zfs/module/spl/spl-kmem-cache.c

1795 lines
50 KiB
C
Raw Normal View History

/*
* Copyright (C) 2007-2010 Lawrence Livermore National Security, LLC.
* Copyright (C) 2007 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
* Written by Brian Behlendorf <behlendorf1@llnl.gov>.
* UCRL-CODE-235197
*
* This file is part of the SPL, Solaris Porting Layer.
* For details, see <http://zfsonlinux.org/>.
*
* The SPL is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation; either version 2 of the License, or (at your
* option) any later version.
*
* The SPL is distributed in the hope that it will be useful, but WITHOUT
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* for more details.
*
* You should have received a copy of the GNU General Public License along
* with the SPL. If not, see <http://www.gnu.org/licenses/>.
*/
#include <sys/kmem.h>
#include <sys/kmem_cache.h>
#include <sys/shrinker.h>
#include <sys/taskq.h>
#include <sys/timer.h>
#include <sys/vmem.h>
#include <sys/wait.h>
#include <linux/slab.h>
#include <linux/swap.h>
#include <linux/prefetch.h>
/*
* Within the scope of spl-kmem.c file the kmem_cache_* definitions
* are removed to allow access to the real Linux slab allocator.
*/
#undef kmem_cache_destroy
#undef kmem_cache_create
#undef kmem_cache_alloc
#undef kmem_cache_free
/*
* Linux 3.16 replaced smp_mb__{before,after}_{atomic,clear}_{dec,inc,bit}()
* with smp_mb__{before,after}_atomic() because they were redundant. This is
* only used inside our SLAB allocator, so we implement an internal wrapper
* here to give us smp_mb__{before,after}_atomic() on older kernels.
*/
#ifndef smp_mb__before_atomic
#define smp_mb__before_atomic(x) smp_mb__before_clear_bit(x)
#endif
#ifndef smp_mb__after_atomic
#define smp_mb__after_atomic(x) smp_mb__after_clear_bit(x)
#endif
/*
* Cache expiration was implemented because it was part of the default Solaris
* kmem_cache behavior. The idea is that per-cpu objects which haven't been
* accessed in several seconds should be returned to the cache. On the other
* hand Linux slabs never move objects back to the slabs unless there is
* memory pressure on the system. By default the Linux method is enabled
* because it has been shown to improve responsiveness on low memory systems.
* This policy may be changed by setting KMC_EXPIRE_AGE or KMC_EXPIRE_MEM.
*/
/* BEGIN CSTYLED */
unsigned int spl_kmem_cache_expire = KMC_EXPIRE_MEM;
EXPORT_SYMBOL(spl_kmem_cache_expire);
module_param(spl_kmem_cache_expire, uint, 0644);
MODULE_PARM_DESC(spl_kmem_cache_expire, "By age (0x1) or low memory (0x2)");
/*
* Cache magazines are an optimization designed to minimize the cost of
* allocating memory. They do this by keeping a per-cpu cache of recently
* freed objects, which can then be reallocated without taking a lock. This
* can improve performance on highly contended caches. However, because
* objects in magazines will prevent otherwise empty slabs from being
* immediately released this may not be ideal for low memory machines.
*
* For this reason spl_kmem_cache_magazine_size can be used to set a maximum
* magazine size. When this value is set to 0 the magazine size will be
* automatically determined based on the object size. Otherwise magazines
* will be limited to 2-256 objects per magazine (i.e per cpu). Magazines
* may never be entirely disabled in this implementation.
*/
unsigned int spl_kmem_cache_magazine_size = 0;
module_param(spl_kmem_cache_magazine_size, uint, 0444);
MODULE_PARM_DESC(spl_kmem_cache_magazine_size,
"Default magazine size (2-256), set automatically (0)");
/*
* The default behavior is to report the number of objects remaining in the
* cache. This allows the Linux VM to repeatedly reclaim objects from the
* cache when memory is low satisfy other memory allocations. Alternately,
* setting this value to KMC_RECLAIM_ONCE limits how aggressively the cache
* is reclaimed. This may increase the likelihood of out of memory events.
*/
unsigned int spl_kmem_cache_reclaim = 0 /* KMC_RECLAIM_ONCE */;
module_param(spl_kmem_cache_reclaim, uint, 0644);
MODULE_PARM_DESC(spl_kmem_cache_reclaim, "Single reclaim pass (0x1)");
unsigned int spl_kmem_cache_obj_per_slab = SPL_KMEM_CACHE_OBJ_PER_SLAB;
module_param(spl_kmem_cache_obj_per_slab, uint, 0644);
MODULE_PARM_DESC(spl_kmem_cache_obj_per_slab, "Number of objects per slab");
unsigned int spl_kmem_cache_obj_per_slab_min = SPL_KMEM_CACHE_OBJ_PER_SLAB_MIN;
module_param(spl_kmem_cache_obj_per_slab_min, uint, 0644);
MODULE_PARM_DESC(spl_kmem_cache_obj_per_slab_min,
"Minimal number of objects per slab");
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
unsigned int spl_kmem_cache_max_size = SPL_KMEM_CACHE_MAX_SIZE;
module_param(spl_kmem_cache_max_size, uint, 0644);
MODULE_PARM_DESC(spl_kmem_cache_max_size, "Maximum size of slab in MB");
/*
* For small objects the Linux slab allocator should be used to make the most
* efficient use of the memory. However, large objects are not supported by
* the Linux slab and therefore the SPL implementation is preferred. A cutoff
* of 16K was determined to be optimal for architectures using 4K pages.
*/
#if PAGE_SIZE == 4096
unsigned int spl_kmem_cache_slab_limit = 16384;
#else
unsigned int spl_kmem_cache_slab_limit = 0;
#endif
module_param(spl_kmem_cache_slab_limit, uint, 0644);
MODULE_PARM_DESC(spl_kmem_cache_slab_limit,
"Objects less than N bytes use the Linux slab");
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
/*
* This value defaults to a threshold designed to avoid allocations which
* have been deemed costly by the kernel.
*/
unsigned int spl_kmem_cache_kmem_limit =
((1 << (PAGE_ALLOC_COSTLY_ORDER - 1)) * PAGE_SIZE) /
SPL_KMEM_CACHE_OBJ_PER_SLAB;
module_param(spl_kmem_cache_kmem_limit, uint, 0644);
MODULE_PARM_DESC(spl_kmem_cache_kmem_limit,
"Objects less than N bytes use the kmalloc");
Fix kmem cache deadlock logic The kmem cache implementation always adds new slabs by dispatching a task to the spl_kmem_cache taskq to perform the allocation. This is done because large slabs must be allocated using vmalloc(). It is possible these allocations will block on IO because the GFP_NOIO flag is not honored. This can result in a deadlock. Therefore, a deadlock detection strategy was implemented to deal with this case. When it is determined, by timeout, that the spl_kmem_cache thread has deadlocked attempting to add a new slab. Then all callers attempting to allocate from the cache fall back to using kmalloc() which does honor all passed flags. This logic was correct but an optimization in the code allowed for a deadlock. Because only slabs backed by vmalloc() can deadlock in the way described above. An optimization was made to only invoke this deadlock detection code for vmalloc() backed caches. This had the advantage of making it easy to distinguish these objects when they were freed. But this isn't strictly safe. If all the spl_kmem_cache threads end up deadlocked than we can't grow any of the other caches either. This can once again result in a deadlock if memory needs to be allocated from one of these other caches to ensure forward progress. The fix here is to remove the optimization which limits this fall back allocation stratagy to vmalloc() backed caches. Doing this means we may need to take the cache lock in spl_kmem_cache_free() call path. But this small cost can be mitigated by ignoring objects with virtual addresses. For good measure the default number of spl_kmem_cache threads has been increased from 1 to 4, and made tunable. This alone wouldn't resolve the original issue since it's still possible for all the threads to be deadlocked. However, it does help responsiveness by ensuring that a single deadlocked spl_kmem_cache thread doesn't block allocations from other caches until the timeout is reached. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2015-01-09 22:00:34 +00:00
/*
* The number of threads available to allocate new slabs for caches. This
* should not need to be tuned but it is available for performance analysis.
*/
unsigned int spl_kmem_cache_kmem_threads = 4;
module_param(spl_kmem_cache_kmem_threads, uint, 0444);
MODULE_PARM_DESC(spl_kmem_cache_kmem_threads,
"Number of spl_kmem_cache threads");
/* END CSTYLED */
Fix kmem cache deadlock logic The kmem cache implementation always adds new slabs by dispatching a task to the spl_kmem_cache taskq to perform the allocation. This is done because large slabs must be allocated using vmalloc(). It is possible these allocations will block on IO because the GFP_NOIO flag is not honored. This can result in a deadlock. Therefore, a deadlock detection strategy was implemented to deal with this case. When it is determined, by timeout, that the spl_kmem_cache thread has deadlocked attempting to add a new slab. Then all callers attempting to allocate from the cache fall back to using kmalloc() which does honor all passed flags. This logic was correct but an optimization in the code allowed for a deadlock. Because only slabs backed by vmalloc() can deadlock in the way described above. An optimization was made to only invoke this deadlock detection code for vmalloc() backed caches. This had the advantage of making it easy to distinguish these objects when they were freed. But this isn't strictly safe. If all the spl_kmem_cache threads end up deadlocked than we can't grow any of the other caches either. This can once again result in a deadlock if memory needs to be allocated from one of these other caches to ensure forward progress. The fix here is to remove the optimization which limits this fall back allocation stratagy to vmalloc() backed caches. Doing this means we may need to take the cache lock in spl_kmem_cache_free() call path. But this small cost can be mitigated by ignoring objects with virtual addresses. For good measure the default number of spl_kmem_cache threads has been increased from 1 to 4, and made tunable. This alone wouldn't resolve the original issue since it's still possible for all the threads to be deadlocked. However, it does help responsiveness by ensuring that a single deadlocked spl_kmem_cache thread doesn't block allocations from other caches until the timeout is reached. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2015-01-09 22:00:34 +00:00
/*
* Slab allocation interfaces
*
* While the Linux slab implementation was inspired by the Solaris
* implementation I cannot use it to emulate the Solaris APIs. I
* require two features which are not provided by the Linux slab.
