After the news that SourceForge as started hijacking the downloads of
various projects they deem to be abandoned I have decided to move
libhugetlbfs to GitHub for source code hosting and librelist.com for the
devel mailing list. I will leave the release tarballs in place for now,
but will prob
On Sat, 07 Feb 2015, Mike Kravetz wrote:
> Using hugeadm to adjust pool sizes with negative memsizes specified using
> the format "-memsize" will not work properly (for example -100M).
> Make sure only positive values are passed to size_to_smaller_unit().
>
> Signed-off-by: Mike Kravetz
Applied
On Tue, 09 Dec 2014, Kevin O'Gorman wrote:
> I'm running a fairly recent Xubuntu, with 3.13 Linux kernel on AMD64
> hardware. Although I'm expecting to run mostly 64-bit executables, I'm a
> bit unnerved to find 5 of the 32-bit tests failing.
>
> Some peculiarities may explain this. Or not.
> -
On Thu, 06 Nov 2014, Gary S. Robertson wrote:
> From: "Gary S. Robertson"
>
> The makefile only recognized i386 or x86_64 arches. Added support
> to recognize i486, i586, i686.
>
> Signed-off-by: Gary S. Robertson
Applied, thanks
Eric
signature.asc
Description: Digital signature
--
On Tue, 25 Nov 2014, Petr Holasek wrote:
> zero_filesize_segment tescase fails on s390x with segfault because program
> header virtual address is set to very low value below zSeries default
> 0x8000 address.
>
> I haven't found a way how to set custom virtual base address on s390x since ld
>
On Fri, 05 Dec 2014, Jan Stancek wrote:
> The low and high limit passed to vm_unmapped_area() in
> hugetlb_get_unmapped_area() is (TASK_UNMAPPED_BASE, TASK_SIZE).
>
> On 64-bit kernel this is defined as:
> #define TASK_UNMAPPED_BASE (test_thread_flag(TIF_31BIT) ? \
>
On Thu, 06 Nov 2014, Gary S. Robertson wrote:
> From: "Gary S. Robertson"
>
> The makefile only recognized i386 or x86_64 arches. Added support
> to recognize i486, i586, i686.
>
> Signed-off-by: Gary S. Robertson
> ---
> Makefile |2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
On Thu, 06 Nov 2014, Gary S. Robertson wrote:
> From: "Gary S. Robertson"
>
> There's 3 issues fixed:
> * typo on MB variable (MB -> $MB)
> * some linker variants are missing (linux and big endian)
> - armelfb_linux_eabi
> - aarch64elfb
> - aarch64linux
> - aarch64linuxb
> * text segment
On Wed, 05 Nov 2014, Jan Stancek wrote:
> libhugetlbfs.so is not linked with elf64lppc.o on ppc64le,
> which makes plt_extrasz() always return 0.
>
> This is causing failures (SIGSEGV) in these tests:
> HUGETLB_ELFMAP=W linkhuge_rw (16M: 64):
> HUGETLB_ELFMAP=RW linkhuge_rw (16M: 64):
> HUG
On Thu, 04 Sep 2014, Dan Horák wrote:
> Distributions usually build their packages natively in pure 32-bit or 64-bit
> environments.
>
> original patch by Anton Arapov (Red Hat)
>
> Signed-off-by: Dan Horák
Applied, thanks
Eric
signature.asc
Description: Digital signature
--
On Tue, 19 Aug 2014, Haren Myneni wrote:
>
> Tests: Remove duplicate mremap-* tests
>
> mremap-expand-slice-collision, mremap-fixed-normal-near-huge and
> mremap-fixed-huge-near-normal tests are listed twice. So removing
> the duplicate entries.
>
> Signed-off-by: Haren Myneni
>
>
Applied,
On Mon, 11 Aug 2014, Gary S. Robertson wrote:
> From: "Gary S. Robertson"
>
> This patch applies the same technique used by Koen Kool in the following patch
> which was accepted by the libhugetlbfs project:
>
> [0a4f6] Add aarch64_be_support 2014-03-31 10:52:37
>
> It modifies the libh
On Tue, 10 Sep 2013, Steve Capper wrote:
> Hello,
> This series adds Aarch64 support and makes some minor tweaks.
