able to destroy the data
structure when unloading.
Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Cc: Randy Dunlap <rdunlap@xenotime.net>
Cc: Dean Nelson <dcn@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Modify the gen_pool allocator (lib/genalloc.c) to utilize a bitmap scheme
instead of the buddy scheme. The purpose of this change is to eliminate
the touching of the actual memory being allocated.
Since the change modifies the interface, a change to the uncached allocator
(arch/ia64/kernel/uncached.c) is also required.
Both Andrey Volkov and Jes Sorenson have expressed a desire that the
gen_pool allocator not write to the memory being managed. See the
following:
http://marc.theaimsgroup.com/?l=linux-kernel&m=113518602713125&w=2http://marc.theaimsgroup.com/?l=linux-kernel&m=113533568827916&w=2
Signed-off-by: Dean Nelson <dcn@sgi.com>
Cc: Andrey Volkov <avolkov@varma-el.com>
Acked-by: Jes Sorensen <jes@trained-monkey.org>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
genalloc improperly stores the sizes of freed chunks, allocates overlapping
memory regions, and oopses after its in-band data is overwritten.
Signed-off-by: Chris Humbert <mahadri-kernel@drigon.com>
Cc: Jes Sorensen <jes@trained-monkey.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch contains the ia64 uncached page allocator and the generic
allocator (genalloc). The uncached allocator was formerly part of the SN2
mspec driver but there are several other users of it so it has been split
off from the driver.
The generic allocator can be used by device driver to manage special memory
etc. The generic allocator is based on the allocator from the sym53c8xx_2
driver.
Various users on ia64 needs uncached memory. The SGI SN architecture requires
it for inter-partition communication between partitions within a large NUMA
cluster. The specific user for this is the XPC code. Another application is
large MPI style applications which use it for synchronization, on SN this can
be done using special 'fetchop' operations but it also benefits non SN
hardware which may use regular uncached memory for this purpose. Performance
of doing this through uncached vs cached memory is pretty substantial. This
is handled by the mspec driver which I will push out in a seperate patch.
Rather than creating a specific allocator for just uncached memory I came up
with genalloc which is a generic purpose allocator that can be used by device
drivers and other subsystems as they please. For instance to handle onboard
device memory. It was derived from the sym53c7xx_2 driver's allocator which
is also an example of a potential user (I am refraining from modifying sym2
right now as it seems to have been under fairly heavy development recently).
On ia64 memory has various properties within a granule, ie. it isn't safe to
access memory as uncached within the same granule as currently has memory
accessed in cached mode. The regular system therefore doesn't utilize memory
in the lower granules which is mixed in with device PAL code etc. The
uncached driver walks the EFI memmap and pulls out the spill uncached pages
and sticks them into the uncached pool. Only after these chunks have been
utilized, will it start converting regular cached memory into uncached memory.
Hence the reason for the EFI related code additions.
Signed-off-by: Jes Sorensen <jes@wildopensource.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