*
* 1) Constructors AND destructors. Recent versions of the Linux
* kernel have removed support for destructors. This is a deal
* breaker for the SPL which contains particularly expensive
* initializers for mutex's, condition variables, etc. We also
* require a minimal level of cleanup for these data types unlike
* many Linux data types which do need to be explicitly destroyed.
*
* 2) Virtual address space backed slab. Callers of the Solaris slab
* expect it to work well for both small are very large allocations.
* Because of memory fragmentation the Linux slab which is backed
* by kmalloc'ed memory performs very badly when confronted with
* large numbers of large allocations. Basing the slab on the
* virtual address space removes the need for contiguous pages
* and greatly improve performance for large allocations.
*
* For these reasons, the SPL has its own slab implementation with
* the needed features. It is not as highly optimized as either the
* Solaris or Linux slabs, but it should get me most of what is
* needed until it can be optimized or obsoleted by another approach.
*
* One serious concern I do have about this method is the relatively
* small virtual address space on 32bit arches. This will seriously
* constrain the size of the slab caches and their performance.
*/
struct list_head spl_kmem_cache_list; /* List of caches */
struct rw_semaphore spl_kmem_cache_sem; /* Cache list lock */
taskq_t *spl_kmem_cache_taskq; /* Task queue for aging / reclaim */
static void spl_cache_shrink(spl_kmem_cache_t *skc, void *obj);
SPL_SHRINKER_CALLBACK_FWD_DECLARE(spl_kmem_cache_generic_shrinker);
SPL_SHRINKER_DECLARE(spl_kmem_cache_shrinker,
spl_kmem_cache_generic_shrinker, KMC_DEFAULT_SEEKS);
static void *
kv_alloc(spl_kmem_cache_t *skc, int size, int flags)
{
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
gfp_t lflags = kmem_flags_convert(flags);
void *ptr;
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
if (skc->skc_flags & KMC_KMEM) {
ASSERT(ISP2(size));
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
ptr = (void *)__get_free_pages(lflags, get_order(size));
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
} else {
ptr = __vmalloc(size, lflags | __GFP_HIGHMEM, PAGE_KERNEL);
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
}
/* Resulting allocated memory will be page aligned */
ASSERT(IS_P2ALIGNED(ptr, PAGE_SIZE));
return (ptr);
}
static void
kv_free(spl_kmem_cache_t *skc, void *ptr, int size)
{
ASSERT(IS_P2ALIGNED(ptr, PAGE_SIZE));
/*
* The Linux direct reclaim path uses this out of band value to
* determine if forward progress is being made. Normally this is
* incremented by kmem_freepages() which is part of the various
* Linux slab implementations. However, since we are using none
* of that infrastructure we are responsible for incrementing it.
*/
if (current->reclaim_state)
current->reclaim_state->reclaimed_slab += size >> PAGE_SHIFT;
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
if (skc->skc_flags & KMC_KMEM) {
ASSERT(ISP2(size));
free_pages((unsigned long)ptr, get_order(size));
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
} else {
vfree(ptr);
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
}
}
/*
* Required space for each aligned sks.
*/
static inline uint32_t
spl_sks_size(spl_kmem_cache_t *skc)
{
return (P2ROUNDUP_TYPED(sizeof (spl_kmem_slab_t),
skc->skc_obj_align, uint32_t));
}
/*
* Required space for each aligned object.
*/
static inline uint32_t
spl_obj_size(spl_kmem_cache_t *skc)
{
uint32_t align = skc->skc_obj_align;
return (P2ROUNDUP_TYPED(skc->skc_obj_size, align, uint32_t) +
P2ROUNDUP_TYPED(sizeof (spl_kmem_obj_t), align, uint32_t));
}
/*
* Lookup the spl_kmem_object_t for an object given that object.
*/
static inline spl_kmem_obj_t *
spl_sko_from_obj(spl_kmem_cache_t *skc, void *obj)
{
return (obj + P2ROUNDUP_TYPED(skc->skc_obj_size,
skc->skc_obj_align, uint32_t));
}
/*
* Required space for each offslab object taking in to account alignment
* restrictions and the power-of-two requirement of kv_alloc().
*/
static inline uint32_t
spl_offslab_size(spl_kmem_cache_t *skc)
{
return (1UL << (fls64(spl_obj_size(skc)) + 1));
}
/*
* It's important that we pack the spl_kmem_obj_t structure and the
* actual objects in to one large address space to minimize the number
* of calls to the allocator. It is far better to do a few large
* allocations and then subdivide it ourselves. Now which allocator
* we use requires balancing a few trade offs.
*
* For small objects we use kmem_alloc() because as long as you are
* only requesting a small number of pages (ideally just one) its cheap.
* However, when you start requesting multiple pages with kmem_alloc()
* it gets increasingly expensive since it requires contiguous pages.
* For this reason we shift to vmem_alloc() for slabs of large objects
* which removes the need for contiguous pages. We do not use
* vmem_alloc() in all cases because there is significant locking
* overhead in __get_vm_area_node(). This function takes a single
* global lock when acquiring an available virtual address range which
* serializes all vmem_alloc()'s for all slab caches. Using slightly
* different allocation functions for small and large objects should
* give us the best of both worlds.
*
* KMC_ONSLAB KMC_OFFSLAB
*
* +------------------------+ +-----------------+
* | spl_kmem_slab_t --+-+ | | spl_kmem_slab_t |---+-+
* | skc_obj_size <-+ | | +-----------------+ | |
* | spl_kmem_obj_t | | | |
* | skc_obj_size <---+ | +-----------------+ | |
* | spl_kmem_obj_t | | | skc_obj_size | <-+ |
* | ... v | | spl_kmem_obj_t | |
* +------------------------+ +-----------------+ v
*/
static spl_kmem_slab_t *
spl_slab_alloc(spl_kmem_cache_t *skc, int flags)
{
spl_kmem_slab_t *sks;
spl_kmem_obj_t *sko, *n;
void *base, *obj;
uint32_t obj_size, offslab_size = 0;
int i, rc = 0;
base = kv_alloc(skc, skc->skc_slab_size, flags);
if (base == NULL)
return (NULL);
sks = (spl_kmem_slab_t *)base;
sks->sks_magic = SKS_MAGIC;
sks->sks_objs = skc->skc_slab_objs;
sks->sks_age = jiffies;
sks->sks_cache = skc;
INIT_LIST_HEAD(&sks->sks_list);
INIT_LIST_HEAD(&sks->sks_free_list);
sks->sks_ref = 0;
obj_size = spl_obj_size(skc);
if (skc->skc_flags & KMC_OFFSLAB)
offslab_size = spl_offslab_size(skc);
for (i = 0; i < sks->sks_objs; i++) {
if (skc->skc_flags & KMC_OFFSLAB) {
obj = kv_alloc(skc, offslab_size, flags);
if (!obj) {
rc = -ENOMEM;
goto out;
}
} else {
obj = base + spl_sks_size(skc) + (i * obj_size);
}
ASSERT(IS_P2ALIGNED(obj, skc->skc_obj_align));
sko = spl_sko_from_obj(skc, obj);
sko->sko_addr = obj;
sko->sko_magic = SKO_MAGIC;
sko->sko_slab = sks;
INIT_LIST_HEAD(&sko->sko_list);
list_add_tail(&sko->sko_list, &sks->sks_free_list);
}
out:
if (rc) {
if (skc->skc_flags & KMC_OFFSLAB)
list_for_each_entry_safe(sko,
n, &sks->sks_free_list, sko_list) {
kv_free(skc, sko->sko_addr, offslab_size);
}
kv_free(skc, base, skc->skc_slab_size);
sks = NULL;
}
return (sks);
}
/*
* Remove a slab from complete or partial list, it must be called with
* the 'skc->skc_lock' held but the actual free must be performed
* outside the lock to prevent deadlocking on vmem addresses.
*/
static void
spl_slab_free(spl_kmem_slab_t *sks,
struct list_head *sks_list, struct list_head *sko_list)
{
spl_kmem_cache_t *skc;
ASSERT(sks->sks_magic == SKS_MAGIC);
ASSERT(sks->sks_ref == 0);
skc = sks->sks_cache;
ASSERT(skc->skc_magic == SKC_MAGIC);
/*
* Update slab/objects counters in the cache, then remove the
* slab from the skc->skc_partial_list. Finally add the slab
* and all its objects in to the private work lists where the
* destructors will be called and the memory freed to the system.
*/
skc->skc_obj_total -= sks->sks_objs;
skc->skc_slab_total--;
list_del(&sks->sks_list);
list_add(&sks->sks_list, sks_list);
list_splice_init(&sks->sks_free_list, sko_list);
}
/*
* Reclaim empty slabs at the end of the partial list.
*/
static void
spl_slab_reclaim(spl_kmem_cache_t *skc)
{
spl_kmem_slab_t *sks, *m;
spl_kmem_obj_t *sko, *n;
LIST_HEAD(sks_list);
LIST_HEAD(sko_list);
uint32_t size = 0;
/*
* Empty slabs and objects must be moved to a private list so they
* can be safely freed outside the spin lock. All empty slabs are
* at the end of skc->skc_partial_list, therefore once a non-empty
* slab is found we can stop scanning.
*/
spin_lock(&skc->skc_lock);
list_for_each_entry_safe_reverse(sks, m,
&skc->skc_partial_list, sks_list) {
if (sks->sks_ref > 0)
break;
spl_slab_free(sks, &sks_list, &sko_list);
}
spin_unlock(&skc->skc_lock);
/*
* The following two loops ensure all the object destructors are
* run, any offslab objects are freed, and the slabs themselves
* are freed. This is all done outside the skc->skc_lock since
* this allows the destructor to sleep, and allows us to perform
* a conditional reschedule when a freeing a large number of
* objects and slabs back to the system.
