On Thu, May 03, 2018 at 11:32:42AM +0200, Mark Kettenis wrote:
> > Date: Thu, 3 May 2018 08:19:01 +0100
> > From: Dimitris Papastamos <s...@2f30.org>
>
> Hi Dimitris,
>
> > On Tue, May 01, 2018 at 11:55:00AM +0200, Mark Kettenis wrote:
> > > So aft
On Tue, May 01, 2018 at 11:55:00AM +0200, Mark Kettenis wrote:
> So after adding a quick hack to mitigate Spectre variant 2 to ARM
> Trusted Firmware (ATF), ARM actually designed a proper solution that
> minimizes the performance loss and makes the presence of the
> workaround detectable. This is
On Thu, May 03, 2018 at 08:19:01AM +0100, Dimitris Papastamos wrote:
> On Tue, May 01, 2018 at 11:55:00AM +0200, Mark Kettenis wrote:
> > So after adding a quick hack to mitigate Spectre variant 2 to ARM
> > Trusted Firmware (ATF), ARM actually designed a proper solution tha
> + case CPU_ID_CORTEX_A15:
> + case CPU_ID_CORTEX_A57:
> + case CPU_ID_CORTEX_A72:
> + /*
> + * Vulnerable; BPIALL is "not effective" so must use
> + * ICIALLU and hope the firmware set the magic bit in
> + * the ACTLR that actually
Hi,
Typo fix, below.
diff --git a/share/man/man5/pf.conf.5 b/share/man/man5/pf.conf.5
index 3559803..955539b 100644
--- a/share/man/man5/pf.conf.5
+++ b/share/man/man5/pf.conf.5
@@ -368,7 +368,7 @@ Translates to the point-to-point interface's peer
address(es).
.Pp
Host names may also have the
On Fri, Dec 30, 2016 at 11:30:10AM -0700, Theo de Raadt wrote:
> I'm wondering if anyone doing an install/upgrade has noticed any
> behaviour changes in the last week...
>
> There's a secret diff being tested :-)
i was building from source but just tried a snapshot
is it a scheduler or
On Tue, Dec 13, 2016 at 12:35:07PM +0100, Martin Pieuchot wrote:
> Hello,
>
> On 10/12/16(Sat) 11:08, Dimitris Papastamos wrote:
> > While I was playing around with the ip4 multicast code, I thought
> > I would attempt to make the membership data structure similar to that
&g
On Sun, Dec 11, 2016 at 07:24:40PM +, Dimitris Papastamos wrote:
> Hi,
>
> I converted the igmp stats to use percpu counters. This work is
> basically the same as what dlg@ did for other parts of the stack.
> I looked at the diff and adjusted it for igmp.
ping
Hi,
I converted the igmp stats to use percpu counters. This work is
basically the same as what dlg@ did for other parts of the stack.
I looked at the diff and adjusted it for igmp.
diff --git a/sys/netinet/igmp.c b/sys/netinet/igmp.c
index 11446ce4188..c2a0a4839b4 100644
---
Hi,
While I was playing around with the ip4 multicast code, I thought
I would attempt to make the membership data structure similar to that
of ip6. This means changing from a dynamic array to a linked list.
The max membership limit has been lifted (I did not see a similar limit
in the ip6
Hi everyone,
I am working on a program that requires putting the node in
all-multicast mode. The program needs to see all the packets going to
solicited-node multicast addresses. I did this by modifying the
interface flags to add IFF_ALLMULTI.
Unfortunately recvmsg() was only returning for the
On Fri, Sep 16, 2016 at 09:09:44AM -0400, Ted Unangst wrote:
> Dimitris Papastamos wrote:
> > Hi everyone,
> >
> > I've put together a patch for 6.0-stable that adds domain name
> > matching support to rebound(8). The patch is quite rough at the
> > moment.
Hi everyone,
I've put together a patch for 6.0-stable that adds domain name
matching support to rebound(8). The patch is quite rough at the
moment.
The config is as follows:
match "local." 10.0.0.53
match "." 8.8.8.8
Requests to foo.local. are sent over to 10.0.0.53, all other
Hi,
I've converted ikectl(8) from getpass to readpassphrase.
===
RCS file: /cvs/src/usr.sbin/ikectl/ikeca.c,v
retrieving revision 1.40
diff -u -p -r1.40 ikeca.c
--- ikeca.c 2 Nov 2015 12:21:27 - 1.40
+++ ikeca.c 3
Hi,
encrypt(1) updated to use readpassphrase(3) instead of getpass(3).
Thought I would grep the tree after the recent commit to passwd(1)
by tedu.
