Re: OpenBSD 6.2 released: Oct 9, 2017

2017-10-09 Thread R0me0 ***
Long Life For Magic Puffer Fish! Without has, OpenBSD in a soul, it would be impossible to achieve the current magnitude of this amazing Operating System. Cheers guys! 2017-10-10 0:59 GMT-03:00 Theo de Raadt : > > newvers.sh in 6.2/sys.tar.gz says it's "-current". it

Re: OpenBSD 6.2 released: Oct 9, 2017

2017-10-09 Thread Theo de Raadt
> newvers.sh in 6.2/sys.tar.gz says it's "-current". it looks strange. .. > and, there is no top(root) CVS/ in 6.2/src.tar.gz. Thanks for noticing these errors. I am going to replace these two files. It may take a few hours for them to propogate.

armv7/panda usb eth addr wondering

2017-10-09 Thread Artturi Alm
Hi, i've been bothered by this: smsc0 at uhub1 port 1 configuration 1 interface 0 "Standard Microsystems SMSC9512/14" rev 2.00/2.00 addr 3 smsc0: address ff:ff:ff:ff:ff:ff that does happen no matter if i run dhcp-command in u-boot shell or not. u-boot does store the addr in env, but i don't

Re: OpenBSD 6.2 released: Oct 9, 2017

2017-10-09 Thread FUKAUMI Naoki
From: FUKAUMI Naoki Subject: Re: OpenBSD 6.2 released: Oct 9, 2017 Date: Tue, 10 Oct 2017 11:53:59 +0900 (JST) > newvers.sh in 6.2/sys.tar.gz says it's "-current". it looks strange. > > $ sha256 sys.tar.gz > SHA256 (sys.tar.gz) = >

Re: efiboot: Restore GOP mode on SetMode() failure

2017-10-09 Thread YASUOKA Masahiko
On Sat, 7 Oct 2017 09:24:20 +0200 Klemens Nanni wrote: > On Sat, Oct 07, 2017 at 01:15:40AM +, YASUOKA Masahiko wrote: >> > See my updated diff for reusing the gopi struct, please. >> >> ok, but the diff seems to be against wrong revision. You seems to >> have other diffs,

Re: OpenBSD 6.2 released: Oct 9, 2017

2017-10-09 Thread FUKAUMI Naoki
Hi, newvers.sh in 6.2/sys.tar.gz says it's "-current". it looks strange. $ sha256 sys.tar.gz SHA256 (sys.tar.gz) = 1fe2c451a6151164a4e97fc07b639413a9846e67406f54578a3ff8ceba9e802f $ tar xzf sys.tar.gz sys/conf/newvers.sh $ grep -C5 vers.c sys/conf/newvers.sh #

Re: OpenBSD 6.2 released: Oct 9, 2017

2017-10-09 Thread Callum R. Davies
Huge congrats to all OpenBSD developers for another rock-solid release, and with so many wonderful improvements to boot! Just a couple of small nits: On Mon, Oct 09, 2017 at 08:44:31AM -0600, Theo de Raadt wrote: > - SYSTEM SOURCE CODE --- > > The

Re: softdep dangling vnode

2017-10-09 Thread Mike Belopuhov
On Mon, Oct 09, 2017 at 22:21 +, Alexander Bluhm wrote: > Hi, > > we sometimes see a panic "unmount: dangling vnode" when rebooting a 6.1 > system with softdep. > > I have hacked some diagnostic panics until I got these traces from the > reboot and update process. > > Reboot: >

[patch] pkg_sign.1 correction

2017-10-09 Thread Bryan Vyhmeister
I was reading pkg_sign.1 and noticed that the description for the -o option seems incorrect. The current pkg_sign.1 text says under the -o option that "Otherwise, unsigned packages are created in the current directory" where I think it should say "Otherwise, signed packages are created in the

softdep dangling vnode

2017-10-09 Thread Alexander Bluhm
Hi, we sometimes see a panic "unmount: dangling vnode" when rebooting a 6.1 system with softdep. I have hacked some diagnostic panics until I got these traces from the reboot and update process. Reboot: sleep_finish() at sleep_finish+0xb1 tsleep() at tsleep+0x154 biowait() at biowait+0x46

Re: pow() returns a negative result on loongson

2017-10-09 Thread Paul Irofti
On Mon, Oct 09, 2017 at 10:39:47PM +0200, Juan Francisco Cantero Hurtado wrote: > Marc Feeley (Gambit Scheme) has been helping me with a bug on Gambit on > Loongson. Apparently the bug is on our side. > > I've created this little test based on his code: > > #include > #include > > int main()

Re: Add SIMPLEQ_REMOVE to sys/queue.h

2017-10-09 Thread Paul Irofti
On Mon, Oct 09, 2017 at 11:21:24PM +0200, Alexander Bluhm wrote: > On Tue, Oct 10, 2017 at 12:07:31AM +0300, Paul Irofti wrote: > > I needed this in a program and searching the kernel I see that this is > > also hacked in a few places (e.g. mpii(4)). > > Why don't you use a double linked list? I