*/
if (skc->skc_flags & KMC_OFFSLAB)
size = spl_offslab_size(skc);
list_for_each_entry_safe(sko, n, &sko_list, sko_list) {
ASSERT(sko->sko_magic == SKO_MAGIC);
if (skc->skc_flags & KMC_OFFSLAB)
kv_free(skc, sko->sko_addr, size);
}
list_for_each_entry_safe(sks, m, &sks_list, sks_list) {
ASSERT(sks->sks_magic == SKS_MAGIC);
kv_free(skc, sks, skc->skc_slab_size);
}
}
static spl_kmem_emergency_t *
spl_emergency_search(struct rb_root *root, void *obj)
{
struct rb_node *node = root->rb_node;
spl_kmem_emergency_t *ske;
unsigned long address = (unsigned long)obj;
while (node) {
ske = container_of(node, spl_kmem_emergency_t, ske_node);
if (address < ske->ske_obj)
node = node->rb_left;
else if (address > ske->ske_obj)
node = node->rb_right;
else
return (ske);
}
return (NULL);
}
static int
spl_emergency_insert(struct rb_root *root, spl_kmem_emergency_t *ske)
{
struct rb_node **new = &(root->rb_node), *parent = NULL;
spl_kmem_emergency_t *ske_tmp;
unsigned long address = ske->ske_obj;
while (*new) {
ske_tmp = container_of(*new, spl_kmem_emergency_t, ske_node);
parent = *new;
if (address < ske_tmp->ske_obj)
new = &((*new)->rb_left);
else if (address > ske_tmp->ske_obj)
new = &((*new)->rb_right);
else
return (0);
}
rb_link_node(&ske->ske_node, parent, new);
rb_insert_color(&ske->ske_node, root);
return (1);
}
/*
* Allocate a single emergency object and track it in a red black tree.
*/
static int
spl_emergency_alloc(spl_kmem_cache_t *skc, int flags, void **obj)
{
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
gfp_t lflags = kmem_flags_convert(flags);
spl_kmem_emergency_t *ske;
int order = get_order(skc->skc_obj_size);
int empty;
/* Last chance use a partial slab if one now exists */
spin_lock(&skc->skc_lock);
empty = list_empty(&skc->skc_partial_list);
spin_unlock(&skc->skc_lock);
if (!empty)
return (-EEXIST);
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
ske = kmalloc(sizeof (*ske), lflags);
if (ske == NULL)
return (-ENOMEM);
ske->ske_obj = __get_free_pages(lflags, order);
if (ske->ske_obj == 0) {
kfree(ske);
return (-ENOMEM);
}
spin_lock(&skc->skc_lock);
empty = spl_emergency_insert(&skc->skc_emergency_tree, ske);
if (likely(empty)) {
skc->skc_obj_total++;
skc->skc_obj_emergency++;
if (skc->skc_obj_emergency > skc->skc_obj_emergency_max)
skc->skc_obj_emergency_max = skc->skc_obj_emergency;
}
spin_unlock(&skc->skc_lock);
if (unlikely(!empty)) {
free_pages(ske->ske_obj, order);
kfree(ske);
return (-EINVAL);
}
*obj = (void *)ske->ske_obj;
return (0);
}
/*
* Locate the passed object in the red black tree and free it.
*/
static int
spl_emergency_free(spl_kmem_cache_t *skc, void *obj)
{
spl_kmem_emergency_t *ske;
int order = get_order(skc->skc_obj_size);
spin_lock(&skc->skc_lock);
ske = spl_emergency_search(&skc->skc_emergency_tree, obj);
Fix kmem cache deadlock logic The kmem cache implementation always adds new slabs by dispatching a task to the spl_kmem_cache taskq to perform the allocation. This is done because large slabs must be allocated using vmalloc(). It is possible these allocations will block on IO because the GFP_NOIO flag is not honored. This can result in a deadlock. Therefore, a deadlock detection strategy was implemented to deal with this case. When it is determined, by timeout, that the spl_kmem_cache thread has deadlocked attempting to add a new slab. Then all callers attempting to allocate from the cache fall back to using kmalloc() which does honor all passed flags. This logic was correct but an optimization in the code allowed for a deadlock. Because only slabs backed by vmalloc() can deadlock in the way described above. An optimization was made to only invoke this deadlock detection code for vmalloc() backed caches. This had the advantage of making it easy to distinguish these objects when they were freed. But this isn't strictly safe. If all the spl_kmem_cache threads end up deadlocked than we can't grow any of the other caches either. This can once again result in a deadlock if memory needs to be allocated from one of these other caches to ensure forward progress. The fix here is to remove the optimization which limits this fall back allocation stratagy to vmalloc() backed caches. Doing this means we may need to take the cache lock in spl_kmem_cache_free() call path. But this small cost can be mitigated by ignoring objects with virtual addresses. For good measure the default number of spl_kmem_cache threads has been increased from 1 to 4, and made tunable. This alone wouldn't resolve the original issue since it's still possible for all the threads to be deadlocked. However, it does help responsiveness by ensuring that a single deadlocked spl_kmem_cache thread doesn't block allocations from other caches until the timeout is reached. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2015-01-09 22:00:34 +00:00
if (ske) {
rb_erase(&ske->ske_node, &skc->skc_emergency_tree);
skc->skc_obj_emergency--;
skc->skc_obj_total--;
}
spin_unlock(&skc->skc_lock);
Fix kmem cache deadlock logic The kmem cache implementation always adds new slabs by dispatching a task to the spl_kmem_cache taskq to perform the allocation. This is done because large slabs must be allocated using vmalloc(). It is possible these allocations will block on IO because the GFP_NOIO flag is not honored. This can result in a deadlock. Therefore, a deadlock detection strategy was implemented to deal with this case. When it is determined, by timeout, that the spl_kmem_cache thread has deadlocked attempting to add a new slab. Then all callers attempting to allocate from the cache fall back to using kmalloc() which does honor all passed flags. This logic was correct but an optimization in the code allowed for a deadlock. Because only slabs backed by vmalloc() can deadlock in the way described above. An optimization was made to only invoke this deadlock detection code for vmalloc() backed caches. This had the advantage of making it easy to distinguish these objects when they were freed. But this isn't strictly safe. If all the spl_kmem_cache threads end up deadlocked than we can't grow any of the other caches either. This can once again result in a deadlock if memory needs to be allocated from one of these other caches to ensure forward progress. The fix here is to remove the optimization which limits this fall back allocation stratagy to vmalloc() backed caches. Doing this means we may need to take the cache lock in spl_kmem_cache_free() call path. But this small cost can be mitigated by ignoring objects with virtual addresses. For good measure the default number of spl_kmem_cache threads has been increased from 1 to 4, and made tunable. This alone wouldn't resolve the original issue since it's still possible for all the threads to be deadlocked. However, it does help responsiveness by ensuring that a single deadlocked spl_kmem_cache thread doesn't block allocations from other caches until the timeout is reached. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2015-01-09 22:00:34 +00:00
if (ske == NULL)
return (-ENOENT);
free_pages(ske->ske_obj, order);
kfree(ske);
return (0);
}
/*
* Release objects from the per-cpu magazine back to their slab. The flush
* argument contains the max number of entries to remove from the magazine.
*/
static void
__spl_cache_flush(spl_kmem_cache_t *skc, spl_kmem_magazine_t *skm, int flush)
{
int i, count = MIN(flush, skm->skm_avail);
ASSERT(skc->skc_magic == SKC_MAGIC);
ASSERT(skm->skm_magic == SKM_MAGIC);
for (i = 0; i < count; i++)
spl_cache_shrink(skc, skm->skm_objs[i]);
skm->skm_avail -= count;
memmove(skm->skm_objs, &(skm->skm_objs[count]),
sizeof (void *) * skm->skm_avail);
}
static void
spl_cache_flush(spl_kmem_cache_t *skc, spl_kmem_magazine_t *skm, int flush)
{
spin_lock(&skc->skc_lock);
__spl_cache_flush(skc, skm, flush);
spin_unlock(&skc->skc_lock);
}
static void
spl_magazine_age(void *data)
{
spl_kmem_cache_t *skc = (spl_kmem_cache_t *)data;
spl_kmem_magazine_t *skm = skc->skc_mag[smp_processor_id()];
ASSERT(skm->skm_magic == SKM_MAGIC);
ASSERT(skm->skm_cpu == smp_processor_id());
ASSERT(irqs_disabled());
/* There are no available objects or they are too young to age out */
if ((skm->skm_avail == 0) ||
time_before(jiffies, skm->skm_age + skc->skc_delay * HZ))
return;
/*
* Because we're executing in interrupt context we may have
* interrupted the holder of this lock. To avoid a potential
* deadlock return if the lock is contended.
*/
if (!spin_trylock(&skc->skc_lock))
return;
__spl_cache_flush(skc, skm, skm->skm_refill);
spin_unlock(&skc->skc_lock);
}
/*
* Called regularly to keep a downward pressure on the cache.
*
* Objects older than skc->skc_delay seconds in the per-cpu magazines will
* be returned to the caches. This is done to prevent idle magazines from
* holding memory which could be better used elsewhere. The delay is
* present to prevent thrashing the magazine.
*
* The newly released objects may result in empty partial slabs. Those
* slabs should be released to the system. Otherwise moving the objects
* out of the magazines is just wasted work.