>
> The first two patches of this series add Aarch64 support to
> libhugetlbfs. (Starting from 3.11-rc1, the Linux Kernel supports
> HugeTLB and THP for ARM64).
>
> Some general chan
On 2013-08-12 06:29, Dan Horák wrote:
> On Sun, 11 Aug 2013 18:21:31 -0400
> Eric B Munson wrote:
>
>> On Thu, 25 Jul 2013, Dan Horák wrote:
>>
>> > Distributions usually build their packages natively in pure 32-bit
>> > or 64-bit environments.
>>
On Tue, 06 Aug 2013, Chunrong Guo wrote:
> Renamed ARCH check to use powerpc (for ppc32) and powerpc64.
>
> Signed-off-by: Kumar Gala
> Signed-off-by: Chunrong Guo
Applied, thanks.
Eric
signature.asc
Description: Digital signature
-
On Tue, 06 Aug 2013, Chunrong Guo wrote:
> Signed-off-by: Yu Zongchun
> Signed-off-by: Chunrong Guo
Applied, thanks.
Eric
signature.asc
Description: Digital signature
--
Get 100% visibility into Java/.NET code with Ap
On Tue, 06 Aug 2013, b40...@freescale.com wrote:
> From: Becky Bruce
>
> One test was being invoked with HUGETLB_MORECORE=es (instead of yes).
>
> Signed-off-by: Becky Bruce
Applied, thanks.
Eric
signature.asc
Description: Digital signature
--
On Tue, 06 Aug 2013, Chunrong Guo wrote:
> Provide an easy means for cross compiling to override the compiler
> without having to know if we intend a 32-bit or 64-bit build.
>
> Signed-off-by: Kumar Gala
> Signed-off-by: Chunrong Guo
Applied, thanks.
Eric
signature.asc
Description: Digital s
On Thu, 25 Jul 2013, Dan Horák wrote:
> Distributions usually build their packages natively in pure 32-bit or 64-bit
> environments.
They might, but this library is built by people other than package
managers. If you would like the ability to build only native
packages, have a look at how ppc64
On Tue, 06 Aug 2013, Chunrong Guo wrote:
> Signed-off-by: Ting Liu
> Signed-off-by: Chunrong Guo
I am not sure why I should take this. An empty leader doesn't help me figure
it out...
> ---
> Makefile | 22 +++---
> 1 files changed, 11 insertions(+), 11 deletions(-)
>
> d
On Tue, 06 Aug 2013, Chunrong Guo wrote:
> The existing library code enforces classic PPC segment restrictons on the
> alignment and allocation of hugepages. BookE Freescale processors
> don't have this restriction and allowing it to remain in place
> is overly restrictive on 32-bit processors.
On Tue, 06 Aug 2013, Chunrong Guo wrote:
> If /etc/mtab is a symlink to /proc/self/mounts,
> mounting fails because addmntent returns an error.
> As a consequence, the Private directory is unusable.
> This is a patch which checks if /etc/mtab is a symlink,
> and if so, doesn't try to call addmnten
On Wed, 31 Jul 2013, Joonsoo Kim wrote:
> Current kernel (v3.10) or older has a problem on no resv mapping.
> If we map a region which is already reserved by other process,
> with MAP_NORESERVE flag, this mapping should be considered reserved
> maping, because allocated page will go into page cach
On Wed, 31 Jul 2013, Joonsoo Kim wrote:
> Current kernel (v3.10) or older has a problem on no resv mapping.
> If there is a region with MAP_SHARED and MAP_NORESERVE, it is considered
> to have reserved pages. So reserved page can be stolen by no resv mapping.
