Re: Enabling DBUS in VmWare Fusion vm shows VMWare + DragonFly BSD is not so great.

[karu.pruun]

On Mon, Jan 6, 2014 at 1:20 AM, Justin Sherrill wrote:

 On Sun, Jan 5, 2014 at 1:54 AM, karu.pruun wrote:

 By the way, fonts look much better when you use infinality patches to
 freetype2

 http://www.infinality.net/blog/

 Unfortunately the latest release is for freetype2-2.4.12, so I ended up
 building a package using dports src and force-installing it back on the
 system every time pkg upgrades it to the new version 2.5, which happens
 every time when you install a package.


 I have not tried this, but there is a 'pkg lock' option that could keep
 you from having to do this forced downgrade every time.



Thanks for this---it worked!

Cheers,

Peeter

--

Re: DPorts: GLib bug crashes applications linked with it

[karu.pruun]

I have a brand new 3.6 install as well and a dports based xfce4 and its
apps are very stable. I remember though that in an older pkgsrc based
install (maybe 3.2?) crashed thunar and many other xfce apps, so it was
mostly unusable.

Peeter

--


On Thu, Jan 9, 2014 at 2:49 AM  wrote:

  Hi,

 My former copy was based in /usr/pkg, while the new one is based in
 /usr/local. /usr/pkg have been renamed before an eventual deletion, to
 avoid any confusion.

 I wonder if it could be related to any locale manner. I compiled the
 packages myself, so if I'm the only one with this problem, it's probably
 related with my environment.

 What version of DragonFly are you using?

 Thanks for the reply.

 SR



 Le 2014-01-08 12:28, Warren Postma a écrit :

   Xfce for its part is hardly usable because the panels and Thunar are
 crashing at startup.

 I have no problems launching XFCE4 panel (the dock area) or file manager
 (thunar) on my brand new install. I suspect you may have some old shared
 libraries or other broken dependencies? Have you tried uninstalling
 all XFCE4 dependencies that were installed using pkgsrc and reinstalling
 them all freshely with dports?

 Warren



 On Tue, Jan 7, 2014 at 10:33 PM, sruss...@prodigeinfo.qc.ca wrote:

 Hi,

 I've just upgraded my DF station from 3.4 to 3.6. In the process, I
 decided to try dport. I compiled it myself, as usual. I need LDAP+Kerberos
 enabled, plus few other things.

 Overall, I'm very glad of it. DF 3.6 is running like a charm and I can
 now access some applications that were unavailable in pkgsrc.

 I had minor problems with the dports packages. For example, I had to
 compile and install the pkgsrc version of pam_krb5 because the version
 provided in dports does not work properly. VLC2 is also not working for me
 and was replaced by umplayer. Stuff like that.

 I have only one significant problem that affects mostly all applications
 using Glib, which includes wxGTK apps, including Seamonkey and LibreOffice.

 Every time that I try to open or save a file, a the popup window shows
 itself and works normally. But after a few seconds, the popup freezes and
 the application sends this message before crashing:

 «GLib (gthread-posix.c): Unexpected error from C library during
 'pthread_mutex_lock': Resource deadlock avoided.  Aborting.»

 Xfce for its part is hardly usable because the panels and Thunar are
 crashing at startup.

 It looks like the problem affects any process that tries to read a
 directory or something related. Most of theses applications are totally
 functionnal otherwise.

 If anyone had this problem or have a clue of what's causing it, please
 reply to me.

 As for the rest, DragonFly really rocks! It got everything in hand to
 take the lead.