*/
static void
spl_cache_age(void *data)
{
spl_kmem_cache_t *skc = (spl_kmem_cache_t *)data;
taskqid_t id = 0;
ASSERT(skc->skc_magic == SKC_MAGIC);
/* Dynamically disabled at run time */
if (!(spl_kmem_cache_expire & KMC_EXPIRE_AGE))
return;
atomic_inc(&skc->skc_ref);
if (!(skc->skc_flags & KMC_NOMAGAZINE))
on_each_cpu(spl_magazine_age, skc, 1);
spl_slab_reclaim(skc);
while (!test_bit(KMC_BIT_DESTROY, &skc->skc_flags) && !id) {
id = taskq_dispatch_delay(
spl_kmem_cache_taskq, spl_cache_age, skc, TQ_SLEEP,
ddi_get_lbolt() + skc->skc_delay / 3 * HZ);
/* Destroy issued after dispatch immediately cancel it */
if (test_bit(KMC_BIT_DESTROY, &skc->skc_flags) && id)
taskq_cancel_id(spl_kmem_cache_taskq, id);
}
spin_lock(&skc->skc_lock);
skc->skc_taskqid = id;
spin_unlock(&skc->skc_lock);
atomic_dec(&skc->skc_ref);
}
/*
* Size a slab based on the size of each aligned object plus spl_kmem_obj_t.
* When on-slab we want to target spl_kmem_cache_obj_per_slab. However,
* for very small objects we may end up with more than this so as not
* to waste space in the minimal allocation of a single page. Also for
* very large objects we may use as few as spl_kmem_cache_obj_per_slab_min,
* lower than this and we will fail.
*/
static int
spl_slab_size(spl_kmem_cache_t *skc, uint32_t *objs, uint32_t *size)
{
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
uint32_t sks_size, obj_size, max_size, tgt_size, tgt_objs;
if (skc->skc_flags & KMC_OFFSLAB) {
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
tgt_objs = spl_kmem_cache_obj_per_slab;
tgt_size = P2ROUNDUP(sizeof (spl_kmem_slab_t), PAGE_SIZE);
if ((skc->skc_flags & KMC_KMEM) &&
(spl_obj_size(skc) > (SPL_MAX_ORDER_NR_PAGES * PAGE_SIZE)))
return (-ENOSPC);
} else {
sks_size = spl_sks_size(skc);
obj_size = spl_obj_size(skc);
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
max_size = (spl_kmem_cache_max_size * 1024 * 1024);
tgt_size = (spl_kmem_cache_obj_per_slab * obj_size + sks_size);
/*
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
* KMC_KMEM slabs are allocated by __get_free_pages() which
* rounds up to the nearest order. Knowing this the size
* should be rounded up to the next power of two with a hard
* maximum defined by the maximum allowed allocation order.
*/
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
if (skc->skc_flags & KMC_KMEM) {
max_size = SPL_MAX_ORDER_NR_PAGES * PAGE_SIZE;
tgt_size = MIN(max_size,
PAGE_SIZE * (1 << MAX(get_order(tgt_size) - 1, 1)));
}
if (tgt_size <= max_size) {
tgt_objs = (tgt_size - sks_size) / obj_size;
} else {
tgt_objs = (max_size - sks_size) / obj_size;
tgt_size = (tgt_objs * obj_size) + sks_size;
}
}
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
if (tgt_objs == 0)
return (-ENOSPC);
*objs = tgt_objs;
*size = tgt_size;
return (0);
}
/*
* Make a guess at reasonable per-cpu magazine size based on the size of
* each object and the cost of caching N of them in each magazine. Long
* term this should really adapt based on an observed usage heuristic.
*/
static int
spl_magazine_size(spl_kmem_cache_t *skc)
{
uint32_t obj_size = spl_obj_size(skc);
int size;
if (spl_kmem_cache_magazine_size > 0)
return (MAX(MIN(spl_kmem_cache_magazine_size, 256), 2));
/* Per-magazine sizes below assume a 4Kib page size */
if (obj_size > (PAGE_SIZE * 256))
size = 4; /* Minimum 4Mib per-magazine */
else if (obj_size > (PAGE_SIZE * 32))
size = 16; /* Minimum 2Mib per-magazine */
else if (obj_size > (PAGE_SIZE))
size = 64; /* Minimum 256Kib per-magazine */
else if (obj_size > (PAGE_SIZE / 4))
size = 128; /* Minimum 128Kib per-magazine */
else
size = 256;
return (size);
}
/*
* Allocate a per-cpu magazine to associate with a specific core.
*/
static spl_kmem_magazine_t *
spl_magazine_alloc(spl_kmem_cache_t *skc, int cpu)
{
spl_kmem_magazine_t *skm;
int size = sizeof (spl_kmem_magazine_t) +
sizeof (void *) * skc->skc_mag_size;
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
skm = kmalloc_node(size, GFP_KERNEL, cpu_to_node(cpu));
if (skm) {
skm->skm_magic = SKM_MAGIC;
skm->skm_avail = 0;
skm->skm_size = skc->skc_mag_size;
skm->skm_refill = skc->skc_mag_refill;
skm->skm_cache = skc;
skm->skm_age = jiffies;
skm->skm_cpu = cpu;
}
return (skm);
}
/*
* Free a per-cpu magazine associated with a specific core.
*/
static void
spl_magazine_free(spl_kmem_magazine_t *skm)
{
ASSERT(skm->skm_magic == SKM_MAGIC);
ASSERT(skm->skm_avail == 0);
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
kfree(skm);
}
/*
* Create all pre-cpu magazines of reasonable sizes.
*/
static int
spl_magazine_create(spl_kmem_cache_t *skc)
{
int i;
if (skc->skc_flags & KMC_NOMAGAZINE)
return (0);
skc->skc_mag = kzalloc(sizeof (spl_kmem_magazine_t *) *
num_possible_cpus(), kmem_flags_convert(KM_SLEEP));
skc->skc_mag_size = spl_magazine_size(skc);
skc->skc_mag_refill = (skc->skc_mag_size + 1) / 2;
for_each_possible_cpu(i) {
skc->skc_mag[i] = spl_magazine_alloc(skc, i);
if (!skc->skc_mag[i]) {
for (i--; i >= 0; i--)
spl_magazine_free(skc->skc_mag[i]);
kfree(skc->skc_mag);
return (-ENOMEM);
}
}
return (0);
}
/*
* Destroy all pre-cpu magazines.
*/
static void
spl_magazine_destroy(spl_kmem_cache_t *skc)
{
spl_kmem_magazine_t *skm;
int i;
if (skc->skc_flags & KMC_NOMAGAZINE)
return;
for_each_possible_cpu(i) {
skm = skc->skc_mag[i];
spl_cache_flush(skc, skm, skm->skm_avail);
spl_magazine_free(skm);
}
kfree(skc->skc_mag);
}
/*
* Create a object cache based on the following arguments:
* name cache name
* size cache object size
* align cache object alignment
* ctor cache object constructor
* dtor cache object destructor
* reclaim cache object reclaim
* priv cache private data for ctor/dtor/reclaim
* vmp unused must be NULL
* flags
* KMC_KMEM Force SPL kmem backed cache
* KMC_VMEM Force SPL vmem backed cache
* KMC_SLAB Force Linux slab backed cache
* KMC_OFFSLAB Locate objects off the slab
* KMC_NOTOUCH Disable cache object aging (unsupported)
* KMC_NODEBUG Disable debugging (unsupported)
* KMC_NOHASH Disable hashing (unsupported)
* KMC_QCACHE Disable qcache (unsupported)
* KMC_NOMAGAZINE Enabled for kmem/vmem, Disabled for Linux slab
*/
spl_kmem_cache_t *
spl_kmem_cache_create(char *name, size_t size, size_t align,
spl_kmem_ctor_t ctor, spl_kmem_dtor_t dtor, spl_kmem_reclaim_t reclaim,
void *priv, void *vmp, int flags)
{
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
gfp_t lflags = kmem_flags_convert(KM_SLEEP);
spl_kmem_cache_t *skc;
int rc;
/*
* Unsupported flags
*/
ASSERT0(flags & KMC_NOMAGAZINE);
ASSERT0(flags & KMC_NOHASH);
ASSERT0(flags & KMC_QCACHE);
ASSERT(vmp == NULL);
might_sleep();
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
skc = kzalloc(sizeof (*skc), lflags);
if (skc == NULL)
return (NULL);
skc->skc_magic = SKC_MAGIC;
skc->skc_name_size = strlen(name) + 1;
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
skc->skc_name = (char *)kmalloc(skc->skc_name_size, lflags);
if (skc->skc_name == NULL) {
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
kfree(skc);
return (NULL);
}
strncpy(skc->skc_name, name, skc->skc_name_size);
skc->skc_ctor = ctor;
skc->skc_dtor = dtor;
skc->skc_reclaim = reclaim;
skc->skc_private = priv;
skc->skc_vmp = vmp;
skc->skc_linux_cache = NULL;
skc->skc_flags = flags;
skc->skc_obj_size = size;
skc->skc_obj_align = SPL_KMEM_CACHE_ALIGN;
skc->skc_delay = SPL_KMEM_CACHE_DELAY;
skc->skc_reap = SPL_KMEM_CACHE_REAP;
atomic_set(&skc->skc_ref, 0);
INIT_LIST_HEAD(&skc->skc_list);
INIT_LIST_HEAD(&skc->skc_complete_list);
INIT_LIST_HEAD(&skc->skc_partial_list);
skc->skc_emergency_tree = RB_ROOT;
spin_lock_init(&skc->skc_lock);
init_waitqueue_head(&skc->skc_waitq);
skc->skc_slab_fail = 0;
skc->skc_slab_create = 0;
skc->skc_slab_destroy = 0;
skc->skc_slab_total = 0;
skc->skc_slab_alloc = 0;
skc->skc_slab_max = 0;
skc->skc_obj_total = 0;
skc->skc_obj_alloc = 0;
skc->skc_obj_max = 0;
skc->skc_obj_deadlock = 0;
skc->skc_obj_emergency = 0;
skc->skc_obj_emergency_max = 0;
/*
* Verify the requested alignment restriction is sane.