Index: encrypt.c
===
RCS file: /cvs/src/usr.bin/encrypt/encrypt.c,v
retrieving
Hi,
I am seeing a build failure on -current when I build xenocara. I've
already built /usr/src so that's up to date.
/usr/xenocara/lib/mesa/src/mesa/main/varray.c:168: error: no previous prototype
for '_mesa_bind_vertex_buffer'
/usr/xenocara/lib/mesa/src/mesa/main/varray.c:280: error: no
On Mon, Apr 18, 2016 at 10:50:46AM +0200, Martin Pieuchot wrote:
> The current goal of the Network SMP effort is to have a single CPU
> process the IP forwarding path in a process context without holding
> the KERNEL_LOCK(). To achieve this goal we're progressively moving
> code from the softnet
On Wed, Mar 30, 2016 at 02:49:06PM +0200, Mike Belopuhov wrote:
> Good day, Dimitris.
>
> Long time ago in a galaxy far far away I've been using this
> alongside the -F option that I've added. While managed
> switches are becoming cheaper, I don't see a reason for a
> working feature to go away,
Hi everyone,
I totally forgot about this patch. At the time it couldn't go in
because the tree was locked. Does it make sense? If so I will check
whether it applies on -current and resubmit.
On Fri, Jul 31, 2015 at 01:55:07PM +0100, Dimitris Papastamos wrote:
> Hi every
On Fri, Jan 15, 2016 at 12:53:32PM -0700, Theo de Raadt wrote:
> Nope.
>
> The manual page is correct.
>
> > Small correction.
> >
> > ===
> > RCS file: /cvs/src/share/man/man7/hier.7,v
> > retrieving revision 1.150
> > diff -u -p
Small correction.
===
RCS file: /cvs/src/share/man/man7/hier.7,v
retrieving revision 1.150
diff -u -p -r1.150 hier.7
--- hier.7 7 Jan 2016 08:58:14 - 1.150
+++ hier.7 15 Jan 2016 19:49:16 -
@@ -49,8 +49,6 @@
On Mon, Dec 28, 2015 at 09:13:34PM +1000, David Gwynne wrote:
> this builds on jmatthew@'s last commit and adds mpsafe tx.
>
> ive beat on it pretty hard, but more eyes/tests are appreciated.
>
> ok?
Been running with this on my router since yesterday without issues.
I will continue to run with
On Sat, Dec 19, 2015 at 01:08:26PM +0100, Stefan Sperling wrote:
> On Fri, Dec 18, 2015 at 05:40:39PM -0500, David Hill wrote:
> > With sthen@'s patch I can associate, dhcp, and use net.
>
> Here's an updated iwn diff with a better approach for Stuart's fix.
>
> Thanks for helping, Stuart, and
On Sat, Dec 05, 2015 at 06:11:51PM +0100, Jonathan Matthew wrote:
> The main interesting bit here is the txeof and start loops, which previously
> operated based on the prod/cons indices and the contents of the tx queue,
> but now just uses the indices as that's the only way to get a consistent
to get some feedback.
Cheers,
Dimitris
>From 86c7ad057137181b560341779c79bca339b30e95 Mon Sep 17 00:00:00 2001
From: Dimitris Papastamos <s...@2f30.org>
Date: Wed, 28 Oct 2015 14:45:03 +
Subject: [PATCH] Add a simple roundrobin load balancing feature to rebound(8)
---
usr.sbin/rebound/
On Tue, Oct 27, 2015 at 05:20:31PM -0400, Ted Unangst wrote:
> Dimitris Papastamos wrote:
> > There was a comment in the code that indicated that it might be worth
> > investigating the use of trees. I have not currently done any kind of
> > serious benchmarking on
Hi,
There was a comment in the code that indicated that it might be worth
investigating the use of trees. I have not currently done any kind of
serious benchmarking on this but I am looking into it.
===
RCS file:
On Wed, Sep 23, 2015 at 12:12:01PM -0400, Ted Unangst wrote:
> Dimitris Papastamos wrote:
> > I did some very crude benchmarking on my machine and I could push
> > about 100k messages per second through a single message queue.
> >
> > At this point, I can start working
second through a single message queue.
At this point, I can start working on manpages + the necessary
userspace glue.
Is this patch going in the right direction?
Thanks,
Dimitris
>From a4a59e77788b6ac9340c5707b04cd924ab23c923 Mon Sep 17 00:00:00 2001
From: Dimitris Papastamos <s...@2f30.org
is that not much
software is using it currently.
Let me know what you think!