 Thanks,

 Stephane Russell

   --
  Stéphane Russell
 Informaticien
 Prodige Informatique, s.a.
  Tél. : 450 323-6224
 Courriel : mailto:sruss...@prodigeinfo.com
 Web : http://www.prodigeinfo.com

 Ce courrier est confidentiel. Toute utilisation ou rediffusion, intégrale
 ou partielle, par une personne autre que le (les) destinataire(s)
 désigné(s) est interdite. Si vous recevez ce courrier électronique par
 erreur, prière d'en aviser l'envoyeur.

hammer pfs-upgrade question

[karu.pruun]

Hello

I wonder if I'm making an elementary mistake. I wanted to migrate /home to
a new disk, so I hammer mirror-copied /pfs/home to a new disk as explained
in man 5 HAMMER:

hammer mirror-copy /olddisk/pfs/home /newdisk/pfs/home

This results in a read-only /newdisk/pfs/home. I tried to upgrade to master
to be able to mount read-write:

hammer pfs-upgrade /newdisk/pfs/home

But this fails:

pfs-upgrade of PFS#1 () failed: Directory not empty

Can the new pfs be upgraded at all or is it that once a slave, always a
slave and read-only?

Cheers,

Peeter

--

Re: bwi status

[karu.pruun]

On Tue, Apr 1, 2014 at 5:29 PM, Sascha Wildner wrote:

 On Tue, 01 Apr 2014 08:04:23 +0200, karu.pruun wrote:

  I tried to compile a kernel with bwi wireless driver support but it
 failed.
 Does anyone know what is the status of the bwi driver? Would it
 support the Broadcom
 bcm4322 chipset (the airport extreme on a macboopro)?


 Our bwi(4) driver was never adjusted for the most recent 802.11 stack
 upgrade. Someone would have to do this work first (or re-port FreeBSD's
 current one) before it could work.



Thanks for the update.

Peeter

--

unreadable pflog file: solved: patch

[karu.pruun]

Hello

Recently I complained that the /var/log/pflog file in DFly 3.6 is
unreadable by tcpdump. I might have found the cause; apologies: I'm not a
developer and not even versed in C so more knowledgeable people might be
able to comment and improve on this. Anyhow, I compared the code of pflogd
of OpenBSD, FreeBSD and DragonFly, and tracked the issue to how DragonFly
writes to the pflog logfile. The patch for DragonFly_RELEASE_3_6 is
included and seems to work so far: tcpdump is now able to output from
/var/log/pflog.

Cheers,

Peeter

--


0001-Bring-in-some-changes-from-FreeBSD.patch
Description: Binary data

Re: 3.8 release plans

[karu.pruun]

On Tue, May 27, 2014 at 3:38 PM, Sascha Wildner wrote:
 On Tue, 27 May 2014 05:19:03 +0200, karu.pruun wrote:

 Although this is not in the checklist, I'd like to point out a commit
 that has not made it into the tree yet

 http://bugs.dragonflybsd.org/issues/2671


 I'll look at it.


Just to clarify that an earlier commit

http://gitweb.dragonflybsd.org/dragonfly.git/commit/2461cb133c394b4a1aac7d2ce9989ffa14ee6141

remedied the problem that tcpdump did not recognize the pflogd log
file at all. The changes were from FreeBSD; the idea is that when
writing to logfile, the struct pcap_sf_pkthdr should be used and not
pcap_pkthdr. I did not notice that when scanning the logfile, that
same struct, pcap_sf_pkthdr, should be used. In short, both the
previous commit and this commit should have been made at once; both
are from FreeBSD's pflogd.

I did not track down why the earlier version with pcap_pkthdr works in
OpenBSD and not in DragonFly or FreeBSD.


 Something I noticed at 3.8.0RC: at boot, the kernel complains that it
 cannot find

 /boot/kernel/acpi
 /boot/kernel/ehci

 and they are not loaded.


 On which box is this? Is it an upgraded system. If not, which ISO/IMG was
 used? Any idea when it started?