*/
if (align) {
VERIFY(ISP2(align));
VERIFY3U(align, >=, SPL_KMEM_CACHE_ALIGN);
VERIFY3U(align, <=, PAGE_SIZE);
skc->skc_obj_align = align;
}
/*
* When no specific type of slab is requested (kmem, vmem, or
* linuxslab) then select a cache type based on the object size
* and default tunables.
*/
if (!(skc->skc_flags & (KMC_KMEM | KMC_VMEM | KMC_SLAB))) {
if (spl_kmem_cache_slab_limit &&
size <= (size_t)spl_kmem_cache_slab_limit) {
/*
* Objects smaller than spl_kmem_cache_slab_limit can
* use the Linux slab for better space-efficiency.
*/
skc->skc_flags |= KMC_SLAB;
} else if (spl_obj_size(skc) <= spl_kmem_cache_kmem_limit) {
/*
* Small objects, less than spl_kmem_cache_kmem_limit
* per object should use kmem because their slabs are
* small.
*/
skc->skc_flags |= KMC_KMEM;
} else {
/*
* All other objects are considered large and are
* placed on vmem backed slabs.
*/
skc->skc_flags |= KMC_VMEM;
}
}
/*
* Given the type of slab allocate the required resources.
*/
if (skc->skc_flags & (KMC_KMEM | KMC_VMEM)) {
rc = spl_slab_size(skc,
&skc->skc_slab_objs, &skc->skc_slab_size);
if (rc)
goto out;
rc = spl_magazine_create(skc);
if (rc)
goto out;
} else {
unsigned long slabflags = 0;
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
if (size > (SPL_MAX_KMEM_ORDER_NR_PAGES * PAGE_SIZE)) {
rc = EINVAL;
goto out;
}
#if defined(SLAB_USERCOPY)
/*
* Required for PAX-enabled kernels if the slab is to be
* used for copying between user and kernel space.
*/
slabflags |= SLAB_USERCOPY;
#endif
#if defined(HAVE_KMEM_CACHE_CREATE_USERCOPY)
/*
* Newer grsec patchset uses kmem_cache_create_usercopy()
* instead of SLAB_USERCOPY flag
*/
skc->skc_linux_cache = kmem_cache_create_usercopy(
skc->skc_name, size, align, slabflags, 0, size, NULL);
#else
skc->skc_linux_cache = kmem_cache_create(
skc->skc_name, size, align, slabflags, NULL);
#endif
if (skc->skc_linux_cache == NULL) {
rc = ENOMEM;
goto out;
}
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
#if defined(HAVE_KMEM_CACHE_ALLOCFLAGS)
skc->skc_linux_cache->allocflags |= __GFP_COMP;
#elif defined(HAVE_KMEM_CACHE_GFPFLAGS)
skc->skc_linux_cache->gfpflags |= __GFP_COMP;
#endif
skc->skc_flags |= KMC_NOMAGAZINE;
}
if (spl_kmem_cache_expire & KMC_EXPIRE_AGE) {
skc->skc_taskqid = taskq_dispatch_delay(spl_kmem_cache_taskq,
spl_cache_age, skc, TQ_SLEEP,
ddi_get_lbolt() + skc->skc_delay / 3 * HZ);
}
down_write(&spl_kmem_cache_sem);
list_add_tail(&skc->skc_list, &spl_kmem_cache_list);
up_write(&spl_kmem_cache_sem);
return (skc);
out:
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
kfree(skc->skc_name);
kfree(skc);
return (NULL);
}
EXPORT_SYMBOL(spl_kmem_cache_create);
/*
* Register a move callback for cache defragmentation.
* XXX: Unimplemented but harmless to stub out for now.
*/
void
spl_kmem_cache_set_move(spl_kmem_cache_t *skc,
kmem_cbrc_t (move)(void *, void *, size_t, void *))
{
ASSERT(move != NULL);
}
EXPORT_SYMBOL(spl_kmem_cache_set_move);
/*
* Destroy a cache and all objects associated with the cache.
*/
void
spl_kmem_cache_destroy(spl_kmem_cache_t *skc)
{
DECLARE_WAIT_QUEUE_HEAD(wq);
taskqid_t id;
ASSERT(skc->skc_magic == SKC_MAGIC);
ASSERT(skc->skc_flags & (KMC_KMEM | KMC_VMEM | KMC_SLAB));
down_write(&spl_kmem_cache_sem);
list_del_init(&skc->skc_list);
up_write(&spl_kmem_cache_sem);
/* Cancel any and wait for any pending delayed tasks */
VERIFY(!test_and_set_bit(KMC_BIT_DESTROY, &skc->skc_flags));
spin_lock(&skc->skc_lock);
id = skc->skc_taskqid;
spin_unlock(&skc->skc_lock);
taskq_cancel_id(spl_kmem_cache_taskq, id);
/*
* Wait until all current callers complete, this is mainly
* to catch the case where a low memory situation triggers a
* cache reaping action which races with this destroy.
*/
wait_event(wq, atomic_read(&skc->skc_ref) == 0);
if (skc->skc_flags & (KMC_KMEM | KMC_VMEM)) {
spl_magazine_destroy(skc);
spl_slab_reclaim(skc);
} else {
ASSERT(skc->skc_flags & KMC_SLAB);
kmem_cache_destroy(skc->skc_linux_cache);
}
spin_lock(&skc->skc_lock);
/*
* Validate there are no objects in use and free all the
* spl_kmem_slab_t, spl_kmem_obj_t, and object buffers.
*/
ASSERT3U(skc->skc_slab_alloc, ==, 0);
ASSERT3U(skc->skc_obj_alloc, ==, 0);
ASSERT3U(skc->skc_slab_total, ==, 0);
ASSERT3U(skc->skc_obj_total, ==, 0);
ASSERT3U(skc->skc_obj_emergency, ==, 0);
ASSERT(list_empty(&skc->skc_complete_list));
spin_unlock(&skc->skc_lock);
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
kfree(skc->skc_name);
kfree(skc);
}
EXPORT_SYMBOL(spl_kmem_cache_destroy);
/*
* Allocate an object from a slab attached to the cache. This is used to
* repopulate the per-cpu magazine caches in batches when they run low.
*/
static void *
spl_cache_obj(spl_kmem_cache_t *skc, spl_kmem_slab_t *sks)
{
spl_kmem_obj_t *sko;
ASSERT(skc->skc_magic == SKC_MAGIC);
ASSERT(sks->sks_magic == SKS_MAGIC);
sko = list_entry(sks->sks_free_list.next, spl_kmem_obj_t, sko_list);
ASSERT(sko->sko_magic == SKO_MAGIC);
ASSERT(sko->sko_addr != NULL);
/* Remove from sks_free_list */
list_del_init(&sko->sko_list);
sks->sks_age = jiffies;
sks->sks_ref++;
skc->skc_obj_alloc++;
/* Track max obj usage statistics */
if (skc->skc_obj_alloc > skc->skc_obj_max)
skc->skc_obj_max = skc->skc_obj_alloc;
/* Track max slab usage statistics */
if (sks->sks_ref == 1) {
skc->skc_slab_alloc++;
if (skc->skc_slab_alloc > skc->skc_slab_max)
skc->skc_slab_max = skc->skc_slab_alloc;
}
return (sko->sko_addr);
}
/*
* Generic slab allocation function to run by the global work queues.
* It is responsible for allocating a new slab, linking it in to the list
* of partial slabs, and then waking any waiters.
*/
static int
__spl_cache_grow(spl_kmem_cache_t *skc, int flags)
{
spl_kmem_slab_t *sks;
fstrans_cookie_t cookie = spl_fstrans_mark();
sks = spl_slab_alloc(skc, flags);
spl_fstrans_unmark(cookie);
spin_lock(&skc->skc_lock);
if (sks) {
skc->skc_slab_total++;
skc->skc_obj_total += sks->sks_objs;
list_add_tail(&sks->sks_list, &skc->skc_partial_list);
smp_mb__before_atomic();
clear_bit(KMC_BIT_DEADLOCKED, &skc->skc_flags);
smp_mb__after_atomic();
wake_up_all(&skc->skc_waitq);
}
spin_unlock(&skc->skc_lock);
return (sks == NULL ? -ENOMEM : 0);
}
static void
spl_cache_grow_work(void *data)
{
spl_kmem_alloc_t *ska = (spl_kmem_alloc_t *)data;
spl_kmem_cache_t *skc = ska->ska_cache;
(void) __spl_cache_grow(skc, ska->ska_flags);
atomic_dec(&skc->skc_ref);
smp_mb__before_atomic();
clear_bit(KMC_BIT_GROWING, &skc->skc_flags);
smp_mb__after_atomic();
kfree(ska);
}
/*
* Returns non-zero when a new slab should be available.
*/
static int
spl_cache_grow_wait(spl_kmem_cache_t *skc)
{
return (!test_bit(KMC_BIT_GROWING, &skc->skc_flags));
}
/*
* No available objects on any slabs, create a new slab. Note that this
* functionality is disabled for KMC_SLAB caches which are backed by the
* Linux slab.
*/
static int
spl_cache_grow(spl_kmem_cache_t *skc, int flags, void **obj)
{
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
int remaining, rc = 0;
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
ASSERT0(flags & ~KM_PUBLIC_MASK);
ASSERT(skc->skc_magic == SKC_MAGIC);
ASSERT((skc->skc_flags & KMC_SLAB) == 0);
might_sleep();
*obj = NULL;
/*
* Before allocating a new slab wait for any reaping to complete and
* then return so the local magazine can be rechecked for new objects.
*/
if (test_bit(KMC_BIT_REAPING, &skc->skc_flags)) {
rc = spl_wait_on_bit(&skc->skc_flags, KMC_BIT_REAPING,
TASK_UNINTERRUPTIBLE);
return (rc ? rc : -EAGAIN);
}
/*
* To reduce the overhead of context switch and improve NUMA locality,
* it tries to allocate a new slab in the current process context with
* KM_NOSLEEP flag. If it fails, it will launch a new taskq to do the
* allocation.