Thanks,
Dimitris
>From 6b2ed6b728de3c2ca8ab17fca527303b44840ba0 Mon Sep 17 00:00:00 2001
From: Dimitris Papastamos <s...@2f30.org>
Date: Tue, 15 Sep 2015 16:54:34 +0100
Subject: [PATCH] Initial implementation of mq_*() i
Hi,
A slightly updated patch. Removed some checks that were
redundant.
Index: kern/init_sysent.c
===
RCS file: /cvs/src/sys/kern/init_sysent.c,v
retrieving revision 1.171
diff -u -p -r1.171 init_sysent.c
--- kern/init_sysent.c 9
Hi everyone,
I've done a basic implementation of posix_fallocate(2). I've looked
at both the FreeBSD and NetBSD code when doing this. My knowledge of
the VFS layer is very limited, so please let me know if anything looks
wrong.
I added a generic function in vfs_vnops.c called vn_fallocate()
On Sat, Sep 12, 2015 at 06:57:48PM +0100, Dimitris Papastamos wrote:
> + switch (fp->f_type) {
> + case DTYPE_VNODE:
> + break;
> + case DTYPE_PIPE:
> + case VFIFO:
VFIFO shouldn't be here, I have moved it further down after vn_lock
to check it against v_type.
Hi all,
I have a trunk failover setup between my wired and wireless
connection. My wired is master and my wireless is slave.
My config is as follows:
==> /etc/hostname.em0 <==
up
==> /etc/hostname.iwn0 <==
nwid FOO
wpakey XXX
up
==> /etc/hostname.trunk0 <==
trunkproto failover
trunkport em0
On Sat, Sep 05, 2015 at 12:04:20PM +0100, Dimitris Papastamos wrote:
> Scenario 1 - boot with em0 active
> 1) pinging via em0 works
> 2) switch to iwn0, ping doesn't work <-- problem
> 3) switch to em0, ping works
> 4) sh /etc/netstart
> 5) pinging via em0 works
> 6) sw
On Sat, Sep 05, 2015 at 12:39:21PM +0100, Stuart Henderson wrote:
> On 2015/09/05 12:13, Dimitris Papastamos wrote:
> > On Sat, Sep 05, 2015 at 12:04:20PM +0100, Dimitris Papastamos wrote:
> > > Scenario 1 - boot with em0 active
> > > 1) pinging via em0 works
> >
Hi,
You can reproduce with ping -s 0 host (it crashes if the packet size
is less than 24).
===
RCS file: /cvs/src/sbin/ping/ping.c,v
retrieving revision 1.123
diff -u -p -r1.123 ping.c
--- ping.c 2 May 2015 18:03:37 -
On Sat, Jul 25, 2015 at 09:20:18PM +0200, Martin Pieuchot wrote:
On 13/07/15(Mon) 14:04, Dimitris Papastamos wrote:
Hi,
I noticed -f in ndp(8) did nothing at all so I've enabled it and
documented the file syntax in the man page.
If it does nothing, I'd say let's kill it.
Index: ndp
On Mon, Jul 13, 2015 at 02:04:40PM +0100, Dimitris Papastamos wrote:
Hi,
I noticed -f in ndp(8) did nothing at all so I've enabled it and
documented the file syntax in the man page.
===
RCS file: /cvs/src/usr.sbin/ndp/ndp.8,v
Hi,
Just a minor stylistic change.
Cheers,
Dimitris
===
RCS file: /cvs/src/usr.bin/doas/parse.y,v
retrieving revision 1.7
diff -u -p -r1.7 parse.y
--- parse.y 21 Jul 2015 11:04:06 - 1.7
+++ parse.y 21 Jul 2015
Hi,
I noticed -f in ndp(8) did nothing at all so I've enabled it and
documented the file syntax in the man page.
===
RCS file: /cvs/src/usr.sbin/ndp/ndp.8,v
retrieving revision 1.33
diff -u -p -r1.33 ndp.8
--- ndp.8 3 Sep 2014
On Mon, Apr 27, 2015 at 11:18:55AM +0200, Martin Pieuchot wrote:
On 22/04/15(Wed) 10:29, Dimitris Papastamos wrote:
[...]
On a side note, What do you think of using a sysctl instead of an ioctl?
I guess an ioctl will do but it would make it easy for people to disable
this at boot time
Below is an updated diff. I will add the documentation bits
soonish once I've tested all possible configurations as you
requested.
I am also planning to rework the structure of usbdevs.c as it is
a bit confusing currently but that will happen incrementally after
this diff.
Index:
On Wed, Apr 22, 2015 at 08:54:34AM +0200, Martin Pieuchot wrote:
On 17/04/15(Fri) 16:47, Dimitris Papastamos wrote:
Hi,
This patch adds an option to usbdevs(8) to disable USB bus probing
at runtime. The operation is restricted to the root user.
