gmane.os.netbsd.ports.evbmips

1 new photo on MyDailyFlog!

girish vg — 2013-01-26T12:09:24

Hi!
I would like you to come and see my latest photos on MyDailyFlog. 

Check out: http://www.mydailyflog.com/go/invite_register/girishgulawani/48268548&stc=71

Thanks!
girish vg



~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
To unsubscribe of this type of email from MyDailyFlog in the future,
please click below:
http://www.mydailyflog.com/un/port-evbmips< at >netbsd.org&md5=6f0a8b90c8d9a86e&bl=16

Please do not reply directly to this email. Questions? Contact us - 
http://www.mydailyflog.com/go/contact_us

MyDailyFlog, Refriendz Ltd. PO BOX 1184, Luton, Bedfordshire, LU1 9AT.

Check out this photo on MyDailyFlog!

girish vg — 2013-01-26T12:09:23

Hi!
I would like to invite you to visit MyDailyFlog and see my latest photos. 

Check out: http://www.mydailyflog.com/go/invite_register/girishgulawani/48268548&stc=16

Cheers!

girish vg


========================================
Got a digital camera?

MyDailyFlog is a personal photo-blogging space where you can easily post
your latest and greatest photos, and share them with your friends and family.

Create your own DailyFlog at www.MyDailyFlog.com





~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Unsubscribe: to opt out of further invitations from your friends to see 
their DailyFlogs, please click below:
http://www.mydailyflog.com/un/port-evbmips< at >netbsd.org&md5=6f0a8b90c8d9a86e&bl=16

Please do not reply directly to this email. Questions? Contact us - 
http://www.mydailyflog.com/go/contact_us

MyDailyFlog, Refriendz Ltd. PO BOX 1184, Luton, Bedfordshire, LU1 9AT.

Netbsd on Atheros AR71xx

mac c — 2012-12-09T03:59:23

Hi,
I would like to run NetBSD on cpu Atheros 7130,
Netbsd 6 has support for this soc?


Marcos

Re: bpfjit

Alexander Nasonov — 2012-11-24T19:58:46

Sure, I can do it but this will make MATLA special only because I
happened to find an installation instruction. I'd be nice to commit
this for all configs that support bpfjit but I don't have any
expertise in mips world.

If anybody has a list of configs that are still actively used and that
are easy to run in an emulator, I can test them and commit.

Alex

Re: bpfjit

Christos Zoulas — 2012-11-24T15:38:05

In article <20121124142253.GA2412< at >x1000.localdomain>,
Alexander Nasonov  <alnsn< at >yandex.ru> wrote:

I would commit that with the options commented out.


christos

bpfjit

Alexander Nasonov — 2012-11-24T14:22:53

After a successful testing of bpfjit on an emulated MALTA board, I
committed sljit support to evbmips. It's disabled my default but you
can easily enable it. To reproduce my setup, you need to apply the patch
below and build with MKSLJIT=yes.
I followed an instruction on gxemul page to install and run my system
under gxemul emulator. I also added "net.bpf.jit=1" to sysctl.conf
while following an installation instruction.
All sljit and bpfjit userspace tests pass and the system can filter
packets using simple rules.

Alex


Enable bpfjit on MALTA ebvmips board.

Index: sys/arch/evbmips/conf/MALTA
===================================================================
RCS file: /cvsroot/src/sys/arch/evbmips/conf/MALTA,v
retrieving revision 1.69
diff -p -u -u -r1.69 MALTA
--- sys/arch/evbmips/conf/MALTA 17 Oct 2012 14:48:11 -0000      1.69
+++ sys/arch/evbmips/conf/MALTA 24 Nov 2012 14:05:30 -0000
< at >< at > -307,6 +307,9 < at >< at > audio*              at clct?
 #pseudo-device md                              # memory disk device
(

How stable is the Atheros AR9344 support?