*
* However, this can't be applied to KVM_VMEM due to a bug that
* __vmalloc() doesn't honor gfp flags in page table allocation.
*/
if (!(skc->skc_flags & KMC_VMEM)) {
rc = __spl_cache_grow(skc, flags | KM_NOSLEEP);
if (rc == 0)
return (0);
}
/*
* This is handled by dispatching a work request to the global work
* queue. This allows us to asynchronously allocate a new slab while
* retaining the ability to safely fall back to a smaller synchronous
* allocations to ensure forward progress is always maintained.
*/
if (test_and_set_bit(KMC_BIT_GROWING, &skc->skc_flags) == 0) {
spl_kmem_alloc_t *ska;
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
ska = kmalloc(sizeof (*ska), kmem_flags_convert(flags));
if (ska == NULL) {
clear_bit_unlock(KMC_BIT_GROWING, &skc->skc_flags);
smp_mb__after_atomic();
wake_up_all(&skc->skc_waitq);
return (-ENOMEM);
}
atomic_inc(&skc->skc_ref);
ska->ska_cache = skc;
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
ska->ska_flags = flags;
taskq_init_ent(&ska->ska_tqe);
taskq_dispatch_ent(spl_kmem_cache_taskq,
spl_cache_grow_work, ska, 0, &ska->ska_tqe);
}
/*
* The goal here is to only detect the rare case where a virtual slab
* allocation has deadlocked. We must be careful to minimize the use
* of emergency objects which are more expensive to track. Therefore,
* we set a very long timeout for the asynchronous allocation and if
* the timeout is reached the cache is flagged as deadlocked. From
* this point only new emergency objects will be allocated until the
* asynchronous allocation completes and clears the deadlocked flag.
*/
if (test_bit(KMC_BIT_DEADLOCKED, &skc->skc_flags)) {
rc = spl_emergency_alloc(skc, flags, obj);
} else {
remaining = wait_event_timeout(skc->skc_waitq,
spl_cache_grow_wait(skc), HZ / 10);
Fix kmem cache deadlock logic The kmem cache implementation always adds new slabs by dispatching a task to the spl_kmem_cache taskq to perform the allocation. This is done because large slabs must be allocated using vmalloc(). It is possible these allocations will block on IO because the GFP_NOIO flag is not honored. This can result in a deadlock. Therefore, a deadlock detection strategy was implemented to deal with this case. When it is determined, by timeout, that the spl_kmem_cache thread has deadlocked attempting to add a new slab. Then all callers attempting to allocate from the cache fall back to using kmalloc() which does honor all passed flags. This logic was correct but an optimization in the code allowed for a deadlock. Because only slabs backed by vmalloc() can deadlock in the way described above. An optimization was made to only invoke this deadlock detection code for vmalloc() backed caches. This had the advantage of making it easy to distinguish these objects when they were freed. But this isn't strictly safe. If all the spl_kmem_cache threads end up deadlocked than we can't grow any of the other caches either. This can once again result in a deadlock if memory needs to be allocated from one of these other caches to ensure forward progress. The fix here is to remove the optimization which limits this fall back allocation stratagy to vmalloc() backed caches. Doing this means we may need to take the cache lock in spl_kmem_cache_free() call path. But this small cost can be mitigated by ignoring objects with virtual addresses. For good measure the default number of spl_kmem_cache threads has been increased from 1 to 4, and made tunable. This alone wouldn't resolve the original issue since it's still possible for all the threads to be deadlocked. However, it does help responsiveness by ensuring that a single deadlocked spl_kmem_cache thread doesn't block allocations from other caches until the timeout is reached. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2015-01-09 22:00:34 +00:00
if (!remaining) {
spin_lock(&skc->skc_lock);
if (test_bit(KMC_BIT_GROWING, &skc->skc_flags)) {
set_bit(KMC_BIT_DEADLOCKED, &skc->skc_flags);
skc->skc_obj_deadlock++;
}
spin_unlock(&skc->skc_lock);
}
rc = -ENOMEM;
}
return (rc);
}
/*
* Refill a per-cpu magazine with objects from the slabs for this cache.
* Ideally the magazine can be repopulated using existing objects which have
* been released, however if we are unable to locate enough free objects new
* slabs of objects will be created. On success NULL is returned, otherwise
* the address of a single emergency object is returned for use by the caller.
*/
static void *
spl_cache_refill(spl_kmem_cache_t *skc, spl_kmem_magazine_t *skm, int flags)
{
spl_kmem_slab_t *sks;
int count = 0, rc, refill;
void *obj = NULL;
ASSERT(skc->skc_magic == SKC_MAGIC);
ASSERT(skm->skm_magic == SKM_MAGIC);
refill = MIN(skm->skm_refill, skm->skm_size - skm->skm_avail);
spin_lock(&skc->skc_lock);
while (refill > 0) {
/* No slabs available we may need to grow the cache */
if (list_empty(&skc->skc_partial_list)) {
spin_unlock(&skc->skc_lock);
local_irq_enable();
rc = spl_cache_grow(skc, flags, &obj);
local_irq_disable();
/* Emergency object for immediate use by caller */
if (rc == 0 && obj != NULL)
return (obj);
if (rc)
goto out;
/* Rescheduled to different CPU skm is not local */
if (skm != skc->skc_mag[smp_processor_id()])
goto out;
/*
* Potentially rescheduled to the same CPU but
* allocations may have occurred from this CPU while
* we were sleeping so recalculate max refill.
*/
refill = MIN(refill, skm->skm_size - skm->skm_avail);
spin_lock(&skc->skc_lock);
continue;
}
/* Grab the next available slab */
sks = list_entry((&skc->skc_partial_list)->next,
spl_kmem_slab_t, sks_list);
ASSERT(sks->sks_magic == SKS_MAGIC);
ASSERT(sks->sks_ref < sks->sks_objs);
ASSERT(!list_empty(&sks->sks_free_list));
/*
* Consume as many objects as needed to refill the requested
* cache. We must also be careful not to overfill it.
*/
while (sks->sks_ref < sks->sks_objs && refill-- > 0 &&
++count) {
ASSERT(skm->skm_avail < skm->skm_size);
ASSERT(count < skm->skm_size);
skm->skm_objs[skm->skm_avail++] =
spl_cache_obj(skc, sks);
}
/* Move slab to skc_complete_list when full */
if (sks->sks_ref == sks->sks_objs) {
list_del(&sks->sks_list);
list_add(&sks->sks_list, &skc->skc_complete_list);
}
}
spin_unlock(&skc->skc_lock);
out:
return (NULL);
}
/*
* Release an object back to the slab from which it came.
*/
static void
spl_cache_shrink(spl_kmem_cache_t *skc, void *obj)
{
spl_kmem_slab_t *sks = NULL;
spl_kmem_obj_t *sko = NULL;
ASSERT(skc->skc_magic == SKC_MAGIC);
sko = spl_sko_from_obj(skc, obj);
ASSERT(sko->sko_magic == SKO_MAGIC);
sks = sko->sko_slab;
ASSERT(sks->sks_magic == SKS_MAGIC);
ASSERT(sks->sks_cache == skc);
list_add(&sko->sko_list, &sks->sks_free_list);
sks->sks_age = jiffies;
sks->sks_ref--;
skc->skc_obj_alloc--;
/*
* Move slab to skc_partial_list when no longer full. Slabs
* are added to the head to keep the partial list is quasi-full
* sorted order. Fuller at the head, emptier at the tail.
*/
if (sks->sks_ref == (sks->sks_objs - 1)) {
list_del(&sks->sks_list);
list_add(&sks->sks_list, &skc->skc_partial_list);
}
/*
* Move empty slabs to the end of the partial list so
* they can be easily found and freed during reclamation.
*/
if (sks->sks_ref == 0) {
list_del(&sks->sks_list);
list_add_tail(&sks->sks_list, &skc->skc_partial_list);
skc->skc_slab_alloc--;
}
}
/*
* Allocate an object from the per-cpu magazine, or if the magazine
* is empty directly allocate from a slab and repopulate the magazine.
*/
void *
spl_kmem_cache_alloc(spl_kmem_cache_t *skc, int flags)
{
spl_kmem_magazine_t *skm;
void *obj = NULL;
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
ASSERT0(flags & ~KM_PUBLIC_MASK);
ASSERT(skc->skc_magic == SKC_MAGIC);
ASSERT(!test_bit(KMC_BIT_DESTROY, &skc->skc_flags));
/*
* Allocate directly from a Linux slab. All optimizations are left
* to the underlying cache we only need to guarantee that KM_SLEEP
* callers will never fail.