I have not seen it with 3.6.x on Virtualbox. I did a fresh install
with dfly-x86_64-3.8.0RC.iso.bz2 on Virtualbox this morning, uname -a

DragonFly dflybox 3.8-RELEASE DragonFly v3.8.0rc-RELEASE #12: Sun May
25 23:15:08 PDT 2014
r...@pkgbox64.dragonflybsd.org:/usr/obj/build/home/justin/src/sys/X86_64_GENERIC
 x86_64

and noticed the message when booting the installed system from hdd
(the install cd loaded both modules fine). After the kernel loads from
hdd, it says:

--
Unable to load /boot/kernel/acpi
Unable to load /boot/kernel/ehci
ACPI autoload failed - inappropriate file type or format
EHCI autoload failed - inappropriate file type or format
--

and then continues fine. kldstat only shows kernel.

Peeter

--

Re: 3.8 release plans

[karu.pruun]

On Tue, May 27, 2014 at 6:51 PM, Sascha Wildner s...@online.de wrote:
 On Tue, 27 May 2014 10:00:15 +0200, karu.pruun karu.pr...@gmail.com wrote:

 On Tue, May 27, 2014 at 3:38 PM, Sascha Wildner wrote:

 On Tue, 27 May 2014 05:19:03 +0200, karu.pruun wrote:

 Something I noticed at 3.8.0RC: at boot, the kernel complains that it
 cannot find

 /boot/kernel/acpi
 /boot/kernel/ehci

 and they are not loaded.



 On which box is this? Is it an upgraded system. If not, which ISO/IMG was
 used? Any idea when it started?



 I have not seen it with 3.6.x on Virtualbox. I did a fresh install
 with dfly-x86_64-3.8.0RC.iso.bz2 on Virtualbox this morning, uname -a

 DragonFly dflybox 3.8-RELEASE DragonFly v3.8.0rc-RELEASE #12: Sun May
 25 23:15:08 PDT 2014

 r...@pkgbox64.dragonflybsd.org:/usr/obj/build/home/justin/src/sys/X86_64_GENERIC
  x86_64

 and noticed the message when booting the installed system from hdd
 (the install cd loaded both modules fine). After the kernel loads from
 hdd, it says:

 --
 Unable to load /boot/kernel/acpi
 Unable to load /boot/kernel/ehci
 ACPI autoload failed - inappropriate file type or format
 EHCI autoload failed - inappropriate file type or format


 It must be a new issue specific to VirtualBox. I just tried

 http://avalon.dragonflybsd.org/iso-images/dfly-x86_64-3.8.0RC.iso.bz2

 in Microsoft's Hyper-V and it _does_ load the acpi and ehci modules when
 booting from the ISO itself and also when booting a system which was
 installed from this ISO.

 I also now remember a guy in the #dragonflybsd IRC channel who had the same
 issue (also on VirtualBox) some weeks ago. I'll look a bit further tonight
 if I can find something.


Thanks for this.

Peeter

--

Re: 3.8 release plans

[karu.pruun]

On Tue, May 27, 2014 at 8:42 PM, Sascha Wildner s...@online.de wrote:
 On Tue, 27 May 2014 13:38:22 +0200, karu.pruun karu.pr...@gmail.com wrote:

 I also now remember a guy in the #dragonflybsd IRC channel who had the
 same
 issue (also on VirtualBox) some weeks ago. I'll look a bit further
 tonight
 if I can find something.



 Thanks for this.


 I can't check here at work, but do you remember (or can check it) if the
 loading acpi/ehci loading issues were only occurring when booting from the
 installed system or if it was also failing when booting the ISO? Note that
 the ISO boot should load dm.ko too.


Indeed, it is weird that when booting the ISO, acpi/ehci and dm load
fine, but when booting the installed system acpi/ehci do not load and
show the errors. And dm does not load either but this happens
silently; there are no error messages or anything.

I forgot to add that this is Virtualbox 4.3.10 r92013 on os x 10.9.