> This test case is to detect this sit
On Wed, 31 Jul 2013, Joonsoo Kim wrote:
> Current kernel (v3.10) or older has a problem in cow optimization. Cow
> optimization is that if we find a page which has only 1 reference from
> this process, we use it directly for private write. This applied to
> anonymous page and a page from page cach
On Tue, 06 Aug 2013, Chunrong Guo wrote:
> Signed-off-by: Zongchun Yu
> Signed-off-by: Chunrong Guo
> ---
The linker scripts for all architectures are deprecated.
--hugetlbfs-align should be used instead. The scripts are still
present for legacy applications that cannot be changed. Because of
On 2013-04-02 05:47, Groden (Jinhu) Li wrote:
> Hi
>
> Libhugetlbfs override the morecore function to implement hugepage
> for
> malloc, but for multi thread version of malloc( PT malloc) ,only the
> main arena use
>
> The morecore ,other arenas use mmap to alloc heap , that mean ,for a
> multith
On 2013-04-02 05:47, Groden (Jinhu) Li wrote:
> Hi
>
> Libhugetlbfs override the morecore function to implement hugepage
> for
> malloc, but for multi thread version of malloc( PT malloc) ,only the
> main arena use
>
> The morecore ,other arenas use mmap to alloc heap , that mean ,for a
> multith
Hi all,
Libhugetlbfs-2.16 has been released and is now available from
sourceforge.net
https://sourceforge.net/projects/libhugetlbfs/files/libhugetlbfs/2.16/
Please see the NEWS file for details on the changes from 2.15
Eric
--
On 2013-03-07 09:41, Gerald Schaefer wrote:
> The strategy to find out TASK_SIZE won't work on s390x anymore,
> starting
> with kernel 3.9. We will dynamically increase the pagetable levels on
> s390x on access beyond TASK_SIZE, effectively increasing TASK_SIZE
> from
> 2^42 to 2^53, but /proc/se
On 2013-03-08 20:26, Dubielzig, Rich wrote:
> Hi,
>
> I know that there was some patch activity on this a few years ago,
> but I just wanted to report that making with parallel jobs is still
> sometimes failing as of 2.14. I have not yet been able to reproduce
> the defect myself, but I have attac
On 2013-02-26 08:33, Gerald Schaefer wrote:
> commit 8c075e33 introduced a regression. The previous sscanf()
> implicitly
> did a null termination on path, while the new code misses the null
> termination. This can result in a failing hugetlbfs_test_path() and
> an
> error message like: "libhuget
On 2013-02-28 09:00, Gerald Schaefer wrote:
> On Tue, 26 Feb 2013 14:33:54 +0100
> Gerald Schaefer wrote:
>
>> commit 8c075e33 introduced a regression. The previous sscanf()
>> implicitly did a null termination on path, while the new code misses
>> the null termination. This can result in a failin
On 2012-12-10 06:01, Jan Stancek wrote:
> mremap-fixed-huge-near-normal testcase mmaps 4*hpage_size -
> getpagesize()
> of anon shared memory. Then it finds area of 3 huge pages aligned to
> hpage_size and munmaps lower and upper areas:
>
>lower +- 3*hpage_size -+ upper
> /
On 2013-02-21 10:57, Steve Capper wrote:
> This patch adds support for the ARM architecture to libhugetlbfs.
>
> We do not add the xBT/xBDT style linker scripts as these have been
> deprecated
> in favour of adjusting the page sizes via command line parameter to
> ld.
>
> Signed-off-by: Steve Cap
On 2013-02-22 05:39, Steve Capper wrote:
> Supply some ARM signal handler code for icache-hygiene, and an ARM
> icache
> flush function.
>
> Signed-off-by: Steve Capper
> ---
Applied, thanks.
Eric
--
Everyone hates slow
Hi all,
libhugetlbfs-2.15 has been released and is now available from sourceforge.net
https://sourceforge.net/projects/libhugetlbfs/files/libhugetlbfs/2.15/
Please see the NEWS file for details on the changes from 2.14
Eric
signature.asc
Description: Digital signature
-
On Fri, 07 Dec 2012, Kokron, Daniel S. (GSFC-610.1)[Computer Sciences
Corporation] wrote:
> This patch resolves clipped LD_LIBRARY_PATH. Using hugectl to enable use of
> hugepage for an application on a system with SLES11 SP1. hugectl was
> clipping my LD_LIBRARY_PATH at 254 characters and pr
On Sat, 01 Dec 2012, Eric B Munson wrote:
> This reverts commit cdd2fa336d9980eb268a775b7cf9c49e3b698cd7.
