From: "Yinghai Lu" <yhlu.kernel@gmail.com>
To: "Gabriel C" <nix.or.die@googlemail.com>
Cc: "Andrew Morton" <akpm@linux-foundation.org>,
"Ingo Molnar" <mingo@elte.hu>, "H. Peter Anvin" <hpa@zytor.com>,
"Thomas Gleixner" <tglx@linutronix.de>,
"Mika Fischer" <mika.fischer@zoopnet.de>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 1/2] x86: mtrr cleanup for converting continuous to discrete layout v7
Date: Wed, 30 Apr 2008 09:26:19 -0700 [thread overview]
Message-ID: <86802c440804300926s42d0da4fs34a9a95d6d466cbe@mail.gmail.com> (raw)
In-Reply-To: <48186040.3050800@googlemail.com>
On Wed, Apr 30, 2008 at 5:04 AM, Gabriel C <nix.or.die@googlemail.com> wrote:
>
> Yinghai Lu wrote:
> > On Tue, Apr 29, 2008 at 9:12 PM, Gabriel C <nix.or.die@googlemail.com> wrote:
> >> Yinghai Lu wrote:
> >> > please try the patches
> >> >
> >> > from
> >> > http://lkml.org/lkml/2008/4/29/754
> >> > http://lkml.org/lkml/2008/4/29/753
> >> >
> >> > in addtion to
> >> >
> >> > http://people.redhat.com/mingo/x86.git/README
> >> > ( it has v8 already)
> >> >
> >> > and try with mtrr_gran_size=128m
> >>
> >> Without any value I get :
> >>
> >> ...
> >>
> >> [ 0.000000] Linux version 2.6.25-x86-latest.git-06598-g6a2c2ff-dirty (crazy@thor) (gcc version 4.3.0 (Frugalware Linux) ) #1 SMP PREEMPT Wed Apr 30 05:51:39 CEST 2008
> >>
> >>
> >> [ 0.000000] Command line: root=/dev/sdb1 ro debug vga=0x317
> >> [ 0.000000] BIOS-provided physical RAM map:
> >> [ 0.000000] BIOS-e820: 0000000000000000 - 000000000009cc00 (usable)
> >> [ 0.000000] BIOS-e820: 000000000009cc00 - 00000000000a0000 (reserved)
> >> [ 0.000000] BIOS-e820: 00000000000e4000 - 0000000000100000 (reserved)
> >> [ 0.000000] BIOS-e820: 0000000000100000 - 00000000cf550000 (usable)
> >> [ 0.000000] BIOS-e820: 00000000cf550000 - 00000000cf55e000 (ACPI data)
> >> [ 0.000000] BIOS-e820: 00000000cf55e000 - 00000000cf5e0000 (ACPI NVS)
> >> [ 0.000000] BIOS-e820: 00000000cf5e0000 - 00000000cf600000 (reserved)
> >> [ 0.000000] BIOS-e820: 00000000fee00000 - 00000000fee01000 (reserved)
> >> [ 0.000000] BIOS-e820: 00000000ffc00000 - 0000000100000000 (reserved)
> >> [ 0.000000] BIOS-e820: 0000000100000000 - 000000012c000000 (usable)
> >> [ 0.000000] Entering add_active_range(0, 0, 156) 0 entries of 256 used
> >> [ 0.000000] Entering add_active_range(0, 256, 849232) 1 entries of 256 used
> >> [ 0.000000] Entering add_active_range(0, 1048576, 1228800) 2 entries of 256 used
> >> [ 0.000000] max_pfn_mapped = 1228800
> >> [ 0.000000] x86 PAT enabled: cpu 0, old 0x7040600070406, new 0x7010600070106
> >> [ 0.000000] After WB checking
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 000000000012c000
> >> [ 0.000000] After UC checking
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 00000000000cf600
> >> [ 0.000000] MTRR MAP PFN: 0000000000100000 - 000000000012c000
> >> [ 0.000000] MTRR MAP PFN: 00000000000cf800 - 00000000000d0000