Peeter

--

Re: 3.8 release plans

[karu.pruun]

On Tue, May 27, 2014 at 9:45 PM, karu.pruun karu.pr...@gmail.com wrote:
 On Tue, May 27, 2014 at 8:42 PM, Sascha Wildner s...@online.de wrote:
 On Tue, 27 May 2014 13:38:22 +0200, karu.pruun karu.pr...@gmail.com wrote:

 I also now remember a guy in the #dragonflybsd IRC channel who had the
 same
 issue (also on VirtualBox) some weeks ago. I'll look a bit further
 tonight
 if I can find something.



 Thanks for this.


 I can't check here at work, but do you remember (or can check it) if the
 loading acpi/ehci loading issues were only occurring when booting from the
 installed system or if it was also failing when booting the ISO? Note that
 the ISO boot should load dm.ko too.


 Indeed, it is weird that when booting the ISO, acpi/ehci and dm load
 fine, but when booting the installed system acpi/ehci do not load and
 show the errors. And dm does not load either but this happens
 silently; there are no error messages or anything.

 I forgot to add that this is Virtualbox 4.3.10 r92013 on os x 10.9.


This seems to be a Virtualbox problem: I just installed DragonFly
3.6.2 anew on Virtualbox 4.3.10 and ran into the same problem. The
other DFly 3.6.2 image that I am using and that runs fine was
installed using an older version of Virtualbox (perhaps 4.2.x); I
should revert back to that one.

Cheers,

Peeter

--

Re: 3.8 release plans

[karu.pruun]

On Wed, May 28, 2014 at 4:51 PM, Zenny wrote:
 Confirmation that iso works fine in VirtualBox 4.3.8 OSE with ACPI
 disabled. But the img file didn't.

 On 5/28/14, Justin Sherrill jus...@shiningsilence.com wrote:
 Hey this boots strangely oh wait it is just this particular version of
 VirtualBox that has the problem has been a reported problem at what feels
 like every release for years.  I'm starting to think that if anyone reports
 issues in VirtualBox, the first step should always be try a different
 version of VirtualBox, or a different virtual system.


Point taken, suppose it's a good idea to test on different machines.

However, now it looks again like a DFly problem. I tested on
Virtualbox 4.1, 4.2, 4.3 and finally reproduced it on a physical
machine too. It only concerns installations where all partitions are
UFS. It does not show for a hammer system (ie root HAMMER and /boot
UFS).

I used the partition layout:

---
/  UFS
swap
/usr  UFS
---

when booting the system from hdd, the acpi/ehci error shows for DFly
3.6.x and 3.8.0RC where both are x86_64. So it appeared already in
3.6.x.

Peeter

--

Re: 3.8.1 set for release on Monday 16-June-2014

[karu.pruun]

Thanks for the update!

I am wondering if you could briefly comment on boot0cfg. This might be
(or might not be?) related to a recent strange error I encountered
while installing DFly on a macbookpro: I wiped clean the harddisk and
installed DFly + FreeBSD using an MBR partitioning scheme. Each has an
MBR partition and since DFly was installed last I hit enter when it
asked if I wanted to install bootblocks. The result is strange: DFly's
boot manager can boot FreeBSD but not DFly (after the boot prompt, the
cursor moves to the next line and that's all, the machine hangs). I
can still boot DFly using the install CD (get to boot prompt and tell
the correct rootdev and vfs.root.mountfrom etc), but it's clumsy.

So I wonder if it's boot0cfg or something related to boot1_64 and
boot2_64 that fails the booting process? I suppose the fact that
macbookpro has EFI does not help either. I am not sure how to track
this further.

Cheers,

Peeter

--


On Fri, Jun 13, 2014 at 2:35 PM, Matthew Dillon
dil...@apollo.backplane.com wrote:
 We will be releasing 3.8.1 on Monday.  It will contain:

 * Temporary csprng fixes while we work on adding a second algorithm
 * USB mouse stall fix for X
 * MXCSR fixes for X and firefox (hopefully they are now fixed).
 * Updated openssl (again)

 These changes are in the -release branch already so anyone who wants
 a jump-start can just compile up a new world/kernel and be done with it.