It would be nice to show if probing
Hi,
This patch adds an option to usbdevs(8) to disable USB bus probing
at runtime. The operation is restricted to the root user.
I am not sure if this approach is sensible or even correct. Some
pointers would be much appreciated.
This was started as part of a reply by mpi on tech@
On Thu, Apr 16, 2015 at 12:06:49PM +0200, Martin Pieuchot wrote:
On 13/04/15(Mon) 17:49, Dimitris Papastamos wrote:
Hi,
Attached below an updated patch that makes the layout a bit more sensible.
It's nice it works.
Now like various students that applied for the GSoC you've done
ping
Hi everyone,
This is a preliminary patch that adds descriptor dumping support
to usbdevs(8). I have attempted to keep any unnecessary changes to
a minimum. I will do some cleanup work incrementally on top of
this patch.
I've tested this on a single machine running -current so more testing
Hi,
Attached below an updated patch that makes the layout a bit more sensible.
Index: usr.sbin/usbdevs/usbdevs.8
===
RCS file: /cvs/src/usr.sbin/usbdevs/usbdevs.8,v
retrieving revision 1.9
diff -u -p -r1.9 usbdevs.8
---
One more thing I forgot to mention, the current printing layout
is not ideal. I'll look into making it a bit more readable.
On Sat, Jan 31, 2015 at 08:25:23PM +1100, Jonathan Gray wrote:
Some test reports with this would be great.
In particular for radeon.
Been running this patch with i915 for 2 days now without any
perceivable issues.
OpenBSD 5.7-beta (GENERIC.MP) #23: Tue Feb 3 17:07:19 GMT 2015
Hi,
This is my attempt at converting bgpd to use siphash. This does
*not* link as is because the userland implementation of siphash
is currently missing.
In certain cases where multiple data is hashed in succession, it might
have been better to use a siphash context and call SipHash24_Update().
Admittedly it would have been better to store the key in the struct
and then use arc4random_buf() only once or so. I am not familiar with the
code so I will investigate this option.
On Thu, Dec 04, 2014 at 12:52:04PM -0500, Ted Unangst wrote:
Ah, so this won't work. That's not how siphash works.
The key needs to last as long as the hash table does, which is
generally as long as the program. Different keys result in different
hashes, which makes it hard to look things up
Hi everyone,
Any interest in this? It is a cleanup change and not intended to have
any functional differences.
Cheers,
Dimitris
On Tue, Nov 11, 2014 at 02:21:35PM +0400, Vadim Zhukov wrote:
Why TAILQ? SLIST should be enough.
SLIST makes sense yes.
Also, moving sys/types.h is wrong.
OK, will fix.
Why do you want to get this in? Any more ac(8) patches coming?
Not really, I was just randomly browsing through the
Hi everyone,
I just built the kernel with XHCI_DEBUG enabled. I am having
the following problem:
1) boot the machine with keyboard and mouse plugged in
2) keyboard does not respond at the login prompt
3) plug it out and plug it back in at the same USB port
4) keyboard does not respond
5) plug
Hi everyone,
I had to type this off the screen. My USB keyboard does not work
at the ddb prompt so I cannot provide more information.
xhci_pipe_open: pipe=0x80493000 addr=2 depth=1 port=9 speed=2
xhci0: dev 1 dci 3 (epAddr=0x81)
xhci0: speed 1 mps 8 rhport 9 route 0x0
xhci0: max ESIT
Hi,
../../../../dev/usb/xhci.c:1109
2c26: 4a 8d 14 ba lea(%rdx,%r15,4),%rdx
2c2a: 48 8d 44 10 01 lea0x1(%rax,%rdx,1),%rax
2c2f: 49 8b 84 c5 b8 05 00mov0x5b8(%r13,%rax,8),%rax
2c36: 00
2c37: 89 08
On Sun, Nov 09, 2014 at 03:01:47PM +0100, Martin Pieuchot wrote:
Thanks for the report. I just committed a fix for this. The problem
was in the code closing the pipe. This would only matter for devices
closing opening multiple times their pipes, like mouses or keyboards
when they are
Hi,
I've attempted to convert the hand-rolled linked lists over to
TAILQ. I've tested this briefly and it seems to work okay.
Does it look good? Let me know if I need to rework anything.
===
RCS file: /cvs/src/usr.sbin/ac/ac.c,v
Hi,
It seems to me that we should not free `pfd' at this point. The saved
max poll fd is not reset to 0 and I do not see any guarantees that `pfd'
will point to valid memory after calling free() here.