>
> Conflicts:
> tests/hugetests.h
> tests/testutils.c
>
> With the removal of sscanf from the setup routines, the malloc tests
> now pass without havi
On Sat, 01 Dec 2012, Eric B Munson wrote:
> commit cdd2fa33 makes th library setup exhaust any heap allocated on
> small pages. This covers but does not address the cause of the
> allocation. As mentioned by Steve Capper, sscanf is the cause of the
> allocation. This patch removes
On 2012-12-07 10:14, Kokron, Daniel S. (GSFC-610.1)[Computer Sciences
Corporation] wrote:
> Eric,
>
> I was using hugectl to enable use of hugepage for an application on a
> system with SLES11 SP1. hugectl was clipping my LD_LIBRARY_PATH at
> 254 characters and preventing execution. NAME_MAX on
On 2012-12-06 15:09, Dan Kokron wrote:
> git diff hugectl.c
> diff --git a/hugectl.c b/hugectl.c
> index 741247f..8f05595 100644
> --- a/hugectl.c
> +++ b/hugectl.c
> @@ -271,7 +271,7 @@ void setup_mappings(int count)
>
> void library_path(char *path)
> {
> - char val[NAME_MAX] = "";
> +
prior to replacing _morecore.
Signed-off-by: Eric B Munson
Cc: Steve Capper
Cc: Jan Stancek
---
hugeutils.c | 31 +++---
kernel-features.c | 64 ++---
2 files changed, 64 insertions(+), 31 deletions(-)
diff --git a
This reverts commit cdd2fa336d9980eb268a775b7cf9c49e3b698cd7.
Conflicts:
tests/hugetests.h
tests/testutils.c
With the removal of sscanf from the setup routines, the malloc tests
now pass without having to exhaust small page allocations.
Signed-off-by: Eric B Munson
Cc: Jan
On 2012-11-15 14:44, telenn barz wrote:
> On Thu, Nov 15, 2012 at 3:34 PM, Eric B Munson [1]>
> wrote:
>
>> On 2012-11-15 03:42, telenn barz wrote:
>>
>>> Making some tests with libhugetlbfs on kernel 2.6.34, backing
>>> automatically the heap, we r
On 2012-11-15 03:42, telenn barz wrote:
> Making some tests with libhugetlbfs on kernel 2.6.34, backing
> automatically the heap, we realized that the fallback mechanism is
> all-or-nothing : if theres not enough hugepages, the whole asked
> memory is backed by base pages.
>
> We expected libhugetl
On Thu, 08 Nov 2012, Gerald Schaefer wrote:
> This patch series adds support for linking libhugetlbfs with the
> "--hugetlbfs-align" option on s390. The deprecated "--hugetlbfs-link"
> option will not be supported on s390, so no custom linker scripts are
> provided. To allow building w/o linker sc
On Thu, 08 Nov 2012, Gerald Schaefer wrote:
> This patch fixes a warning about an empty address range that will be issued
> when the offset in prepare_segment() is smaller than page_size. In that case,
> check_range_empty() would be called with len == 0 (the page_size down-aligned
> offset), trigg
On Mon, 10 Sep 2012, Jan Stancek wrote:
> This testcase failed on some setups (ppc64):
> Starting testcase "./obj64/mremap-expand-slice-collision", pid 10840
> do_readback(0x2000100, 0x100, "huge above")
> do_readback(0x1ff, 0x1, "normal below")
> Attempting to remap...