*/
if (skc->skc_flags & KMC_SLAB) {
struct kmem_cache *slc = skc->skc_linux_cache;
do {
Refactor generic memory allocation interfaces This patch achieves the following goals: 1. It replaces the preprocessor kmem flag to gfp flag mapping with proper translation logic. This eliminates the potential for surprises that were previously possible where kmem flags were mapped to gfp flags. 2. It maps vmem_alloc() allocations to kmem_alloc() for allocations sized less than or equal to the newly-added spl_kmem_alloc_max parameter. This ensures that small allocations will not contend on a single global lock, large allocations can still be handled, and potentially limited virtual address space will not be squandered. This behavior is entirely different than under Illumos due to different memory management strategies employed by the respective kernels. However, this functionally provides the semantics required. 3. The --disable-debug-kmem, --enable-debug-kmem (default), and --enable-debug-kmem-tracking allocators have been unified in to a single spl_kmem_alloc_impl() allocation function. This was done to simplify the code and make it more maintainable. 4. Improve portability by exposing an implementation of the memory allocations functions that can be safely used in the same way they are used on Illumos. Specifically, callers may safely use KM_SLEEP in contexts which perform filesystem IO. This allows us to eliminate an entire class of Linux specific changes which were previously required to avoid deadlocking the system. This change will be largely transparent to existing callers but there are a few caveats: 1. Because the headers were refactored and extraneous includes removed callers may find they need to explicitly add additional #includes. In particular, kmem_cache.h must now be explicitly includes to access the SPL's kmem cache implementation. This behavior is different from Illumos but it was done to avoid always masking the Linux slab functions when kmem.h is included. 2. Callers, like Lustre, which made assumptions about the definitions of KM_SLEEP, KM_NOSLEEP, and KM_PUSHPAGE will need to be updated. Other callers such as ZFS which did not will not require changes. 3. KM_PUSHPAGE is no longer overloaded to imply GFP_NOIO. It retains its original meaning of allowing allocations to access reserved memory. KM_PUSHPAGE callers can be converted back to KM_SLEEP. 4. The KM_NODEBUG flags has been retired and the default warning threshold increased to 32k. 5. The kmem_virt() functions has been removed. For callers which need to distinguish between a physical and virtual address use is_vmalloc_addr(). Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-08 20:37:14 +00:00
obj = kmem_cache_alloc(slc, kmem_flags_convert(flags));
} while ((obj == NULL) && !(flags & KM_NOSLEEP));
2019-10-18 17:24:28 +00:00
if (obj != NULL) {
/*
* Even though we leave everything up to the
* underlying cache we still keep track of
* how many objects we've allocated in it for
* better debuggability.
*/
spin_lock(&skc->skc_lock);
skc->skc_obj_alloc++;
spin_unlock(&skc->skc_lock);
}
goto ret;
}
local_irq_disable();
restart:
/*
* Safe to update per-cpu structure without lock, but
* in the restart case we must be careful to reacquire
* the local magazine since this may have changed
* when we need to grow the cache.
*/
skm = skc->skc_mag[smp_processor_id()];
ASSERT(skm->skm_magic == SKM_MAGIC);
if (likely(skm->skm_avail)) {
/* Object available in CPU cache, use it */
obj = skm->skm_objs[--skm->skm_avail];
skm->skm_age = jiffies;
} else {
obj = spl_cache_refill(skc, skm, flags);
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
if ((obj == NULL) && !(flags & KM_NOSLEEP))
goto restart;
Refine slab cache sizing This change is designed to improve the memory utilization of slabs by more carefully setting their size. The way the code currently works is problematic for slabs which contain large objects (>1MB). This is due to slabs being unconditionally rounded up to a power of two which may result in unused space at the end of the slab. The reason the existing code rounds up every slab is because it assumes it will backed by the buddy allocator. Since the buddy allocator can only performs power of two allocations this is desirable because it avoids wasting any space. However, this logic breaks down if slab is backed by vmalloc() which operates at a page level granularity. In this case, the optimal thing to do is calculate the minimum required slab size given certain constraints (object size, alignment, objects/slab, etc). Therefore, this patch reworks the spl_slab_size() function so that it sizes KMC_KMEM slabs differently than KMC_VMEM slabs. KMC_KMEM slabs are rounded up to the nearest power of two, and KMC_VMEM slabs are allowed to be the minimum required size. This change also reduces the default number of objects per slab. This reduces how much memory a single cache object can pin, which can result in significant memory saving for highly fragmented caches. But depending on the workload it may result in slabs being allocated and freed more frequently. In practice, this has been shown to be a better default for most workloads. Also the maximum slab size has been reduced to 4MB on 32-bit systems. Due to the limited virtual address space it's critical the we be as frugal as possible. A limit of 4M still lets us reasonably comfortably allocate a limited number of 1MB objects. Finally, the kmem:slab_small and kmem:slab_large SPLAT tests were extended to provide better test coverage of various object sizes and alignments. Caches are created with random parameters and their basic functionality is verified by allocating several slabs worth of objects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-12-15 22:06:18 +00:00
local_irq_enable();
goto ret;
}
local_irq_enable();
ASSERT(obj);
ASSERT(IS_P2ALIGNED(obj, skc->skc_obj_align));
ret:
/* Pre-emptively migrate object to CPU L1 cache */
if (obj) {
if (obj && skc->skc_ctor)
skc->skc_ctor(obj, skc->skc_private, flags);
else
prefetchw(obj);
}
return (obj);
}
EXPORT_SYMBOL(spl_kmem_cache_alloc);
/*
* Free an object back to the local per-cpu magazine, there is no
* guarantee that this is the same magazine the object was originally
* allocated from. We may need to flush entire from the magazine
* back to the slabs to make space.
*/
void
spl_kmem_cache_free(spl_kmem_cache_t *skc, void *obj)
{
spl_kmem_magazine_t *skm;
unsigned long flags;
int do_reclaim = 0;
Fix kmem cache deadlock logic The kmem cache implementation always adds new slabs by dispatching a task to the spl_kmem_cache taskq to perform the allocation. This is done because large slabs must be allocated using vmalloc(). It is possible these allocations will block on IO because the GFP_NOIO flag is not honored. This can result in a deadlock. Therefore, a deadlock detection strategy was implemented to deal with this case. When it is determined, by timeout, that the spl_kmem_cache thread has deadlocked attempting to add a new slab. Then all callers attempting to allocate from the cache fall back to using kmalloc() which does honor all passed flags. This logic was correct but an optimization in the code allowed for a deadlock. Because only slabs backed by vmalloc() can deadlock in the way described above. An optimization was made to only invoke this deadlock detection code for vmalloc() backed caches. This had the advantage of making it easy to distinguish these objects when they were freed. But this isn't strictly safe. If all the spl_kmem_cache threads end up deadlocked than we can't grow any of the other caches either. This can once again result in a deadlock if memory needs to be allocated from one of these other caches to ensure forward progress. The fix here is to remove the optimization which limits this fall back allocation stratagy to vmalloc() backed caches. Doing this means we may need to take the cache lock in spl_kmem_cache_free() call path. But this small cost can be mitigated by ignoring objects with virtual addresses. For good measure the default number of spl_kmem_cache threads has been increased from 1 to 4, and made tunable. This alone wouldn't resolve the original issue since it's still possible for all the threads to be deadlocked. However, it does help responsiveness by ensuring that a single deadlocked spl_kmem_cache thread doesn't block allocations from other caches until the timeout is reached. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2015-01-09 22:00:34 +00:00
int do_emergency = 0;
ASSERT(skc->skc_magic == SKC_MAGIC);
ASSERT(!test_bit(KMC_BIT_DESTROY, &skc->skc_flags));
/*
* Run the destructor
*/
if (skc->skc_dtor)
skc->skc_dtor(obj, skc->skc_private);
/*
* Free the object from the Linux underlying Linux slab.
*/
if (skc->skc_flags & KMC_SLAB) {
kmem_cache_free(skc->skc_linux_cache, obj);
2019-10-18 17:24:28 +00:00
spin_lock(&skc->skc_lock);
skc->skc_obj_alloc--;
spin_unlock(&skc->skc_lock);
return;
}
/*
Fix kmem cache deadlock logic The kmem cache implementation always adds new slabs by dispatching a task to the spl_kmem_cache taskq to perform the allocation. This is done because large slabs must be allocated using vmalloc(). It is possible these allocations will block on IO because the GFP_NOIO flag is not honored. This can result in a deadlock. Therefore, a deadlock detection strategy was implemented to deal with this case. When it is determined, by timeout, that the spl_kmem_cache thread has deadlocked attempting to add a new slab. Then all callers attempting to allocate from the cache fall back to using kmalloc() which does honor all passed flags. This logic was correct but an optimization in the code allowed for a deadlock. Because only slabs backed by vmalloc() can deadlock in the way described above. An optimization was made to only invoke this deadlock detection code for vmalloc() backed caches. This had the advantage of making it easy to distinguish these objects when they were freed. But this isn't strictly safe. If all the spl_kmem_cache threads end up deadlocked than we can't grow any of the other caches either. This can once again result in a deadlock if memory needs to be allocated from one of these other caches to ensure forward progress. The fix here is to remove the optimization which limits this fall back allocation stratagy to vmalloc() backed caches. Doing this means we may need to take the cache lock in spl_kmem_cache_free() call path. But this small cost can be mitigated by ignoring objects with virtual addresses. For good measure the default number of spl_kmem_cache threads has been increased from 1 to 4, and made tunable. This alone wouldn't resolve the original issue since it's still possible for all the threads to be deadlocked. However, it does help responsiveness by ensuring that a single deadlocked spl_kmem_cache thread doesn't block allocations from other caches until the timeout is reached. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2015-01-09 22:00:34 +00:00
* While a cache has outstanding emergency objects all freed objects
* must be checked. However, since emergency objects will never use
* a virtual address these objects can be safely excluded as an
* optimization.
*/
Fix kmem cache deadlock logic The kmem cache implementation always adds new slabs by dispatching a task to the spl_kmem_cache taskq to perform the allocation. This is done because large slabs must be allocated using vmalloc(). It is possible these allocations will block on IO because the GFP_NOIO flag is not honored. This can result in a deadlock. Therefore, a deadlock detection strategy was implemented to deal with this case. When it is determined, by timeout, that the spl_kmem_cache thread has deadlocked attempting to add a new slab. Then all callers attempting to allocate from the cache fall back to using kmalloc() which does honor all passed flags. This logic was correct but an optimization in the code allowed for a deadlock. Because only slabs backed by vmalloc() can deadlock in the way described above. An optimization was made to only invoke this deadlock detection code for vmalloc() backed caches. This had the advantage of making it easy to distinguish these objects when they were freed. But this isn't strictly safe. If all the spl_kmem_cache threads end up deadlocked than we can't grow any of the other caches either. This can once again result in a deadlock if memory needs to be allocated from one of these other caches to ensure forward progress. The fix here is to remove the optimization which limits this fall back allocation stratagy to vmalloc() backed caches. Doing this means we may need to take the cache lock in spl_kmem_cache_free() call path. But this small cost can be mitigated by ignoring objects with virtual addresses. For good measure the default number of spl_kmem_cache threads has been increased from 1 to 4, and made tunable. This alone wouldn't resolve the original issue since it's still possible for all the threads to be deadlocked. However, it does help responsiveness by ensuring that a single deadlocked spl_kmem_cache thread doesn't block allocations from other caches until the timeout is reached. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2015-01-09 22:00:34 +00:00
if (!is_vmalloc_addr(obj)) {
spin_lock(&skc->skc_lock);
do_emergency = (skc->skc_obj_emergency > 0);
spin_unlock(&skc->skc_lock);
if (do_emergency && (spl_emergency_free(skc, obj) == 0))
return;
}
local_irq_save(flags);
/*
* Safe to update per-cpu structure without lock, but
* no remote memory allocation tracking is being performed
* it is entirely possible to allocate an object from one
* CPU cache and return it to another.
