[PATCH 3/6] compcache: TLSF Allocator interface

[PATCH 3/6] compcache: TLSF Allocator interface

[Nitin Gupta]

Two Level Segregate Fit (TLSF) Allocator is used to allocate memory for
variable size compressed pages. Its fast and gives low fragmentation.
Following links give details on this allocator:
 - http://rtportal.upv.es/rtmalloc/files/tlsf_paper_spe_2007.pdf
 - http://code.google.com/p/compcache/wiki/TLSFAllocator

This kernel port of TLSF (v2.3.2) introduces several changes but underlying
algorithm remains the same.

Changelog TLSF v2.3.2 vs this kernel port
 - Pool now dynamically expands/shrinks.
   It is collection of contiguous memory regions.
 - Changes to pool create interface as a result of above change.
 - Collect and export stats (/proc/tlsfinfo)
 - Cleanups: kernel coding style, added comments, macros -> static inline, etc.

Signed-off-by: Nitin Gupta <nitingupta910 at gmail dot com>
---

 include/linux/tlsf.h |   93 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 93 insertions(+), 0 deletions(-)

diff --git a/include/linux/tlsf.h b/include/linux/tlsf.h
new file mode 100644
index 0000000..ef8092c
--- /dev/null
+++ b/include/linux/tlsf.h
@@ -0,0 +1,93 @@
+/*
+ * Two Levels Segregate Fit memory allocator (TLSF)
+ * Version 2.3.2
+ *
+ * Written by Miguel Masmano Tello <mimastel@doctor.upv.es>
+ *
+ * Thanks to Ismael Ripoll for his suggestions and reviews
+ *
+ * Copyright (C) 2007, 2006, 2005, 2004
+ *
+ * This code is released using a dual license strategy: GPL/LGPL
+ * You can choose the licence that better fits your requirements.
+ *
+ * Released under the terms of the GNU General Public License Version 2.0
+ * Released under the terms of the GNU Lesser General Public License Version 2.1
+ *
+ * This is kernel port of TLSF allocator.
+ * Original code can be found at: http://rtportal.upv.es/rtmalloc/
+ * 	- Nitin Gupta (nitingupta910 at gmail dot com)
+ */
+
+#ifndef _TLSF_H_
+#define _TLSF_H_
+
+typedef void* (get_memory)(size_t bytes);
+typedef void (put_memory)(void *ptr);
+
+/**
+ * tlsf_create_memory_pool - create dynamic memory pool
+ * @name: name of the pool
+ * @get_mem: callback function used to expand pool
+ * @put_mem: callback function used to shrink pool
+ * @init_size: inital pool size (in bytes)
+ * @max_size: maximum pool size (in bytes) - set this as 0 for no limit
+ * @grow_size: amount of memory (in bytes) added to pool whenever required
+ *
+ * All size values are rounded up to next page boundary.
+ */
+extern void *tlsf_create_memory_pool(const char *name,
+					get_memory get_mem,
+					put_memory put_mem,
+					size_t init_size,
+					size_t max_size,
+					size_t grow_size);
+/**
+ * tlsf_destory_memory_pool - cleanup given pool
+ * @mem_pool: Pool to be destroyed
+ *
+ * Data structures associated with pool are freed.
+ * All memory allocated from pool must be freed before
+ * destorying it.
+ */
+extern void tlsf_destroy_memory_pool(void *mem_pool);
+
+/**
+ * tlsf_malloc - allocate memory from given pool
+ * @size: no. of bytes
+ * @mem_pool: pool to allocate from
+ */
+extern void *tlsf_malloc(size_t size, void *mem_pool);
+
+/**
+ * tlsf_calloc - allocate and zero-out memory from given pool
+ * @size: no. of bytes
+ * @mem_pool: pool to allocate from
+ */
+extern void *tlsf_calloc(size_t nelem, size_t elem_size, void *mem_pool);
+
+/**
+ * tlsf_free - free memory from given pool
+ * @ptr: address of memory to be freed
+ * @mem_pool: pool to free from
+ */
+extern void tlsf_free(void *ptr, void *mem_pool);
+
+/**
+ * tlsf_get_used_size - get memory currently used by given pool
+ *
+ * Used memory includes stored data + metadata + internal fragmentation
+ */
+extern size_t tlsf_get_used_size(void *mem_pool);
+
+/**
+ * tlsf_get_total_size - get total memory currently allocated for given pool
+ *
+ * This is the total memory currently allocated for this pool which includes
+ * used size + free size.
+ *
+ * (Total - Used) is good indicator of memory efficiency of allocator.
+ */
+extern size_t tlsf_get_total_size(void *mem_pool);
+
+#endif

