headers
NetBSD source update | 1 Sep 2010 05:13
Picon

daily CVS update output


Updating src tree:
P src/bin/pax/ar_io.c
P src/bin/pax/options.c
U src/doc/CHANGES
P src/include/rmt.h
P src/lib/librmt/rmtlib.c
U src/sys/arch/arm/omap/omapfb.c
U src/sys/arch/arm/omap/omapfbreg.h
P src/sys/arch/arm/pic/pic.c
P src/sys/arch/dreamcast/dev/gdrom.c
P src/sys/arch/evbarm/conf/BEAGLEBOARD
P src/sys/arch/sparc/dev/cgfourteen.c
P src/sys/arch/sparc/dev/cgfourteenvar.h
P src/sys/arch/sparc64/dev/gfb.c
P src/sys/ddb/db_input.c
P src/tests/h_macros.h
P src/usr.bin/audiocfg/Makefile
U src/usr.bin/audiocfg/audiocfg.1

Updating xsrc tree:

Killing core files:

Running the SUP scanner:
SUP Scan for current starting at Wed Sep  1 03:02:02 2010
SUP Scan for current completed at Wed Sep  1 03:02:28 2010
SUP Scan for mirror starting at Wed Sep  1 03:02:28 2010
SUP Scan for mirror completed at Wed Sep  1 03:05:08 2010
(Continue reading)

Permalink | Reply | 0 comments |
headers
Thomas Klausner | 1 Sep 2010 15:17
Picon

Re: firefox slowness -- X server problem?

On Mon, Aug 23, 2010 at 07:30:41PM +0300, Jukka Ruohonen wrote:
> > Going back to a kernel from 20100702, firefox was fast again.
> > I looked at the dmesg differences and found that I had enabled acpicpu
> > in the meantime. I rebuilt a kernel using sources from yesterday, but
> > disabling acpicpu, and firefox is still fast; with a kernel from
> > yesterday and acpicpu enabled, it is slow.
> 
> This should be fixed now.

I think I already replied to this, but it's still in my inbox.

In case I hadn't: yes, it's fixed now, thank you!
 Thomas
Permalink | Reply | 0 comments |
headers
Aran Clauson | 1 Sep 2010 20:54
Picon

Re: Amd64 Build error

All,
	Sorry for the noise.  Bad /etc/mk.conf.

Aran

> All,
> 	I am getting the following error when building distribution from current (yesterday) on amd64:
> 
> /home/NetBSD/src/../obj/gnu/usr.bin/gdb6/libgdb/libgdb.a(tui-io.o): In function `tui_setup_io':
> tui-io.c:(.text+0x211): undefined reference to `readline_echoing_p'
> tui-io.c:(.text+0x249): undefined reference to `readline_echoing_p'
> tui-io.c:(.text+0x377): undefined reference to `readline_echoing_p'
> /home/NetBSD/src/../obj/gnu/usr.bin/gdb6/bfd/libbfd.a(cache.o): In function `bfd_open_file':
> cache.c:(.text+0x29e): undefined reference to `unlink_if_ordinary'
> collect2: ld returned 1 exit status
> 
> Any ideas?
> 
> Aran
>
Permalink | Reply | 0 comments |
headers
Aran Clauson | 1 Sep 2010 20:05
Picon

Amd64 Build error

All,
	I am getting the following error when building distribution from current (yesterday) on amd64:

/home/NetBSD/src/../obj/gnu/usr.bin/gdb6/libgdb/libgdb.a(tui-io.o): In function `tui_setup_io':
tui-io.c:(.text+0x211): undefined reference to `readline_echoing_p'
tui-io.c:(.text+0x249): undefined reference to `readline_echoing_p'
tui-io.c:(.text+0x377): undefined reference to `readline_echoing_p'
/home/NetBSD/src/../obj/gnu/usr.bin/gdb6/bfd/libbfd.a(cache.o): In function `bfd_open_file':
cache.c:(.text+0x29e): undefined reference to `unlink_if_ordinary'
collect2: ld returned 1 exit status

Any ideas?