> >> [ 0.000000] After sorting
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 00000000000cf600
> >> [ 0.000000] MTRR MAP PFN: 00000000000cf800 - 00000000000d0000
> >> [ 0.000000] MTRR MAP PFN: 0000000000100000 - 000000000012c000
> >> [ 0.000000] range0: 0000000000000000 - 00000000c0000000
> >> [ 0.000000] Setting variable MTRR 0, base: 0MB, range: 2048MB, type WB
> >> [ 0.000000] Setting variable MTRR 1, base: 2048MB, range: 1024MB, type WB
> >> [ 0.000000] range: 00000000c0000000 - 00000000cf600000
> >>
> >> [ 0.000000] Setting variable MTRR 2, base: 3072MB, range: 128MB, type WB
> >> [ 0.000000] Setting variable MTRR 3, base: 3200MB, range: 64MB, type WB
> >> [ 0.000000] Setting variable MTRR 4, base: 3264MB, range: 32MB, type WB
> >> [ 0.000000] Setting variable MTRR 5, base: 3296MB, range: 16MB, type WB
> >> [ 0.000000] Setting variable MTRR 6, base: 3312MB, range: 4MB, type WB
> >> [ 0.000000] Setting variable MTRR 7, base: 3316MB, range: 2MB, type WB
> >> [ 0.000000] range0: 00000000cf800000 - 00000000df800000
> >> [ 0.000000] range: 00000000df800000 - 00000000d0000000
> >>
> >> [ 0.000000] DONE variable MTRRs
> >> [ 0.000000] x86 PAT enabled: cpu 0, old 0x7010600070106, new 0x7010600070106
> >> [ 0.000000] After WB checking
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 00000000000cf600
> >>
> >> [ 0.000000] After UC checking
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 00000000000cf600
> >>
> >> [ 0.000000] After sorting
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 00000000000cf600
> >> [ 0.000000] WARNING: BIOS bug: CPU MTRRs don't cover all of memory, losing 704MB of RAM.
> >>
> >> [ 0.000000] update e820 for mtrr
> >>
> >> [ 0.000000] modified physical RAM map:
> >> [ 0.000000] modified: 0000000000000000 - 000000000009cc00 (usable)
> >> [ 0.000000] modified: 000000000009cc00 - 00000000000a0000 (reserved)
> >> [ 0.000000] modified: 00000000000e4000 - 0000000000100000 (reserved)
> >> [ 0.000000] modified: 0000000000100000 - 00000000cf550000 (usable)
> >> [ 0.000000] modified: 00000000cf550000 - 00000000cf55e000 (ACPI data)
> >> [ 0.000000] modified: 00000000cf55e000 - 00000000cf5e0000 (ACPI NVS)
> >> [ 0.000000] modified: 00000000cf5e0000 - 00000000cf600000 (reserved)
> >> [ 0.000000] modified: 00000000fee00000 - 00000000fee01000 (reserved)
> >>
> >> [ 0.000000] modified: 00000000ffc00000 - 000000012c000000 (reserved)
> >> [ 0.000000] Entering add_active_range(0, 0, 156) 3 entries of 256 used
> >> [ 0.000000] Entering add_active_range(0, 256, 849232) 3 entries of 256 used
> >>
> >> [ 0.000000] max_pfn_mapped = 1228800
> >> [ 0.000000] init_memory_mapping
> >>
> >> ...
> >>
> >> [ 20.984343] [drm] Initialized i915 1.6.0 20060119 on minor 0
> >> [ 21.450368] mtrr: no more MTRRs available
> >>
> >> ...
> >>
> >> with mtrr_gran_size=128m I get :
> >>
> >> ...