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 3/6] compcache: TLSF Allocator interface

[Peter Zijlstra]

On Mon, 2008-03-24 at 20:34 +0530, Nitin Gupta wrote:
> Two Level Segregate Fit (TLSF) Allocator is used to allocate memory for
> variable size compressed pages. Its fast and gives low fragmentation.
> Following links give details on this allocator:
>  - http://rtportal.upv.es/rtmalloc/files/tlsf_paper_spe_2007.pdf
>  - http://code.google.com/p/compcache/wiki/TLSFAllocator
> 
> This kernel port of TLSF (v2.3.2) introduces several changes but underlying
> algorithm remains the same.
> 
> Changelog TLSF v2.3.2 vs this kernel port
>  - Pool now dynamically expands/shrinks.
>    It is collection of contiguous memory regions.
>  - Changes to pool create interface as a result of above change.
>  - Collect and export stats (/proc/tlsfinfo)
>  - Cleanups: kernel coding style, added comments, macros -> static inline, etc.

Can you explain why you need this allocator, why don't the current
kernel allocators work for you?




--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 3/6] compcache: TLSF Allocator interface

[Nitin Gupta]

On Mon, Mar 24, 2008 at 10:26 PM, Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:
> On Mon, 2008-03-24 at 20:34 +0530, Nitin Gupta wrote:
>  > Two Level Segregate Fit (TLSF) Allocator is used to allocate memory for
>  > variable size compressed pages. Its fast and gives low fragmentation.
>  > Following links give details on this allocator:
>  >  - http://rtportal.upv.es/rtmalloc/files/tlsf_paper_spe_2007.pdf
>  >  - http://code.google.com/p/compcache/wiki/TLSFAllocator
>  >
>  > This kernel port of TLSF (v2.3.2) introduces several changes but underlying
>  > algorithm remains the same.
>  >
>  > Changelog TLSF v2.3.2 vs this kernel port
>  >  - Pool now dynamically expands/shrinks.
>  >    It is collection of contiguous memory regions.
>  >  - Changes to pool create interface as a result of above change.
>  >  - Collect and export stats (/proc/tlsfinfo)
>  >  - Cleanups: kernel coding style, added comments, macros -> static inline, etc.
>
>  Can you explain why you need this allocator, why don't the current
>  kernel allocators work for you?
>
>

kmalloc() allocates one of pre-defined sizes (as defined in
kmalloc_sizes.h). This will surely cause severe fragmentation with
these variable sized compressed pages.

Whereas, TLSF maintains very fine grained size lists. In all the
workloads I tested, it showed <5% fragmentation. Also, its very simple
as just ~700 LOC.

- Nitin

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 3/6] compcache: TLSF Allocator interface

[Peter Zijlstra]

On Mon, 2008-03-24 at 23:04 +0530, Nitin Gupta wrote:
> On Mon, Mar 24, 2008 at 10:26 PM, Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:
> > On Mon, 2008-03-24 at 20:34 +0530, Nitin Gupta wrote:
> >  > Two Level Segregate Fit (TLSF) Allocator is used to allocate memory for
> >  > variable size compressed pages. Its fast and gives low fragmentation.
> >  > Following links give details on this allocator:
> >  >  - http://rtportal.upv.es/rtmalloc/files/tlsf_paper_spe_2007.pdf
> >  >  - http://code.google.com/p/compcache/wiki/TLSFAllocator
> >  >
> >  > This kernel port of TLSF (v2.3.2) introduces several changes but underlying
> >  > algorithm remains the same.
> >  >
> >  > Changelog TLSF v2.3.2 vs this kernel port
> >  >  - Pool now dynamically expands/shrinks.
> >  >    It is collection of contiguous memory regions.
> >  >  - Changes to pool create interface as a result of above change.
> >  >  - Collect and export stats (/proc/tlsfinfo)
> >  >  - Cleanups: kernel coding style, added comments, macros -> static inline, etc.
> >
> >  Can you explain why you need this allocator, why don't the current
> >  kernel allocators work for you?
> >
> >
> 
> kmalloc() allocates one of pre-defined sizes (as defined in
> kmalloc_sizes.h). This will surely cause severe fragmentation with
> these variable sized compressed pages.
> 
> Whereas, TLSF maintains very fine grained size lists. In all the
> workloads I tested, it showed <5% fragmentation. Also, its very simple
> as just ~700 LOC.

Yeah, it also suffers from a horrible coding style, can use excessive
amounts of vmalloc space, isn't hooked into the reclaim process as an
allocator should be and has a severe lack of per-cpu data making it a
pretty big bottleneck on anything with more than a few cores.