Lloyd Parkes — 2012-10-06T22:08:49

Hi all,
I'm looking at replacing my NetBSD based home router with a MikroTik RB2011UAS-2HnD because I need more throughput and the routerboards are cheap. I'm sure the included RouterOS works fine, but I wouldn't mind running NetBSD instead. I realise that NetBSD hasn't been ported to RB2011 devices, but they are cheap enough that I'm willing to buy one to tinker with. 

So, how stable is the Atheros AR9344 support? 

Cheers,
Lloyd

Re: System time and ToD clock?

John D. Baker — 2012-06-22T04:50:11

Having gotten around to UTSL, I see that evbmips/isa/mcclock_isa.c
asserts that PMON uses a base+offset strategy for handing year values,
with an epoch year of 1920.  That would explain the "1992" assertion of
gNewSense.

Maybe this was true of older versions of PMON or the implementation on
other boards, but that used in in the Lemote systems (PMON2000, IIUC)
seems to use a different epoch (as seen in OpenBSD/loongson files).

Perhaps the setting of "sc->sc_year0" can be made conditional on the
machine type (either at compile time or runtime)?

System time and ToD clock?

John D. Baker — 2012-06-21T07:04:54

Not having seen any documentation about this (and I'd not yet gotten to
UTSL), I was a bit surprised when gNewSense reported that my Yeeloong's
ToD clock was set to 1992, making the filesystems have dates in the
future.

Dropping into PMON at the next boot confirmed that the day/month/time
was correct, but was 20 years in the past.  I set it back to the current
year.

When I booted NetBSD again, it declared "Preposterous time in ToD clock"
and initialized system time from the filesystem timestamps.  Shortly
therafter 'ntpd' was started to keep time sane (thanks to now-working
ioctls).

When, after a panic (freeing free inode on an ext2fs filesystem), I
again booted gNewSense to clean up.  Once again it reported a year of
1992.  Returning to NetBSD once more elicited no warnings.

This is reminiscent of the mvme68k port where system ToD clock holds
only a 2-digit year taken as an offset from the epoch (1968 in the
case of NetBSD/mvme68k).

Is this also the case for the evbmips port?  While this might be fine

Re: sshd session hang, bad file descriptor data

Christos Zoulas — 2012-06-09T01:19:20

In article <Pine.NEB.4.64.1206081415270.22128< at >david.technoskunk.fur>,
John D. Baker <jdbaker< at >mylinuxisp.com> wrote:

This is exactly what I see on my shark using current. It is probably
not the same bug, but there is a time during the negotiation process
where all the signals are blocked. can you run ssh in gdb?

christos

sshd session hang, bad file descriptor data

John D. Baker — 2012-06-08T19:30:44

Attempts to connect to my Yeeloong via ssh from an i386 client still
hang.  I'm not getting any console messages, but that may be a side
effect from my 'amd' investigation.

The symptoms from the client end remain the same.  The server prompts
for password and no further operations are possible on the client end.
the client terminal is locked into an unresponsive network connection.
The usual "~." doesn't work.  In a graphical environment, one must
destroy the window.  (I haven't tried from a non-graphical client, but
I expect one must log in to the client from elsewhere and kill the
frozen session.)

Below are what may be relevant snippets of ktruss output generated with
'sudo ktruss -i -p <pid> | tee sshd.ktruss'.  In this instance, the PID
of the master sshd process was 630.  There were no messages relevant
to this logged in /var/log/messages.

The output is some different than before.  Before, I saw a distinct
pattern of user-after-close.  This time, an attempt is made to close
a file descriptor of -1 (0

'amd' Invalid argument data

John D. Baker — 2012-06-08T18:54:23

So, getting back to the problem with 'amd' failing to mount my
various automount points with "Invalid argument" reports, below
is the relevant excerpt from '/var/log/messages':