> >>
> >> [ 0.000000] Linux version 2.6.25-x86-latest.git-06598-g6a2c2ff-dirty (crazy@thor) (gcc version 4.3.0 (Frugalware Linux) ) #1 SMP PREEMPT Wed Apr 30 05:51:39 CEST 2008
> >> [ 0.000000] Command line: root=/dev/sdb1 ro debug vga=0x317 mtrr_gran_size=128m
> >>
> >>
> >> [ 0.000000] BIOS-provided physical RAM map:
> >> [ 0.000000] BIOS-e820: 0000000000000000 - 000000000009cc00 (usable)
> >> [ 0.000000] BIOS-e820: 000000000009cc00 - 00000000000a0000 (reserved)
> >> [ 0.000000] BIOS-e820: 00000000000e4000 - 0000000000100000 (reserved)
> >> [ 0.000000] BIOS-e820: 0000000000100000 - 00000000cf550000 (usable)
> >> [ 0.000000] BIOS-e820: 00000000cf550000 - 00000000cf55e000 (ACPI data)
> >> [ 0.000000] BIOS-e820: 00000000cf55e000 - 00000000cf5e0000 (ACPI NVS)
> >> [ 0.000000] BIOS-e820: 00000000cf5e0000 - 00000000cf600000 (reserved)
> >> [ 0.000000] BIOS-e820: 00000000fee00000 - 00000000fee01000 (reserved)
> >> [ 0.000000] BIOS-e820: 00000000ffc00000 - 0000000100000000 (reserved)
> >> [ 0.000000] BIOS-e820: 0000000100000000 - 000000012c000000 (usable)
> >> [ 0.000000] Entering add_active_range(0, 0, 156) 0 entries of 256 used
> >> [ 0.000000] Entering add_active_range(0, 256, 849232) 1 entries of 256 used
> >> [ 0.000000] Entering add_active_range(0, 1048576, 1228800) 2 entries of 256 used
> >> [ 0.000000] max_pfn_mapped = 1228800
> >> [ 0.000000] x86 PAT enabled: cpu 0, old 0x7040600070406, new 0x7010600070106
> >> [ 0.000000] After WB checking
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 000000000012c000
> >> [ 0.000000] After UC checking
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 00000000000cf600
> >> [ 0.000000] MTRR MAP PFN: 0000000000100000 - 000000000012c000
> >> [ 0.000000] MTRR MAP PFN: 00000000000cf800 - 00000000000d0000
> >> [ 0.000000] After sorting
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 00000000000cf600
> >> [ 0.000000] MTRR MAP PFN: 00000000000cf800 - 00000000000d0000
> >> [ 0.000000] MTRR MAP PFN: 0000000000100000 - 000000000012c000
> >> [ 0.000000] range0: 0000000000000000 - 00000000c0000000
> >> [ 0.000000] Setting variable MTRR 0, base: 0MB, range: 2048MB, type WB
> >> [ 0.000000] Setting variable MTRR 1, base: 2048MB, range: 1024MB, type WB
> >> [ 0.000000] range: 00000000c0000000 - 00000000c8000000
> >> [ 0.000000] Setting variable MTRR 2, base: 3072MB, range: 128MB, type WB
> >> [ 0.000000] rangeX: 00000000d0000000 - 00000000d0000000
> >> [ 0.000000] range0: 0000000100000000 - 0000000120000000
> >> [ 0.000000] Setting variable MTRR 3, base: 4096MB, range: 512MB, type WB
> >> [ 0.000000] range: 0000000120000000 - 0000000128000000
> >> [ 0.000000] Setting variable MTRR 4, base: 4608MB, range: 128MB, type WB
> >> [ 0.000000] DONE variable MTRRs
> >> [ 0.000000] x86 PAT enabled: cpu 0, old 0x7010600070106, new 0x7010600070106
> >> [ 0.000000] After WB checking
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 00000000000c8000
> >> [ 0.000000] MTRR MAP PFN: 0000000000100000 - 0000000000128000
> >> [ 0.000000] After UC checking
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 00000000000c8000
> >> [ 0.000000] MTRR MAP PFN: 0000000000100000 - 0000000000128000
> >> [ 0.000000] After sorting
> >> [ 0.000000] MTRR MAP PFN: 0000000000000000 - 00000000000c8000
> >> [ 0.000000] MTRR MAP PFN: 0000000000100000 - 0000000000128000
> >> [ 0.000000] WARNING: BIOS bug: CPU MTRRs don't cover all of memory, losing 181MB of RAM.
> >>
> >> [ 0.000000] update e820 for mtrr
> >>
> >> [ 0.000000] modified physical RAM map:
> >> [ 0.000000] modified: 0000000000000000 - 000000000009cc00 (usable)
> >> [ 0.000000] modified: 000000000009cc00 - 00000000000a0000 (reserved)
> >> [ 0.000000] modified: 00000000000e4000 - 0000000000100000 (reserved)
> >> [ 0.000000] modified: 0000000000100000 - 00000000c8000000 (usable)
> >> [ 0.000000] modified: 00000000c8000000 - 00000000cf550000 (reserved)
> >> [ 0.000000] modified: 00000000cf550000 - 00000000cf55e000 (ACPI data)
> >> [ 0.000000] modified: 00000000cf55e000 - 00000000cf5e0000 (ACPI NVS)
> >> [ 0.000000] modified: 00000000cf5e0000 - 00000000cf600000 (reserved)
> >> [ 0.000000] modified: 00000000fee00000 - 00000000fee01000 (reserved)
> >>
> >> [ 0.000000] modified: 00000000ffc00000 - 0000000100000000 (reserved)
> >> [ 0.000000] modified: 0000000100000000 - 0000000128000000 (usable)
> >> [ 0.000000] modified: 0000000128000000 - 000000012c000000 (reserved)
> >>
> >> [ 0.000000] Entering add_active_range(0, 0, 156) 3 entries of 256 used
> >> [ 0.000000] Entering add_active_range(0, 256, 819200) 3 entries of 256 used
> >> [ 0.000000] Entering add_active_range(0, 1048576, 1212416) 3 entries of 256 used
> >>
> >> [ 0.000000] max_pfn_mapped = 1228800
> >> [ 0.000000] init_memory_mapping
> >>
> >> ...