Other code that follows the same style, like mountd(8) and rpc.rstatd(8)
do not do this.
OK?
diff -N sys/mkfifo.c
--- /dev/null 1 Jan 1970 00:00:00 -
+++ sys/mkfifo.c23 Oct 2014 14:14:13 -
@@ -0,0 +1,24 @@
+/*
+ * Copyright (c) 2014 Dimitris Papastamos s...@2f30.org
+ *
+ * Permission to use, copy, modify, and distribute this software for any
+ * purpose with or without fee
On Thu, Oct 23, 2014 at 11:12:02AM -0400, Ted Unangst wrote:
On Thu, Oct 23, 2014 at 15:16, Dimitris Papastamos wrote:
Hi everyone,
This is my attempt to add libc wrappers for mkfifo() and
mkfifoat() using mknodat().
Is there a reason? This would seem to do nothing except complicate
Hi everyone,
I've tested this on my home setup and it seems to work.
Let me know if anything else needs working.
Cheers,
Dimitris
===
RCS file: /cvs/src/sbin/mountd/mountd.c,v
retrieving revision 1.76
diff -u -p -r1.76 mountd.c
On Mon, Oct 06, 2014 at 04:56:08PM -0600, Todd C. Miller wrote:
The following diff should fix it.
- todd
Index: sys/miscfs/fifofs/fifo_vnops.c
===
RCS file: /home/cvs/openbsd/src/sys/miscfs/fifofs/fifo_vnops.c,v
retrieving
Hi everyone,
I am using select(2) on a FIFO fd and monitoring
for readability. select(2) doesn't return after the writer
exits.
The same piece of code marks the fd as readable on Linux.
Not sure which behaviour is correct though.
Doing a very similar test with stdin and closing it before
the
On Mon, Oct 06, 2014 at 01:08:57PM +0100, Dimitris Papastamos wrote:
The same piece of code marks the fd as readable on Linux.
I tested on NetBSD and it also seems to mark the fd readable.
Doing a very similar test with stdin and closing it before
the call to select(2) behaves as expected
On Sat, Aug 02, 2014 at 11:07:30PM +0200, Fritjof Bornebusch wrote:
On Wed, Jul 30, 2014 at 09:26:54PM +0100, Dimitris Papastamos wrote:
On Wed, Jul 30, 2014 at 10:19:19PM +0200, Fritjof Bornebusch wrote:
+ tzone = (int)strtonum(h, -23, 23, errstr);
The explicit cast
On Wed, Jul 30, 2014 at 10:19:19PM +0200, Fritjof Bornebusch wrote:
+ tzone = (int)strtonum(h, -23, 23, errstr);
The explicit cast is not needed here.
On Wed, May 07, 2014 at 12:13:42PM -0700, patrick keshishian wrote:
On Wed, May 07, 2014 at 02:47:32PM +0100, Dimitris Papastamos wrote:
Hi,
This piece of code now uses poll() instead of select().
I have not got round to test this yet, but I will as soon as I have
a working setup
Hi all,
* No need to check for NULL before free.
* No need to cast the return value of the allocation functions.
* Do not use malloc() + memset(), simply use calloc()
Tested against the regression tests for pqueue.
ok?
===
---
Hi,
Simplify pqueue_find() to the very essential bits. This change should
exhibit no functional changes or side-effects.
Tested against the regression tests for pqueue.
ok?
===
--- pqueue.orig/pqueue.c
+++ pqueue/pqueue.c
@@
Hi,
* Use sizeof() instead of hardcoding the size.
Tested against the regression tests for pqueue.
ok?
===
--- pqueue.orig/pqueue.c
+++ pqueue/pqueue.c
@@ -160,7 +160,8 @@ pqueue_find(pqueue_s *pq, unsigned char
pitem
Hi,
This piece of code now uses poll() instead of select().
I have not got round to test this yet, but I will as soon as I have
a working setup.
Thoughts?
===
RCS file: /cvs/src/usr.sbin/amd/amd/info_hes.c,v
retrieving revision
On Tue, Apr 29, 2014 at 07:57:32PM -0600, Theo de Raadt wrote:
Don't bother with diffs to b_sock.c. Instead, if you have code
which uses it, talk to krw.
There is a monster diff coming which rewrites it all.
And by the way, all that code disapears and is replaced by 2 lines.
That's
Not sure this is sensible as it encourages people to simply
update the table.
I was inclined to remove the code entirely but I am not sure
what broken systems might rely on this.
Only build tested.
Thoughts?
Index: b_sock.c
===
77 matches
Mail list logo