On Mon, 01 Oct 2012, Jan Stancek wrote:
> mremap-expand-slice-collision mmaps page of shared anon memory,
> which is then passed to mremap to extend this mapping to two pages:
> mremap(q, page_size, 2*page_size, 0);
>
> The mapping is extended, but backing shmem file is not, so accessing
> seco
On Fri, 14 Sep 2012, Jan Stancek wrote:
> huge_at_4GB_normal_below, huge_below_4GB_normal_above, straddle_4GB
> are tests to test corner cases on powerpc. powerpc manages memory
> using slices: low (0-4G) and high (4G-).
>
> These tests are using MAP_FIXED and try to mmap hugepage at various
> lo
On Wed, 12 Sep 2012, Jan Stancek wrote:
> huge_at_4GB_normal_below, huge_below_4GB_normal_above, straddle_4GB
> are tests to test corner cases on powerpc. powerpc manages memory
> using slices: low (0-4G) and high (4G-).
>
> These tests are using MAP_FIXED and try to mmap hugepage at various
> lo
Announcing libhugetlbfs 2.14. This is mostly a test suite fix release, please
see the NEWS file for more information. The tarball is available at
http://libhugetlbfs.sourceforge.net
Thanks,
Eric
signature.asc
Description: Digital signature
--
On Fri, 07 Sep 2012, Jan Stancek wrote:
> What is fake on one platform can be actual page_size
> on some other platform. This testcase failed on ppc64
> with 64K pagesize:
> gethugepagesizes (16M: 32):FAILLine 397: Duplicate size 65536 at 0/2
> gethugepagesizes (16M: 64):FAILLine 3
On Fri, 07 Sep 2012, Jan Stancek wrote:
> stack_grow_into_huge allocates large chunks
> on memory and do_child() will eventually fail with SIGSEGV.
>
> If system is set up to dump core, it can take significant
> time until this action is completed. For example if ABRT
> runs on such system, kerne
On Thu, 06 Sep 2012, Jan Stancek wrote:
> First allocation in malloc tests is just 1024 bytes.
> If there is enough free mem in malloc arenas, this first
> allocation can be satisfied without call to MORECORE()
> and testcase will fail.
>
> LD_PRELOAD=libhugetlbfs.so HUGETLB_MORECORE=yes malloc (
On Thu, 06 Sep 2012, Jan Stancek wrote:
> do_readahead() requires a_ops->readpage to be != NULL:
> 568 static ssize_t
> 569 do_readahead(struct address_space *mapping, struct file *filp,
> 570 pgoff_t index, unsigned long nr)
> 571 {
> 572 if (!mapping || !mapping->a_ops || !m
On 2012-07-09 00:47, Sebastian Kuzminsky wrote:
> Hi folks, Im playing around with huge pages on x86-64, and Im running
> into behavior Im not understanding.
>
> Im on Linux 3.2 (Ubuntu Precise), with libhugetlbfs 2.11.
>
> Heres a representative part of my /proc/cpuinfo, particularly note
> that t
On Tue, 12 Jun 2012 21:13:42 -0400, Josh Boyer wrote:
> On Tue, Jun 12, 2012 at 5:18 PM, Eric B Munson
> wrote:
>> On Tue, 12 Jun 2012 15:34:32 -0400, Josh Boyer wrote:
>>>
>>> Hi,
>>>
>>> I've been poking around at the libhugetlbfs testca
On Thu, 14 Jun 2012 17:40:31 +0200, telenn barz wrote:
> Hi all,
>
> This post is a request for clarification on some features of
> libhugetlbfs. I realize that this mailing list is not intended for
> this kind of help demand, but after searching unsuccessfully in the
> mailing list archive, readin
On Fri, 15 Jun 2012 11:48:47 -0400, starlight.201...@binnacle.cx wrote:
> Mostly answered these myself. Posting
> to help future searchers:
>
> (1) yes
>
> (2) yes
>
> In both cases it seems one should always set
> HUGETLB_MORECORE=yes in the environment as
> I don't see how that can be made the d
On Tue, 12 Jun 2012 15:34:32 -0400, Josh Boyer wrote:
> Hi,
>
> I've been poking around at the libhugetlbfs testcases and there are a
> small number of failures that I can't figure out. The primary one is
> the malloc test. From all appearances, it seems that the custom
> hugetlbfs_morecore funct
On Sat, 10 Mar 2012, Eric B Munson wrote:
> If obtain_prepared_file fails for any reason, the last fd in our hugetlb
> segment
> table will be 0 when the error clean up starts because obtain_prepared_file
> does not write to the segment structure in error cases. Because the cleanu
On Mon, 12 Mar 2012, Mel Gorman wrote:
> Map profile events for the i386/core_i7 and add a sensible default for
> i386 in general.