Aran
Permalink | Reply | 1 comment |
headers
NetBSD source update | 2 Sep 2010 05:19
Picon

daily CVS update output


Updating src tree:
P src/crypto/external/bsd/netpgp/dist/tst
P src/crypto/external/bsd/netpgp/dist/src/hkpd/Makefile
P src/crypto/external/bsd/netpgp/dist/src/hkpd/hkpd.c
P src/crypto/external/bsd/netpgp/dist/src/hkpd/main.c
P src/crypto/external/bsd/netpgp/dist/src/lib/create.c
P src/crypto/external/bsd/netpgp/dist/src/lib/keyring.c
P src/crypto/external/bsd/netpgp/dist/src/lib/netpgp.c
P src/crypto/external/bsd/netpgp/dist/src/lib/version.h
P src/crypto/external/bsd/netpgp/dist/src/netpgp/netpgp.c
P src/crypto/external/bsd/netpgp/dist/src/netpgpkeys/netpgpkeys.c
P src/crypto/external/bsd/netpgp/dist/src/netpgpverify/verify.c
P src/crypto/external/bsd/netpgp/netpgp/Makefile
P src/distrib/sets/lists/man/mi
P src/distrib/sets/lists/tests/mi
P src/external/bsd/ntp/bin/ntpdc/Makefile
P src/external/bsd/ntp/bin/ntpq/Makefile
P src/external/bsd/ntp/include/config.h
P src/gnu/usr.bin/groff/tmac/mdoc.local
P src/lib/Makefile
P src/lib/libm/src/k_standard.c
P src/lib/libp2k/p2k.c
P src/lib/libperfuse/libperfuse.3
P src/lib/libperfuse/ops.c
P src/lib/libperfuse/perfuse_if.h
P src/lib/libperfuse/perfuse_priv.h
P src/lib/libperfuse/subr.c
P src/lib/libukfs/ukfs.c
P src/sys/arch/arm/omap/files.omap2
(Continue reading)

Permalink | Reply | 0 comments |
headers
NetBSD source update | 3 Sep 2010 05:07
Picon

daily CVS update output


Updating src tree:
P src/crypto/external/bsd/netpgp/Makefile
P src/crypto/external/bsd/netpgp/dist/configure
P src/crypto/external/bsd/netpgp/dist/configure.ac
P src/crypto/external/bsd/netpgp/dist/include/netpgp.h
U src/crypto/external/bsd/netpgp/dist/src/hkpclient/hkpc.1
U src/crypto/external/bsd/netpgp/dist/src/hkpclient/hkpc.c
U src/crypto/external/bsd/netpgp/dist/src/hkpclient/hkpc.h
U src/crypto/external/bsd/netpgp/dist/src/hkpclient/main.c
P src/crypto/external/bsd/netpgp/dist/src/hkpd/Makefile
U src/crypto/external/bsd/netpgp/dist/src/hkpd/hkpd.1
cvs update: src/crypto/external/bsd/netpgp/dist/src/hkpd/hkpd.8 is no longer in the repository
P src/crypto/external/bsd/netpgp/dist/src/lib/netpgp.c
P src/crypto/external/bsd/netpgp/dist/src/lib/version.h
P src/crypto/external/bsd/netpgp/dist/src/netpgp/Makefile
P src/crypto/external/bsd/netpgp/dist/src/netpgpkeys/netpgpkeys.c
P src/crypto/external/bsd/netpgp/dist/src/netpgpverify/Makefile
U src/crypto/external/bsd/netpgp/hkpc/Makefile
U src/crypto/external/bsd/netpgp/hkpd/Makefile
P src/crypto/external/bsd/netpgp/netpgpkeys/Makefile
P src/distrib/sets/lists/comp/mi
P src/lib/libperfuse/ops.c
P src/lib/librump/Makefile
P src/lib/librump/rump.3
U src/lib/librump/rump_lwproc.3
P src/sys/arch/sparc/stand/common/promdev.c
P src/sys/lib/libsa/loadfile_elf32.c
P src/tests/dev/scsipi/Makefile
P src/tests/fs/nfs/nfsservice/rumpnfsd.c
(Continue reading)

Permalink | Reply | 0 comments |
headers
Jes | 3 Sep 2010 11:16
Picon
Gravatar

estd and P-States

Hi all:

I've been searching information about Enhanced Speed Step and P-States. I currently use 'estd' to control
the cpu frequency, using '-s' option (smooth). I would launch 'estd' with '-b' option when the laptop is
running on battery, but I'm not sure if this is supported. Why? well, I can't find any 'sysctl machdep'
related to P-States.