Now, it might be needed, might work better, and the scalability issue
might not be a problem when used for swap, but still, you don't treat
any of these points in your changelog.

FWIW, please split up the patches in a sane way. This series looks like
it wants to be 2 or 3 patches. The first introducing all of TLSF (this
split per file is horrible). The second doing all of the block device,
and a possible last doing documentation and such.

Also, how bad was kmalloc() compared to this TLSF, we need numbers :-)

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 3/6] compcache: TLSF Allocator interface

[Nitin Gupta]

On Tue, Mar 25, 2008 at 12:26 AM, Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:
>
> On Mon, 2008-03-24 at 23:04 +0530, Nitin Gupta wrote:
>  > On Mon, Mar 24, 2008 at 10:26 PM, Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:
>  > > On Mon, 2008-03-24 at 20:34 +0530, Nitin Gupta wrote:
>  > >  > Two Level Segregate Fit (TLSF) Allocator is used to allocate memory for
>  > >  > variable size compressed pages. Its fast and gives low fragmentation.
>  > >  > Following links give details on this allocator:
>  > >  >  - http://rtportal.upv.es/rtmalloc/files/tlsf_paper_spe_2007.pdf
>  > >  >  - http://code.google.com/p/compcache/wiki/TLSFAllocator
>  > >  >
>  > >  > This kernel port of TLSF (v2.3.2) introduces several changes but underlying
>  > >  > algorithm remains the same.
>  > >  >
>  > >  > Changelog TLSF v2.3.2 vs this kernel port
>  > >  >  - Pool now dynamically expands/shrinks.
>  > >  >    It is collection of contiguous memory regions.
>  > >  >  - Changes to pool create interface as a result of above change.
>  > >  >  - Collect and export stats (/proc/tlsfinfo)
>  > >  >  - Cleanups: kernel coding style, added comments, macros -> static inline, etc.
>  > >
>  > >  Can you explain why you need this allocator, why don't the current
>  > >  kernel allocators work for you?
>  > >
>  > >
>  >
>  > kmalloc() allocates one of pre-defined sizes (as defined in
>  > kmalloc_sizes.h). This will surely cause severe fragmentation with
>  > these variable sized compressed pages.
>  >
>  > Whereas, TLSF maintains very fine grained size lists. In all the
>  > workloads I tested, it showed <5% fragmentation. Also, its very simple
>  > as just ~700 LOC.
>
>  Yeah, it also suffers from a horrible coding style, can use excessive
>  amounts of vmalloc space, isn't hooked into the reclaim process as an
>  allocator should be and has a severe lack of per-cpu data making it a
>  pretty big bottleneck on anything with more than a few cores.
>
>  Now, it might be needed, might work better, and the scalability issue
>  might not be a problem when used for swap, but still, you don't treat
>  any of these points in your changelog.

Currently, this TLSF implementation is not scalable at all (and thats
why it depends on EMBEDDED).

>
>  FWIW, please split up the patches in a sane way. This series looks like
>  it wants to be 2 or 3 patches. The first introducing all of TLSF (this
>  split per file is horrible). The second doing all of the block device,
>  and a possible last doing documentation and such.
>
>  Also, how bad was kmalloc() compared to this TLSF, we need numbers :-)
>
>

Ok, I will get them and present here.

Thanks,
Nitin

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 3/6] compcache: TLSF Allocator interface

[Nitin Gupta]

On Tue, Mar 25, 2008 at 12:26 AM, Peter Zijlstra <a.p.zijlstra@chello.nl> wrote:

>  Yeah, it also suffers from a horrible coding style, can use excessive
>  amounts of vmalloc space, isn't hooked into the reclaim process as an
>  allocator should be and has a severe lack of per-cpu data making it a
>  pretty big bottleneck on anything with more than a few cores.
>
>  Now, it might be needed, might work better, and the scalability issue
>  might not be a problem when used for swap, but still, you don't treat
>  any of these points in your changelog.
>


I will add these points to changelog.
This project is meant for small systems only. So, scalability is not an issue.


>  FWIW, please split up the patches in a sane way. This series looks like
>  it wants to be 2 or 3 patches. The first introducing all of TLSF (this
>  split per file is horrible). The second doing all of the block device,
>  and a possible last doing documentation and such.
>

Ok. I will resend with better splitting.


>  Also, how bad was kmalloc() compared to this TLSF, we need numbers :-)
>
>

I have posted performance numbers at:
http://code.google.com/p/compcache/wiki/AllocatorsComparison

Data Summary:

Peak Memory Usage:

    * Ideal: 24947 KB
    * TLSF: 25377 KB
    * KMalloc(SLUB): 36483 KB

So, KMalloc uses ~43% more memory than TLSF!


- Nitin