> >>
> >> I will try tomorrow some more boot options but now I need some sleep ;)
> >
> > thanks. let's try different mtrr_chunk_size/mtrr_gran_size to get
> > back more ram.
> > under mtrr_gran_size=128m, does the your X server work well..., fast or slow?
>
> Yes X is fine and fast , it is even fine ( slower from my felling ) when I lose the 704MB.
> In general with x86-latest.git tree things seems faster on that box , maybe there are some other bug fixes too.
>
> I've tested some mtrr_chunk_size/mtrr_gran_size combos, dmesg's are uploaded there :
>
> http://frugalware.org/~crazy/dmesg/mtrr/
>
> Also setting lower values on mtrr_gran_size seems to give more RAM back ,
> mtrr_chunk_size 256/512 eats 704 MB and 128 doesn't seems to do something ?
>
> Other things I noticed ( probably you could add a note about in kernel-parameter.txt or some doc file ):
>
> Setting mtrr_gran_size to high , on my box >=512m hangs the box on boot ,
> setting it to low , on my box <=8m , will cause X to die with such a message :
>
> xf86MapVidMem: Could not mmap framebuffer (0xd0000000,0x10000000) (Invalid argument)
>
> If you want I can test such values for mtrr_chunk_size too , just let me know.
>
> To be honest I'm even fine when losing 700 - 800 MB as long X and everything else does work.
> The other alternative for me for that problem without your patches will be to buy new ram ( 2 x 1G )
> and then I lose near 2,3G compared to now or live with broken X until xorg-server will support and
> *work fine* with PAT ( most probably not that soon ).
great. I am still working on
1. auto detect optimal chunk_size, gran_size
2. get back all RAM, at least only lose 1 or 2 M
YH
next prev parent reply other threads:[~2008-04-30 16:26 UTC|newest]
Thread overview: 89+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-04-28 6:37 [PATCH] x86: mtrr cleanup for converting continuous to discrete layout Yinghai Lu
2008-04-28 9:06 ` [PATCH] x86: mtrr cleanup for converting continuous to discrete layout v2 Yinghai Lu
2008-04-28 13:08 ` Ingo Molnar
2008-04-28 13:49 ` Arjan van de Ven
2008-04-28 15:28 ` Mika Fischer
2008-04-28 5:50 ` Arjan van de Ven
2008-04-28 16:01 ` Gabriel C
2008-04-28 16:28 ` Mika Fischer
2008-04-28 19:44 ` [PATCH] x86: mtrr cleanup for converting continuous to discrete layout v3 Yinghai Lu
2008-04-28 20:15 ` Ingo Molnar
2008-04-28 20:18 ` Yinghai Lu
2008-04-28 20:29 ` Ingo Molnar
2008-04-28 20:16 ` [PATCH] x86: mtrr cleanup for converting continuous to discrete layout v4 Yinghai Lu
2008-04-28 22:05 ` [PATCH] x86: mtrr cleanup for converting continuous to discrete layout v5 Yinghai Lu
2008-04-28 22:36 ` Randy Dunlap
2008-04-28 22:47 ` Yinghai Lu
2008-04-29 2:42 ` Andrew Morton
2008-04-29 3:01 ` Yinghai Lu
[not found] ` <200804290157.30651.yhlu.kernel@gmail.com>
2008-04-29 8:59 ` [PATCH 2/2] x86: fix trimming e820 with MTRR holes Yinghai Lu
2008-04-29 11:35 ` Ingo Molnar
2008-04-29 17:18 ` Yinghai Lu
2008-04-29 17:20 ` Yinghai Lu
2008-04-30 3:25 ` [PATCH] x86: fix trimming e820 with MTRR holes. - fix Yinghai Lu