Jun  8 13:21:59 chalk amd[263]: '/home': mount: Invalid argument
Jun  8 13:21:59 chalk amd[685]: '/m': mount: Invalid argument
Jun  8 13:22:04 chalk /netbsd: ktrace timeout
Jun  8 13:22:04 chalk amd[264]: '/net': mount: Invalid argument
Jun  8 13:22:04 chalk amd[264]: '/net': mount: Invalid argument
Jun  8 13:22:05 chalk amd[264]: amfs_toplvl_mount: amfs_mount failed: Invalid argument
Jun  8 13:22:05 chalk amd[881]: /net: mount (amfs_cont): Invalid argument
Jun  8 13:22:10 chalk amd[263]: '/home': mount: Invalid argument
Jun  8 13:22:10 chalk amd[685]: '/m': mount: Invalid argument
Jun  8 13:22:10 chalk amd[685]: amfs_toplvl_mount: amfs_mount failed: Invalid argument
Jun  8 13:22:10 chalk amd[881]: /m: mount (amfs_cont): Invalid argument
Jun  8 13:22:10 chalk amd[263]: amfs_toplvl_mount: amfs_mount failed: Invalid argument
Jun  8 13:

Latest try with Yeeloong

John D. Baker — 2012-06-08T02:57:30

I've finally managed to build a more recent release and update.  Glad to
see all those wireless drivers ported from OpenBSD.

Before getting back to poking at the compat32 and sshd problems, I was
updating local pkgsrc tree and the machine panicked claiming that it
was asked to free an already free block.  I'll need to set ddb.onpanic=1
so it can stop long enough for me to transcribe.  In the meantime, some
details:

The filesystem was ext2fs (shared with gnewSense on hard disk) under
heavy I/O load.

On reboot, drop into gnewSense so its fsck can clean up, then boot
NetBSD again.  Swap space is also shared, but savecore couldn't
find anything and complained about "bad namelist" or some such.  I'll
try to extract the errors soon.

Re: Lemote Yeeloong observations

Christos Zoulas — 2012-05-14T19:23:41

In article <Pine.NEB.4.64.1205140447270.12898< at >david.technoskunk.fur>,
John D. Baker <jdbaker< at >mylinuxisp.com> wrote:

Yes, it seems that some 32 bit translations are maybe wrong. Perhaps
I should try to run a 32 bit userland with a 64 bit kernel on amd64
and see how that goes :-)

But providing kdump snippets of the failed syscalls can show us where
the problems lie.

christos

Re: Lemote Yeeloong observations

John D. Baker — 2012-05-14T09:57:41

On Fri, 11 May 2012 16:59:52 +0000 (UTC), christos< at >astron.com
(Christos Zoulas) wrote:


[snip]

I saw the equivalent come through source-changes< at >.  Updating and building
again, this problem appears to be solved.

A similar problem appears to be affecting 'amd'.  It spews out a series
of "Invalid argument" complaints for each automount point when starting
up.  I'll try to get details soon-ish (travelling soon).

There remains the problem of 'sshd' attempting to operate on a file
descriptor after it has already been closed.

Re: Lemote Yeeloong observations

Christos Zoulas — 2012-05-11T16:59:52

In article <jojfua$7u1$1< at >dough.gmane.org>,
Christos Zoulas <christos< at >astron.com> wrote:

Does this patch fix the problem for you?

christos

Index: netbsd32_ioctl.c
===================================================================
RCS file: /cvsroot/src/sys/compat/netbsd32/netbsd32_ioctl.c,v
retrieving revision 1.64
diff -u -p -u -r1.64 netbsd32_ioctl.c
--- netbsd32_ioctl.c6 Oct 2011 03:19:32 -00001.64
+++ netbsd32_ioctl.c11 May 2012 17:01:26 -0000
< at >< at > -52,6 +52,7 < at >< at > __KERNEL_RCSID(0, "$NetBSD: netbsd32_ioc
 #include <sys/kmem.h>
 #include <sys/envsys.h>
 #include <sys/wdog.h>
+#include <sys/clockctl.h>
 
 #ifdef __sparc__
 #include <dev/sun/fbio.h>
< at >< at > -76,63 +77,6 < at >< at > __KERNEL_RCSID(0, "$NetBSD: netbsd32_ioc
 