>
> Signed-off-by: Mel Gorman
Applied, thanks.
Eric
signature.asc
Description: Digital signature
-
it will try
to close stdout (fd 0). This patch makes the cleanup for loop decrement the
index before starting to close file descriptors.
Reported-by: Dan Scales
Signed-off-by: Eric B Munson
---
elflink.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/elflink.c b
On Thu, 08 Mar 2012, Andrew Hastings wrote:
> Add missing man pages for ld.hugetlbfs.1, hugetlbfs_find_path.3,
> hugetlbfs_find_path_for_size.3, gethugepagesize.3, hugetlbfs_test_path.3,
> hugetlbfs_unlinked_fd.3, and hugetlbfs_unlinked_fd_for_size.3. Add cross
> references to the new man pages.
Hi all,
Libhugetlbfs-2.13 has been released and is now available from sourceforge.net
https://sourceforge.net/projects/libhugetlbfs/files/libhugetlbfs/2.13/
Please see the NEWS file for details on the changes from 2.12
Eric
signature.asc
Description: Digital signature
-
On Fri, 2 Mar 2012 13:46:30 -0800 (PST), Dan Scales wrote:
> Hi Eric,
>
> Sorry to contact you directly, but the libhugetlbfs mailing list on
> sourceforge seems to be inactive and overrun with spam.
>
> The libhugetlbfs code is great, and we have been able to use it to
> map text of executables in
On Fri, 23 Dec 2011 20:29:10 +0800, Easley wrote:
> I changed #define ELEMENTS 1024*1024*64 to #define ELEMENTS
> 1024*1024*20,
>
> and it takes effect. Thank you.
> ---
> And there is another question. It's on another machine
> I take the following steps:
> --
> [root@ ~]# echo 10
On Sun, 11 Dec 2011, Eric B Munson wrote:
> Currently the library will select the first hugetlbfs mount it finds for a
> given page size for usage. This can lead to frustrating failures when root
> owns the first mount and a user has a later mount in the list. This patch
> makes
On Wed, 07 Dec 2011, Eric B Munson wrote:
> Changes from V1:
> * Add note to HOWTO that the HEAPBASE option is not used with THP morecore
> * Change warning in thp_morecore to more accurately reflect the problem
> discovered
>
> Transparent huge pages (THP) make using hug
. Mounts
that are not usable will be skipped.
Signed-off-by: Eric B Munson
Cc: Andrew Lutomirski
Cc: David Gibson
---
Changes from V1:
Use access(2) instead of stat+permission checking
hugeutils.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/hugeutils.c b
On Sat, 10 Dec 2011, David Gibson wrote:
> On Fri, Dec 09, 2011 at 02:20:17PM -0500, Eric B Munson wrote:
> > Currently the library will select the first hugetlbfs mount it finds for a
> > given page size for usage. This can lead to frustrating failures when root
> > owns
. Mounts
that are not usable will be skipped.
Signed-off-by: Eric B Munson
Cc: Andrew Lutomirski
---
hugeutils.c | 26 +-
1 files changed, 25 insertions(+), 1 deletions(-)
diff --git a/hugeutils.c b/hugeutils.c
index b28bd0a..5b16a31 100644
--- a/hugeutils.c
+++ b
On Tue, 08 Nov 2011, Bill Carson wrote:
> Original fixed address at 0x8000 will fail when calling shmat on ARM.
> Use the address system provide for us will prevent this failure.