2008-04-30 12:09 ` Ingo Molnar
2008-04-29 9:00 ` [PATCH 1/2] x86: mtrr cleanup for converting continuous to discrete layout v7 Yinghai Lu
2008-04-29 9:47 ` Gabriel C
2008-04-29 10:30 ` Yinghai Lu
2008-04-29 10:56 ` Yinghai Lu
2008-04-29 11:26 ` Ingo Molnar
2008-04-29 11:51 ` Gabriel C
2008-04-29 17:11 ` Yinghai Lu
2008-04-29 20:25 ` Gabriel C
2008-04-29 21:49 ` Yinghai Lu
2008-04-29 23:56 ` Gabriel C
2008-04-30 0:06 ` Gabriel C
2008-04-30 0:38 ` Yinghai Lu
2008-04-30 1:02 ` Gabriel C
2008-04-30 3:00 ` Yinghai Lu
2008-04-30 3:29 ` Yinghai Lu
2008-04-30 4:12 ` Gabriel C
2008-04-30 4:25 ` Yinghai Lu
2008-04-30 12:04 ` Gabriel C
2008-04-30 16:26 ` Yinghai Lu [this message]
2008-04-30 0:13 ` Yinghai Lu
2008-04-29 10:52 ` [PATCH 1/2] x86: mtrr cleanup for converting continuous to discrete layout v8 Yinghai Lu
2008-04-29 13:07 ` Ingo Molnar
2008-04-29 17:25 ` Yinghai Lu
2008-04-29 20:46 ` Randy Dunlap
2008-04-29 21:54 ` Yinghai Lu
2008-04-30 3:25 ` [PATCH] x86: mtrr cleanup for converting continuous to discrete layout v8 - fix Yinghai Lu
2008-04-30 12:09 ` Ingo Molnar
2008-05-01 8:00 ` [PATCH] x86: mtrr cleanup for converting continuous to discrete - auto detect Yinghai Lu
2008-05-01 11:45 ` Gabriel C
2008-05-02 0:06 ` Yinghai Lu
2008-05-02 0:29 ` Gabriel C
2008-05-02 0:35 ` Yinghai Lu
2008-05-02 1:18 ` Gabriel C
2008-05-02 1:55 ` Yinghai Lu
2008-05-01 12:09 ` Mika Fischer
2008-05-01 16:35 ` Yinghai Lu
2008-05-01 16:59 ` Mika Fischer
2008-05-01 17:40 ` Yinghai Lu
2008-05-01 15:09 ` Randy Dunlap
2008-05-01 16:38 ` Yinghai Lu
2008-05-01 18:57 ` [PATCH] x86: mtrr cleanup for converting continuous to discrete - auto detect v2 Yinghai Lu
2008-05-01 19:42 ` H. Peter Anvin
2008-05-01 21:02 ` Yinghai Lu
2008-05-01 21:10 ` H. Peter Anvin
2008-05-01 21:20 ` Yinghai Lu
2008-05-01 21:26 ` H. Peter Anvin
2008-05-01 21:31 ` Yinghai Lu
2008-05-01 21:33 ` H. Peter Anvin
2008-05-01 21:44 ` Yinghai Lu
2008-05-01 21:49 ` H. Peter Anvin
2008-05-01 22:52 ` Yinghai Lu
2008-05-01 22:57 ` H. Peter Anvin
2008-05-01 23:10 ` Yinghai Lu
2008-05-02 0:52 ` [PATCH] x86: mtrr cleanup for converting continuous to discrete - auto detect v3 Yinghai Lu
2008-05-02 9:40 ` [PATCH] x86: mtrr cleanup for converting continuous to discrete - auto detect v4 Yinghai Lu
2008-04-29 19:00 ` [PATCH] x86: mtrr cleanup for converting continuous to discrete layout v5 Eric W. Biederman
2008-04-29 20:04 ` Yinghai Lu
2008-04-29 20:29 ` Eric W. Biederman
2008-04-29 21:57 ` Yinghai Lu
2008-04-29 22:09 ` Ingo Molnar
2008-04-29 22:18 ` Yinghai Lu
2008-04-29 22:14 ` Eric W. Biederman
2008-04-29 22:54 ` Thomas Gleixner
2008-04-30 1:16 ` Eric W. Biederman
2008-04-30 9:57 ` Alan Cox
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=86802c440804300926s42d0da4fs34a9a95d6d466cbe@mail.gmail.com \
--to=yhlu.kernel@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=hpa@zytor.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mika.fischer@zoopnet.de \
--cc=mingo@elte.hu \
--cc=nix.or.die@googlemail.com \
--cc=tglx@linutronix.de \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®