 #include <dev/vndvar.h>
 
-/* prototypes for the converters */
-static inline void netbsd32_to_partinfo(struct netbsd32_partinfo *,
-  struct partinfo *, u_long);
-#if 0
-static inline void netbsd32_to_format_op(struct netbsd32_format_op *,
-   struct format_op *, u_long);
-#endif
-stati

Re: Lemote Yeeloong observations

Christos Zoulas — 2012-05-11T16:49:14

In article <20120511071141.GC28474< at >mail.duskware.de>,
Martin Husemann  <martin< at >duskware.de> wrote:

The issue is that the ioctls are not being translated.

christos

Re: Lemote Yeeloong observations

Martin Husemann — 2012-05-11T07:11:41

This seems strange, the address should be that of a global variable in ntpd,
and thus should be properly aligned.

Could you instrument the code in src/sys/compat/netbsd32/netbsd32_time.c
and see if the EFAULT comes from the return (error); in line 155 and print
the pointer value?

Something like:

if ((error = copyin(SCARG_P32(uap, tp), &ntv32, sizeof(ntv32)))) {
printf("error %d pointer %p\n", error, (void*)SCARG_P32(uap, tp));
return error;
}

Martin

Re: Lemote Yeeloong observations

John D. Baker — 2012-05-11T05:52:49

On Thu, 10 May 2012 18:42:28 +0000 (UTC), christos%astron.com< at >localhost
(Christos Zoulas) wrote:


Trying to attach 'sudo ktrace -i -p <pid>' exited immediately, leaving
an empty trace file.  Using 'ktruss' instead and 'tee'ing its output
to a file (-o outfile caused immediate exit as well) produced the
following output:

   http://bobdbob.com/~jdbaker/nbsd-debug/sshd.ktruss.gz

which indicates that one of the child processes forked to handle the
connection obtains a file descriptor via netbsd32___socket(),
netbsd32_connect()s with it, does a netbsd32_sendto() on it,
netbsd32_close()s it _and then attempts to perform netbsd32_ioctl() on it
it_, resulting in EBADF.  It further attempts to netbsd32_dup() it,
again resulting in EBADF.  Possible compiler pessimization reorder some
calls?


Likewise, produced:

   http://bobdbob.com/~jdbaker/nbsd-debug/ntpd.ktruss.gz

Searching for EINVAL revealed calls to netbsd32_ioctl() with commands
"CLOCKCTL_CLOCK_SETTIME",  "CLOCKCTL_SETTIMEOFDAY", and "_IOWR" as
failing in

Re: Lemote Yeeloong observations

Christos Zoulas — 2012-05-10T18:42:28

In article <Pine.NEB.4.64.1205092317360.6809< at >david.technoskunk.fur>,
John D. Baker <jdbaker< at >mylinuxisp.com> wrote:

You don't need PTYFS if your kernel has COMPAT_BSDPTY.
ktrace it and see what it is trying to dup


Something is wrong with the ioctl? ktrace it.

christos

Re: Lemote Yeeloong observations

John D. Baker — 2012-05-10T04:46:47


The updated "security/sudo" runs without problems.  Adapting the patch
to "wip/sudo" allows it to work as well.  It would appear that although
built as "evbmips64-el", userland is actually 32-bit?  That would seem
to explain the behavior similar to sparc which doesn't occur on sparc64.

There appears to be a problem with pty allocation--at least as far as
'sshd' is concerned.  Attempting to connect to the Yeeloong (LOONGSON
kernel) hangs at the client end and produces the following on console:

   sshd[pid]: error: do_exec_pty: Dup #1: Bad file descriptor

I have the standard complement of /dev/pty??.  Is ptyfs required instead?

Also, running 'ntpd' produces a regular cascade of:

   ntpd[pid]: step-systime: Invalid argument

Not sure what this indicates.  Is 'clockctl' not included?  I'll have
to check the kernel config.

Xorg complains about a missing symbol/feature.  Log file here:

   http://bobdbob.com/~jdbaker/nbsd-debug/startx-fail-yeeloong.txt

Attempting to build "lang/perl" fails to produce an e