> Moreover, add cleanup when shmat failed to release reserved huge pages
> otherwise paranoid pool check will fail
Transparent huge pages (THP) have two tunables in sysfs used to control the
behavior of khugepaged. This patch adds the support for controlling these
tunables via hugeadm options.
Signed-off-by: Eric B Munson
Acked-by: Andrew Hastings
Cc: a...@cray.com
Cc: da...@gibson.dropbear.id.au
This patch adds controls for the new THP related functionality in the library.
Signed-off-by: Eric B Munson
Acked-by: Andrew Hastings
Cc: a...@cray.com
Cc: da...@gibson.dropbear.id.au
---
hugectl.c | 17 -
man/hugectl.8 |5 +
2 files changed, 21 insertions(+), 1
. This allows the kernel allocator to start the new
area as a huge page rather than wait for promotion.
Cc: a...@cray.com
Cc: da...@gibson.dropbear.id.au
Eric B Munson (3):
Add support for THP in morecore
Add support to hugeadm for configuring transparent huge pages
Add controls to hugectl for
/Documentation/transhuge.txt.
Signed-off-by: Eric B Munson
Cc: a...@cray.com
Cc: da...@gibson.dropbear.id.au
---
Changes from V1:
Note that MORECORE_HEAPBASE has no effect when using thp_morecore
Change warning when heapbase has changed outside of libhuge in thp_morecore
HOWTO
On Mon, 14 Nov 2011, bill4carson wrote:
>
>
> On 2011年11月14日 16:00, Easley wrote:
> >
> > I add "hugepagesz=1M hugepages=32" to grub.conf.
> > The "hugepagesz=16M" doesn't take effect.
> > The "hugepages=32" take effect.
> > What type is hardware support that?
> >
> > Any help will be appreciat
On Thu, 13 Oct 2011, Andrew Lutomirski wrote:
> I'm using Fedora 15 and I ran hugeadm --create-user-mounts=luto (and
> created a pool). The result is:
>
> $ cat /proc/self/mounts |grep huge
> systemd-1 /dev/hugepages autofs
> rw,relatime,fd=35,pgrp=1,timeout=300,minproto=5,maxproto=5,direct 0 0
On Thu, 10 Nov 2011, Luigi Semenzato wrote:
> Greetings,
>
> I am considering the use of libhugetlbfs for some application. We
> have a large binary that also needs to start quickly. I could not
> find a high-level overview of the strategy employed for placing the
> text segment in huge pages a
On Thu, 10 Nov 2011 17:27:40 +0800, Easley wrote:
> Thank you very much.
> I still have a question. How can I know that have been take effect?
> I take steps as follow:
> --
> mkdir /hugetlb
> mount -t hugetlbfs hugetlbfs /hugetlb
> HUGETLB_MORECORE=yes LD_PRELOAD=libhugetlbfs.so hu
On Thu, 10 Nov 2011 10:41:21 +0800, Easley wrote:
> Hi,
>
> I run "grep Huge /proc/meminfo" in rhel5.4 .
> it shows:
> HugePages_Total: 0
> HugePages_Free: 0
> HugePages_Rsvd: 0
> Hugepagesize: 2048 kB
>
> It says hugepagesize is 2M, how can I change the size to 16M?
> Any help will be appreciated.
On Thu, 22 Sep 2011 17:33:16 +0100, Mel Gorman wrote:
> On Thu, Sep 22, 2011 at 11:43:23AM +0100, Mel Gorman wrote:
>> Distributions typically install vmlinux gzipped but oprofile_start
>> assumes it is not gzipped because that is what would occur for a
>> manual kernel install. oprofile can work
On Wed, 20 Jul 2011, Andrew Hastings wrote:
> Eric,
>
> On 07/11/11 10:51, Eric B Munson wrote:
> > Transparent huge pages (THP) make using huge pages on x86(_64) incredibly
> > easy.
> > However, the standard glibc malloc is not setup to optimixe allocations to
Transparent huge pages (THP) have two tunables in sysfs used to control the
behavior of khugepaged. This patch adds the support for controlling these
tunables via hugeadm options.