In '/etc/powerd/scripts/acadapter' I've seen this piece of code:

        # Enable power saving mode for speedstep CPUs
        /sbin/sysctl -w machdep.speedstep_state=0 >/dev/null 2>&1

But in my laptop there's no such 'machdep.speedstep_state'.

My laptop runs 5.99.39 i386 with 'options ENHANCED_SPEEDSTEP' in the
kernel.

Is there any information about how to configure the C2 state when on battery in NetBSD 5.99?

Thanks in advance,

Jes
Permalink | Reply | 5 comments |
headers
Jukka Ruohonen | 3 Sep 2010 14:22
Picon
Picon
Favicon

Re: estd and P-States

On Fri, Sep 03, 2010 at 11:16:02AM +0200, Jes wrote:
> I've been searching information about Enhanced Speed Step and P-States. I
> currently use 'estd' to control the cpu frequency, using '-s' option
> (smooth).  I would launch 'estd' with '-b' option when the laptop is
> running on battery, but I'm not sure if this is supported.

You can easily achieve this by modifying the mentioned acadapter-script. For
instance, when the "released" event is called in powerd(8), stop the estd(1)
daemon and relaunch it with the desired option.

> Why?  well, I can't find any 'sysctl machdep' related to P-States.

Unfortunately there has been little planning with different P-state
technologies. Depending on the machine, the sysctl-node is either
machdep.speedstep_state, machdep.est, or machdep.powernow. Possibly some
non-x86 architectures add their own buttons to the mess.

Hopefully these sysctl-knobs and the estd(1) daemon will be some day
deprecated in favour of machine-independent in-kernel implementation.

> In '/etc/powerd/scripts/acadapter' I've seen this piece of code:
> 
>         # Enable power saving mode for speedstep CPUs
>         /sbin/sysctl -w machdep.speedstep_state=0 >/dev/null 2>&1
> 
> But in my laptop there's no such 'machdep.speedstep_state'.

Admittedly, the powerd(8) scripts need some adjustment. (For instance, a
comment in the acadapter-script says that it is a "generic script", but the
script itself goes and does some x86 and vendor-specific things.) I recommend
(Continue reading)

Permalink | Reply | 4 comments |
headers
Jes | 3 Sep 2010 16:45
Picon
Gravatar

Re: estd and P-States

Thanks for your quick answer Jukka. I plan to code some custom scripts
to deal with 'estd' for ac/bat modes. 

> > Is there any information about how to configure the C2 state when on
> > battery in NetBSD 5.99?
> 
> Anything above C1 is not currently supported on NetBSD. As you run
> -current, I recommned using the new acpicpu(4) driver (which is
> presently uncommented in the normal GENERIC kernels).
> 
> - Jukka.

Ok, I've activated 'acpicpu' in the kernel, but now... what? :) Sorry,
I don't know how to deal with. I can see it in 'dmesg':

acpicpu0 at acpi0 (CPU0): ACPI CPU
acpicpu1 at acpi0 (CPU1): ACPI CPU

And I've seen in a web page that someone did:

	sysctl -w machdep.idle-mechanism=mwait

But in my case I always get the same response:

	sysctl -w machdep.idle-mechanism=mwait
	sysctl: mwait: string too long

I've tried it with 'halt' and 'hlt' values, but no luck.

Thanks in adavance,
(Continue reading)

Permalink | Reply | 3 comments |
headers
Jukka Ruohonen | 3 Sep 2010 17:48
Picon
Picon
Favicon

Re: estd and P-States

On Fri, Sep 03, 2010 at 04:45:03PM +0200, Jes wrote:
> 	sysctl -w machdep.idle-mechanism=mwait
> 	sysctl: mwait: string too long
> 
> I've tried it with 'halt' and 'hlt' values, but no luck.

As I wrote, C-states are not currently supported. (Well, C0 and C1 are, but
that this is semantics.) Even when these will be supported, you will not
need to do anything to use these. That sysctl(8) is only for informative
purposes.

- Jukka.
Permalink | Reply | 2 comments |

Gmane