From: Nishanth Aravamudan <nacc@us.ibm.com>
To: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <gregkh@suse.de>, Nick Piggin <npiggin@suse.de>,
wli@holomorphy.com, agl@us.ibm.com, luick@cray.com,
Lee.Schermerhorn@hp.com, linux-mm@kvack.org
Subject: Re: [RFC][PATCH] hugetlb: add information and interface in sysfs [Was Re: [RFC][PATCH 4/5] Documentation: add node files to sysfs ABI]
Date: Mon, 28 Apr 2008 13:52:00 -0700 [thread overview]
Message-ID: <20080428205200.GA4386@us.ibm.com> (raw)
In-Reply-To: <Pine.LNX.4.64.0804281328300.31163@schroedinger.engr.sgi.com>
On 28.04.2008 [13:31:00 -0700], Christoph Lameter wrote:
> On Sat, 26 Apr 2008, Greg KH wrote:
>
> > Also, why use a "units" here, just always use the lowest unit, and
> > userspace can convert from kB to GB if needed.
>
> Additional complications will come about because IA64 supports
> varying hugetlb sizes from 4kb to 1GB.
What "complications" do you mean? It's a small function indeed to
convert from the directory name to the corresponding "human-named" size,
e.g. hugepages-1048576 to "1 GB". And such a function will probably
exist in libhugetlbfs at some point, for applications to use, if they
like.
A potential problem I do see is for a 32-bit binary running on a 64-bit
kernel and is one we've run against for 32-bit binaries with 16G pages
available. The 32-bit binary can't actually store the size of the
hugepage in an unsigned long, so we have to remember how big of a value
we can represent (i.e., max_hugepage_size_in_kb) and check what's
obtained from /proc/meminfo against that. Not ideal, for sure.
> Also we would at some point like to add support for 1TB hugepages
> (that may depend on the presence of a special device that handles
> these).
I also don't see a limitation here? For 32-bit programs, we'll see
1073741824 and know we can't convert that into a valid value in bytes.
More importnatly, I think the fact that IA64 supports multiple hugepage
sizes is a reason *for* moving to sysfs for this information? However, I
think we may need to massage the IA64-specific bits of the kernel to
actually support multiple hugepage size pools being available at
run-time? That is, with the current kernel, we can only support one
hugepagesize at run-time, due to VHPT restrictions?
Thanks,
Nish
--
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>
next prev parent reply other threads:[~2008-04-28 20:52 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-04-11 23:44 [PATCH 1/5] hugetlb: numafy several functions Nishanth Aravamudan
2008-04-11 23:47 ` [RFC][PATCH 2/5] " Nishanth Aravamudan
2008-04-11 23:47 ` [PATCH 3/5] hugetlb: interleave dequeueing of huge pages Nishanth Aravamudan
2008-04-11 23:49 ` [RFC][PATCH 4/5] Documentation: add node files to sysfs ABI Nishanth Aravamudan
2008-04-11 23:50 ` [RFC][PATCH 5/5] Documentation: update ABI and hugetlbpage.txt for per-node files Nishanth Aravamudan
2008-04-11 23:56 ` [RFC][PATCH 4/5] Documentation: add node files to sysfs ABI Greg KH
2008-04-12 0:27 ` Nishanth Aravamudan
2008-04-12 9:41 ` Nick Piggin
2008-04-12 10:26 ` Christoph Lameter
2008-04-14 21:09 ` Nishanth Aravamudan
2008-04-13 3:41 ` Greg KH
2008-04-14 21:05 ` Nishanth Aravamudan
2008-04-17 23:16 ` Nishanth Aravamudan
2008-04-17 23:22 ` Christoph Lameter
2008-04-17 23:36 ` Nishanth Aravamudan
2008-04-17 23:39 ` Christoph Lameter
2008-04-18 6:04 ` Nishanth Aravamudan
2008-04-18 17:27 ` Nishanth Aravamudan
2008-04-20 2:24 ` Greg KH
2008-04-21 16:43 ` Nishanth Aravamudan
2008-04-20 2:21 ` Greg KH
2008-04-21 6:06 ` Christoph Lameter
2008-04-21 16:41 ` Nishanth Aravamudan
2008-04-22 5:14 ` Nick Piggin
2008-04-22 16:56 ` Nishanth Aravamudan
2008-04-23 1:03 ` Nick Piggin
2008-04-23 18:32 ` Nishanth Aravamudan
2008-04-23 19:07 ` Adam Litke
2008-04-24 7:13 ` Nick Piggin
2008-04-24 15:54 ` Nishanth Aravamudan
2008-04-27 3:49 ` [RFC][PATCH] hugetlb: add information and interface in sysfs [Was Re: [RFC][PATCH 4/5] Documentation: add node files to sysfs ABI] Nishanth Aravamudan
2008-04-27 5:10 ` Greg KH
2008-04-28 17:22 ` Nishanth Aravamudan
2008-04-28 17:29 ` Greg KH
2008-04-29 17:11 ` Nishanth Aravamudan
2008-04-29 17:22 ` Greg KH
2008-04-29 18:14 ` Nishanth Aravamudan
2008-04-29 18:26 ` Greg KH
2008-04-29 23:48 ` Nishanth Aravamudan
2008-05-01 3:07 ` Greg KH
2008-05-01 18:25 ` Nishanth Aravamudan
2008-04-30 19:19 ` Nishanth Aravamudan
2008-05-01 3:08 ` Greg KH
2008-05-02 17:58 ` Nishanth Aravamudan
2008-04-28 20:31 ` Christoph Lameter
2008-04-28 20:52 ` Nishanth Aravamudan [this message]
2008-04-28 21:29 ` Christoph Lameter
2008-04-29 16:43 ` Nishanth Aravamudan
2008-04-29 17:01 ` Christoph Lameter
2008-04-14 14:52 ` [RFC][PATCH 2/5] hugetlb: numafy several functions Adam Litke
2008-04-14 21:10 ` Nishanth Aravamudan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20080428205200.GA4386@us.ibm.com \
--to=nacc@us.ibm.com \
--cc=Lee.Schermerhorn@hp.com \
--cc=agl@us.ibm.com \
--cc=clameter@sgi.com \
--cc=gregkh@suse.de \
--cc=linux-mm@kvack.org \
--cc=luick@cray.com \
--cc=npiggin@suse.de \
--cc=wli@holomorphy.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.