*/
skm = skc->skc_mag[smp_processor_id()];
ASSERT(skm->skm_magic == SKM_MAGIC);
/*
* Per-CPU cache full, flush it to make space for this object,
* this may result in an empty slab which can be reclaimed once
* interrupts are re-enabled.
*/
if (unlikely(skm->skm_avail >= skm->skm_size)) {
spl_cache_flush(skc, skm, skm->skm_refill);
do_reclaim = 1;
}
/* Available space in cache, use it */
skm->skm_objs[skm->skm_avail++] = obj;
local_irq_restore(flags);
if (do_reclaim)
spl_slab_reclaim(skc);
}
EXPORT_SYMBOL(spl_kmem_cache_free);
/*
* The generic shrinker function for all caches. Under Linux a shrinker
* may not be tightly coupled with a slab cache. In fact Linux always
* systematically tries calling all registered shrinker callbacks which
* report that they contain unused objects. Because of this we only
* register one shrinker function in the shim layer for all slab caches.
* We always attempt to shrink all caches when this generic shrinker
* is called.
*
* If sc->nr_to_scan is zero, the caller is requesting a query of the
* number of objects which can potentially be freed. If it is nonzero,
* the request is to free that many objects.
*
* Linux kernels >= 3.12 have the count_objects and scan_objects callbacks
* in struct shrinker and also require the shrinker to return the number
* of objects freed.
*
* Older kernels require the shrinker to return the number of freeable
* objects following the freeing of nr_to_free.
*
* Linux semantics differ from those under Solaris, which are to
* free all available objects which may (and probably will) be more
* objects than the requested nr_to_scan.
*/
static spl_shrinker_t
__spl_kmem_cache_generic_shrinker(struct shrinker *shrink,
struct shrink_control *sc)
{
spl_kmem_cache_t *skc;
int alloc = 0;
/*
* No shrinking in a transaction context. Can cause deadlocks.
*/
if (sc->nr_to_scan && spl_fstrans_check())
return (SHRINK_STOP);
down_read(&spl_kmem_cache_sem);
list_for_each_entry(skc, &spl_kmem_cache_list, skc_list) {
if (sc->nr_to_scan) {
#ifdef HAVE_SPLIT_SHRINKER_CALLBACK
uint64_t oldalloc = skc->skc_obj_alloc;
spl_kmem_cache_reap_now(skc,
MAX(sc->nr_to_scan>>fls64(skc->skc_slab_objs), 1));
if (oldalloc > skc->skc_obj_alloc)
alloc += oldalloc - skc->skc_obj_alloc;
#else
spl_kmem_cache_reap_now(skc,
MAX(sc->nr_to_scan>>fls64(skc->skc_slab_objs), 1));
alloc += skc->skc_obj_alloc;
#endif /* HAVE_SPLIT_SHRINKER_CALLBACK */
} else {
/* Request to query number of freeable objects */
alloc += skc->skc_obj_alloc;
}
}
up_read(&spl_kmem_cache_sem);
/*
* When KMC_RECLAIM_ONCE is set allow only a single reclaim pass.
* This functionality only exists to work around a rare issue where
* shrink_slabs() is repeatedly invoked by many cores causing the
* system to thrash.
*/
if ((spl_kmem_cache_reclaim & KMC_RECLAIM_ONCE) && sc->nr_to_scan)
return (SHRINK_STOP);
return (MAX(alloc, 0));
}
SPL_SHRINKER_CALLBACK_WRAPPER(spl_kmem_cache_generic_shrinker);
/*
* Call the registered reclaim function for a cache. Depending on how
* many and which objects are released it may simply repopulate the
* local magazine which will then need to age-out. Objects which cannot
* fit in the magazine we will be released back to their slabs which will
* also need to age out before being release. This is all just best
* effort and we do not want to thrash creating and destroying slabs.
*/
void
spl_kmem_cache_reap_now(spl_kmem_cache_t *skc, int count)
{
ASSERT(skc->skc_magic == SKC_MAGIC);
ASSERT(!test_bit(KMC_BIT_DESTROY, &skc->skc_flags));
atomic_inc(&skc->skc_ref);
/*
* Execute the registered reclaim callback if it exists.
*/
if (skc->skc_flags & KMC_SLAB) {
if (skc->skc_reclaim)
skc->skc_reclaim(skc->skc_private);
goto out;
}
/*
* Prevent concurrent cache reaping when contended.
*/
if (test_and_set_bit(KMC_BIT_REAPING, &skc->skc_flags))
goto out;
/*
* When a reclaim function is available it may be invoked repeatedly
* until at least a single slab can be freed. This ensures that we
* do free memory back to the system. This helps minimize the chance
* of an OOM event when the bulk of memory is used by the slab.
*
* When free slabs are already available the reclaim callback will be
* skipped. Additionally, if no forward progress is detected despite
* a reclaim function the cache will be skipped to avoid deadlock.
*
* Longer term this would be the correct place to add the code which
* repacks the slabs in order minimize fragmentation.
*/
if (skc->skc_reclaim) {
uint64_t objects = UINT64_MAX;
int do_reclaim;
do {
spin_lock(&skc->skc_lock);
do_reclaim =
(skc->skc_slab_total > 0) &&
((skc->skc_slab_total-skc->skc_slab_alloc) == 0) &&
(skc->skc_obj_alloc < objects);
objects = skc->skc_obj_alloc;
spin_unlock(&skc->skc_lock);
if (do_reclaim)
skc->skc_reclaim(skc->skc_private);
} while (do_reclaim);
}
/* Reclaim from the magazine and free all now empty slabs. */
if (spl_kmem_cache_expire & KMC_EXPIRE_MEM) {
spl_kmem_magazine_t *skm;
unsigned long irq_flags;
local_irq_save(irq_flags);
skm = skc->skc_mag[smp_processor_id()];
spl_cache_flush(skc, skm, skm->skm_avail);
local_irq_restore(irq_flags);
}
spl_slab_reclaim(skc);
clear_bit_unlock(KMC_BIT_REAPING, &skc->skc_flags);
smp_mb__after_atomic();
wake_up_bit(&skc->skc_flags, KMC_BIT_REAPING);
out:
atomic_dec(&skc->skc_ref);
}
EXPORT_SYMBOL(spl_kmem_cache_reap_now);
OpenZFS 9284 - arc_reclaim_thread has 2 jobs Following the fix for 9018 (Replace kmem_cache_reap_now() with kmem_cache_reap_soon), the arc_reclaim_thread() no longer blocks while reaping. However, the code is still confusing and error-prone, because this thread has two responsibilities. We should instead separate this into two threads each with their own responsibility: 1. keep `arc_size` under `arc_c`, by calling `arc_adjust()`, which improves `arc_is_overflowing()` 2. keep enough free memory in the system, by calling `arc_kmem_reap_now()` plus `arc_shrink()`, which improves `arc_available_memory()`. Furthermore, we can use the zthr infrastructure to separate the "should we do something" from "do it" parts of the logic, and normalize the start up / shut down of the threads. Authored by: Brad Lewis <brad.lewis@delphix.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Serapheim Dimitropoulos <serapheim@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed by: Dan McDonald <danmcd@joyent.com> Reviewed by: Tim Kordas <tim.kordas@joyent.com> Reviewed by: Tim Chase <tim@chase2k.com> Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: Brad Lewis <brad.lewis@delphix.com> Signed-off-by: Brad Lewis <brad.lewis@delphix.com> OpenZFS-issue: https://www.illumos.org/issues/9284 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/de753e34f9 Closes #8165
2017-03-15 23:41:52 +00:00
/*
* This is stubbed out for code consistency with other platforms. There
* is existing logic to prevent concurrent reaping so while this is ugly
* it should do no harm.
*/
int
spl_kmem_cache_reap_active()
{
return (0);
}
EXPORT_SYMBOL(spl_kmem_cache_reap_active);
/*
* Reap all free slabs from all registered caches.
*/
void
spl_kmem_reap(void)
{
struct shrink_control sc;
sc.nr_to_scan = KMC_REAP_CHUNK;
sc.gfp_mask = GFP_KERNEL;
(void) __spl_kmem_cache_generic_shrinker(NULL, &sc);
}
EXPORT_SYMBOL(spl_kmem_reap);
int
spl_kmem_cache_init(void)
{
init_rwsem(&spl_kmem_cache_sem);
INIT_LIST_HEAD(&spl_kmem_cache_list);
spl_kmem_cache_taskq = taskq_create("spl_kmem_cache",
spl_kmem_cache_kmem_threads, maxclsyspri,
spl_kmem_cache_kmem_threads * 8, INT_MAX,
TASKQ_PREPOPULATE | TASKQ_DYNAMIC);
spl_register_shrinker(&spl_kmem_cache_shrinker);
return (0);
}
void
spl_kmem_cache_fini(void)
{
spl_unregister_shrinker(&spl_kmem_cache_shrinker);
taskq_destroy(spl_kmem_cache_taskq);
}