Re: Add SIMPLEQ_REMOVE to sys/queue.h

2017-10-09 Thread Alexander Bluhm
On Tue, Oct 10, 2017 at 12:07:31AM +0300, Paul Irofti wrote: > I needed this in a program and searching the kernel I see that this is > also hacked in a few places (e.g. mpii(4)). Why don't you use a double linked list? I consider that removing with a loop is a hack instead of using as suitable

Add SIMPLEQ_REMOVE to sys/queue.h

2017-10-09 Thread Paul Irofti
I needed this in a program and searching the kernel I see that this is also hacked in a few places (e.g. mpii(4)). Here is a small regression test --- #include #include #include SIMPLEQ_HEAD(listhead, entry) head =

pow() returns a negative result on loongson

2017-10-09 Thread Juan Francisco Cantero Hurtado
Marc Feeley (Gambit Scheme) has been helping me with a bug on Gambit on Loongson. Apparently the bug is on our side. I've created this little test based on his code: #include #include int main() { double x = 0.5; double y = 1074.0; printf("x=%.20g y=%.20g

rw locks vs memory barriers and adaptive spinning

2017-10-09 Thread Mateusz Guzik
Hello, I was looking at rw lock code out of curiosity and noticed you always do membar_enter which on MP-enabled amd64 kernel translates to mfence. This makes the entire business a little bit slower. Interestingly you already have relevant macros for amd64: #define membar_enter_after_atomic()

Re: Installer support to fetch/verify bsd.rd for release upgrade

2017-10-09 Thread Robert Peichaer
On Sun, Oct 08, 2017 at 09:56:15AM +, Robert Peichaer wrote: > Up to now, the upgrade procedure from one to the next release meant > that you had to manually download and verify the new ramdisk kernel. > > What about if you just needed to boot into the existing bsd.rd and > it would support

Re: document tar -vv

2017-10-09 Thread Ingo Schwarze
Hi Todd, Todd C. Miller wrote on Mon, Oct 09, 2017 at 09:35:44AM -0600: > On Mon, 09 Oct 2017 17:27:46 +0200, Ingo Schwarze wrote: >> Given that tar(1) is not even specified by POSIX, i don't see >> how documenting it might do harm. The option was introduced >> here: >> >> date: 1997/01/24

Re: document tar -vv

2017-10-09 Thread Todd C. Miller
On Mon, 09 Oct 2017 17:27:46 +0200, Ingo Schwarze wrote: > Given that tar(1) is not even specified by POSIX, i don't see > how documenting it might do harm. The option was introduced > here: > > date: 1997/01/24 19:41:23; author: millert; > Support multiple -v options like GNU tar (> 1 -v

document tar -vv

2017-10-09 Thread Ingo Schwarze
Hi, Philippe Meunier wrote on Mon, Oct 09, 2017 at 08:03:12AM -0400: > leo_...@volny.cz wrote: >> % tar cvvf - > On a related note, it would be nice if tar(1)'s man page indicated > that the -v option can be specified more than once to get extra > information. Until seeing this discussion

Re: pms: tests with older Elantech touchpads needed

2017-10-09 Thread Stefan Sperling
On Mon, Oct 09, 2017 at 12:41:18AM +0200, Ulf Brosziewski wrote: > This patch adapts the Elantech handlers in pms to the new touchpad > infrastructure of wsmouse. The changes concern models that use the older > protocol versions 1, 2, and 3. You don't find that hardware at the next > corner

OpenBSD 6.2 released: Oct 9, 2017

2017-10-09 Thread Theo de Raadt
- OpenBSD 6.2 RELEASED - October 9, 2017. We are pleased to announce the official release of OpenBSD 6.2. This is our 43rd release. We remain proud of OpenBSD's record of

ifioctl() cleanup

2017-10-09 Thread Martin Pieuchot
Here's a small cleanup to make it easy to push the NET_LOCK() further down. Diff below factorize all ioctl requests need root privileges in one switch () block. It moves SIOCIFAFATTACH/SIOCIFAFDETACH where it belongs, to the block that keeps track of the timestamps of the last change. It gets

Re: pthread_once: don't deadlock on cancel from init_routine

2017-10-09 Thread Scott Cheloha
Hi, > On Oct 1, 2017, at 9:30 PM, Scott Cheloha wrote: > > Per this bit from pthread_once(3): > >> The function pthread_once() is not a cancellation point. >> However, if init_routine() is a cancellation point and is >> cancelled, the effect on once_control is as if

Re: kq_count earlier!