Signed-off-by: Eric B Munson
---
hugeadm.c | 105
/Documentation/transhuge.txt.
Signed-off-by: Eric B Munson
---
HOWTO |7 +++-
hugeutils.c |9 +
libhugetlbfs_internal.h |1 +
morecore.c | 84 --
4 files changed, 96 insertions(+), 5 deletions
originally used to
back the heap with hugetlbfs huge pages to ensure that the heap is grown in
huge page sized increments. This allows the kernel allocator to start the new
area as a huge page rather than wait for promotion.
Eric B Munson (3):
Add support for THP in morecore
Add support to hugeadm
This patch adds controls for the new THP related functionality in the library.
Signed-off-by: Eric B Munson
---
hugectl.c | 17 -
man/hugectl.8 |5 +
2 files changed, 21 insertions(+), 1 deletions(-)
diff --git a/hugectl.c b/hugectl.c
index e2a9828..741247f 100644
On Mon, 25 Jul 2011, Matthew McClintock wrote:
> Signed-off-by: Matthew McClintock
> ---
> Not 100% sure about this fix, but it does remove a build failure when
> build with -j100
Applied, thanks.
Eric
>
> Makefile |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git
On Mon, 25 Jul 2011, McClintock Matthew-B29882 wrote:
> On Mon, Jul 25, 2011 at 3:11 PM, Eric B Munson wrote:
> > Unfortunately this doesn't work for me for a -j values. There is a subtle
> > timing issue with the linker script and the old linker tests. When I build
> &g
On Mon, 25 Jul 2011, Matthew McClintock wrote:
> Signed-off-by: Matthew McClintock
> ---
> Not 100% sure about this fix, but it does remove a build failure when
> build with -j100
>
> Makefile |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/Makefile b/Makefile
>
/Documentation/transhuge.txt.
Signed-off-by: Eric B Munson
---
HOWTO |7 +++-
hugeutils.c |4 ++
libhugetlbfs_internal.h |1 +
morecore.c | 84 --
4 files changed, 91 insertions(+), 5 deletions(-)
diff
originally used to
back the heap with hugetlbfs huge pages to ensure that the heap is grown in
huge page sized increments. This allows the kernel allocator to start the new
area as a huge page rather than wait for promotion.
Eric B Munson (3):
Add support for THP in morecore
Add support to hugeadm
This patch adds controls for the new THP related functionality in the library.
Signed-off-by: Eric B Munson
---
hugectl.c | 17 -
man/hugectl.8 | 11 +++
2 files changed, 27 insertions(+), 1 deletions(-)
diff --git a/hugectl.c b/hugectl.c
index e2a9828..741247f
Transparent huge pages (THP) have two tunables in sysfs used to control the
behavior of khugepaged. This patch adds the support for controlling these
tunables via hugeadm options.
Signed-off-by: Eric B Munson
---
hugeadm.c | 105
On Wed, 15 Jun 2011, Andrew Hastings wrote:
> Eric B Munson wrote:
> >It is true that you will end up with huge pages for the heap eventually
> >without
> >changing anything. khugepaged will promote areas it finds that are 2MB
> >aligned, etc. However, it is possible
On Tue, 14 Jun 2011, David Gibson wrote:
> On Mon, Jun 13, 2011 at 09:06:16AM -0500, Andrew Hastings wrote:
> > On 06/08/2011 09:40 PM, Eric B Munson wrote:
> > > On Wed, Jun 8, 2011 at 8:50 PM, David Gibson
> > > wrote:
> [snip]
> > Actually, in thinking
On Wed, Jun 8, 2011 at 8:50 PM, David Gibson
wrote:
> On Wed, Jun 08, 2011 at 11:16:42AM -0500, Andrew Hastings wrote:
>> On 06/07/11 11:08, Eric B Munson wrote:
>> > On Fri, 03 Jun 2011, Andrew Hastings wrote:
> [snip]
>> >> Might it make sense for HUGETLB_M
1 - 100 of 691 matches
Mail list logo