2017-10-09 Thread Martin Pieuchot
On 09/10/17(Mon) 14:35, Joerg Sonnenberger wrote: > On Mon, Oct 09, 2017 at 12:27:50PM +0200, Martin Pieuchot wrote: > > On 09/10/17(Mon) 12:22, Joerg Sonnenberger wrote: > > > On Mon, Oct 09, 2017 at 12:15:46PM +0200, Martin Pieuchot wrote: > > > > Diff below is a small cleanup to keep the

Re: kq_count earlier!

2017-10-09 Thread Joerg Sonnenberger
On Mon, Oct 09, 2017 at 12:27:50PM +0200, Martin Pieuchot wrote: > On 09/10/17(Mon) 12:22, Joerg Sonnenberger wrote: > > On Mon, Oct 09, 2017 at 12:15:46PM +0200, Martin Pieuchot wrote: > > > Diff below is a small cleanup to keep the accounting of events in > > > sync with the number of events on

Re: Move 'struct kevent' storage to the stack

2017-10-09 Thread Mark Kettenis
> Date: Mon, 9 Oct 2017 12:06:53 +0200 > From: Martin Pieuchot > > On 09/10/17(Mon) 11:55, Mark Kettenis wrote: > > > Date: Mon, 9 Oct 2017 11:48:56 +0200 > > > From: Martin Pieuchot > > > > > > This is the first step to allow kqueue_scan() to be executed in

Re: kq_count earlier!

2017-10-09 Thread Martin Pieuchot
On 09/10/17(Mon) 12:22, Joerg Sonnenberger wrote: > On Mon, Oct 09, 2017 at 12:15:46PM +0200, Martin Pieuchot wrote: > > Diff below is a small cleanup to keep the accounting of events in > > sync with the number of events on the list. This is a noop for the > > moment, but it's small & easy part

Re: kq_count earlier!

2017-10-09 Thread Joerg Sonnenberger
On Mon, Oct 09, 2017 at 12:15:46PM +0200, Martin Pieuchot wrote: > Diff below is a small cleanup to keep the accounting of events in > sync with the number of events on the list. This is a noop for the > moment, but it's small & easy part to review of my upcoming diff. Well, not counting the

kq_count earlier!

2017-10-09 Thread Martin Pieuchot
To prevent an infinite loop, threads looking for events inside kqueue_scan(), insert a `marker' in the list. Such markers are not accounted and they are removed from the list as soon as the thread is finished or goes to sleep. Diff below is a small cleanup to keep the accounting of events in

Re: Move 'struct kevent' storage to the stack

2017-10-09 Thread Martin Pieuchot
On 09/10/17(Mon) 11:55, Mark Kettenis wrote: > > Date: Mon, 9 Oct 2017 11:48:56 +0200 > > From: Martin Pieuchot > > > > This is the first step to allow kqueue_scan() to be executed in > > parallel. I'm currently focusing on data structure ownership with > > regard to sleep. I

Re: Move 'struct kevent' storage to the stack

2017-10-09 Thread Mark Kettenis
> Date: Mon, 9 Oct 2017 11:48:56 +0200 > From: Martin Pieuchot > > This is the first step to allow kqueue_scan() to be executed in > parallel. I'm currently focusing on data structure ownership with > regard to sleep. I don't plan to get rid of the KERNEL_LOCK() for > the

Move 'struct kevent' storage to the stack

2017-10-09 Thread Martin Pieuchot
This is the first step to allow kqueue_scan() to be executed in parallel. I'm currently focusing on data structure ownership with regard to sleep. I don't plan to get rid of the KERNEL_LOCK() for the moment. I don't think it makes sense unless we can unlock the kevent(2) syscall. The diff

Re: rtsx(4): add ADMA support

2017-10-09 Thread Stefan Sperling
On Sun, Oct 08, 2017 at 09:49:41PM +0200, Stefan Sperling wrote: > The sdmmc stack provides DMA buffers for data transfers nowadays. > This was not the case yet when rtsx(4) was written so this driver > is still using a lame memcpy bounce-buffer approach. > > With this diff regular data transfers

Re: armv7 bcopyinout.S vs bcopyinout_xscale.S

2017-10-09 Thread Mark Kettenis
> Date: Mon, 9 Oct 2017 08:45:55 +0300 > From: Artturi Alm > > Hi, > > > has anyone looked at the netbsd xscale-versions of bcopyin/bcopyout/kcopy? > > this is from netbsd bcopyinout.S: > #if defined(__XSCALE__) || defined(_ARM_ARCH_6) > /* > * armv6 and v7 have pld

Re: bc(1) examples

2017-10-09 Thread Otto Moerbeek
On Sun, Oct 08, 2017 at 11:10:06PM +0200, Jan Stary wrote: > On Oct 08 11:31:16, o...@drijf.net wrote: > > On Fri, Oct 06, 2017 at 02:12:01PM +0200, Jan Stary wrote: > > > > > Isn't "4 * a(1)" a more natural incarnation of pi than "2 * a(2^1)"? > > > > The point of this example is to (also)