 The delay is so we can review a few other things in master that we might
 want to cherry-pick and to check for and clean up remaining loose ends.

 I will also use the time to look at boot0cfg and nmalloc cache coloring
 but adjustments to those might not get into 3.8.1.  boot0cfg... we might
 just have to say that it shouldn't be used.  It's a pretty bad hack.
 nmalloc... we might not have time to test the requested performance
 enhancement for the 3.8.1 roll.

 -Matt
 Matthew Dillon
 dil...@backplane.com

Re: Supermicro MBD-A1SAi-2550F

[karu.pruun]

Correct me if Im wrong but I understand you don't even see kernel
booting? If true, then seems not related to the new usb stack since
this kicks in much later in the booting process. The hang is
immediately after the boot prompt; I've seen this since 4 yrs ago on
my laptop (macbookpro) and believe it could be due to (U)EFI that I
see your machine also has. Also, might be more related to boot (ie
boot1 + boot2) rather than boot0 since I have a situation where DFly
and FBSD both live in the MBR partitions and DFly's boot0 is able to
boot FBSD, but not itself! Exactly the same phenomenon: I get the boot
prompt (where actually the partitions are F1 = DFly, F3 = FBSD)

F1  DF/FBSD
F3  DF/FBSD

and then hitting F3 boots FBSD, but F1 hangs. I can still boot DFly
with an ISO image (get the prompt, set currdev etc and boot), but it's
cumbersome.

I don't know what the fix is; would like to hear if anyone has a solution.

In the future (when I'll have some more time), I'd be interested in
working towards integrating UEFI booting into DFly.

Btw, when you'll get your machine to boot, can you tell if it works
with the new usb?

Peeter

--



 On Sat, Jun 21, 2014 at 2:23 PM, Roman Neuhauser neuhau...@sigpipe.cz wrote:
 having just bought a Supermicro MBD-A1SAi-2550F motherboard[1]
 i proceeded and failed to boot 3.8.1 and 3.8.0 from several
 different USB sticks.  FreeBSD-10.0-RELEASE-amd64-memstick.img
 boots normally.

 the system hangs on exit from the F1/F5 menu (boot0 i believe?).

 PXE should work but i still prefer having a means of booting
 from a USB stick.  what can i do to help fix this?

 http://www.supermicro.com/products/motherboard/atom/x10/a1sai-2550f.cfm

 --
 roman

Re: 3.8.1 freeze at boot

[karu.pruun]

Found the location in ohci.c where the system hangs

http://bugs.dragonflybsd.org/issues/2686

Peeter

--




On Wed, Jun 18, 2014 at 7:35 PM, karu.pruun karu.pr...@gmail.com wrote:
 Hello

 While DFly 3.6.x x86_64 boots and runs fine on my laptop
 (macbookpro5,5), the 3.8.1 install iso freezes at boot. (acpi/ehci
 disabled in both cases.) These are the last lines before freezing
 [boot_verbose=YES]:

 --- ---
 pci0: processor (vendor 0x10de, dev 0x0aa3) at device 3.5 irq 14
 ohci0.pci0.pcib0.legacy0.nexus0.root0
 ohci0: nVidia nForce MCP79 USB Controller [tentative] mem
 0xd3488000-0xd3488fff irq 11 at device 4.0 on pci0
 ohci0: Reserved 0x1000 bytes for rid 0x10 type 3 at 0xd3488000
 ---FREEZE---

 Since 3.6.x boots and runs fine (with acpi disabled) on this machine,
 my hypothesis was that I could have 3.8.1 with oldusb. So upgraded
 3.6.2 to 3.8.1 via git and buildworld/buildkernel with WANT_OLDUSB=yes
 in /etc/make.conf. The system now panics at boot, showing
 [boot_verbose=YES]:

 ---3.8.1-oldusb---
 fwohci0: 1394 Open Host Controller Interface [tentative] mem
 0xd310-0xd3100fff irq 7 at device 0.0 on pci4
 fwohci0: latency timer 0 - 32.
 fwohci0: cache size 64 - 64.
 fwohci0: Reserved 0x1000 bytes for rid 0x10 type 3 at 0xd310
 ---panic---

 and then giving the db prompt but because it freezes, or at least the
 keyboard freezes, I can't call dumpsys. So unfortunately there's no
 crash dump. I could then boot and run the system with the old 3.6.2
 kernel.

 I wonder if I have a wrong kernel configuration for oldusb? The
 relevant section reads:

 ---X86_64_GENERIC-OLDUSB---
 # USB support
 #
 # NOTE: If you enable 'oldusb' you must also disable 'usb' and rebuild
 # the world with WANT_OLDUSB=true in /etc/make.conf, in addition
 # to rebuilding the kernel.
 #
 device oldusb # use this instead of usb for the old stack
 #device natausb # ATA-over-USB support (oldusb)
 #device ugen # Generic (oldusb)
 #device uscanner # Scanners (oldusb)
 #device urio # Diamond Rio MP3 Player (oldusb)
 #device aue # ADMtek USB ethernet (oldusb)
 #device cue # CATC USB ethernet (oldusb)
 #device kue # Kawasaki LSI USB ethernet (oldusb)
 #device rue # RealTek 8150 based USB ethernet (oldusb)

 ##device usb # USB Bus (required)
 device uhci # UHCI PCI-USB interface
 device ohci # OHCI PCI-USB interface
 device uhid # Human Interface Devices
 device ukbd # Keyboard
 device ulpt # Printer
 device umass # Disks/Mass storage - Requires scbus and da
 device ums # Mouse
 # USB Ethernet, requires mii
 device axe # ASIX Electronics USB Ethernet
 # USB wireless NICs, requires wlan, wlan_amrr
 #device rum # Ralink Technology RT2501USB/RT2601USB
 #device ural # Ralink Technology RT2500USB

 # FireWire support
 device firewire # FireWire bus code
 device sbp # SCSI over FireWire (Requires scbus and da)
 device fwe # Ethernet over FireWire (non-standard!)
 ---

 I also tried booting FreeBSD 10.0 to see how the (new) usb stack
 fares. It boots fine. I attach the verbose dmesg for both FBSD 10 and
 DFly 3.6.2 just in case.

 I wonder if there's a way to get a debug or dumpsys for 3.8? Could
 vkernel help or is it isolated from hardware?

 Thanks

 Peeter

 --


 Attached:

  - FBSD 10.0 verbose dmesg [acpi enabled]
  - DFly 3.6.2 verbose dmesg [no acpi/ehci]

Re: Supermicro MBD-A1SAi-2550F

[karu.pruun]

On Mon, Jun 23, 2014 at 12:36 AM, Roman Neuhauser neuhau...@sigpipe.cz wrote:
 The hang is immediately after the boot prompt; I've seen this since 4
 yrs ago on my laptop (macbookpro) and believe it could be due to
 (U)EFI that I see your machine also has.

 it's all set to legacy mode, i don't think this is the culprit.
 (i tried with UEFI, with indistinguishable result.


Sure agreed. What I meant was that if you have a UEFI machine then
even when it emulates BIOS style booting, then curiously DFly does not
boot whereas FBSD does.

Peeter

--

Re: Supermicro MBD-A1SAi-2550F

[karu.pruun]

 prompt lsdev
 prompt set vfs.root.mountfrom=ufs:da0s1a
 prompt boot

 OK lsdev
 cd devices:
 disk devices:
 disk0:   BIOS drive C:
 *   disk0s1a: FFS


This is the root device.


 disk1:   BIOS drive D:
   disk1s1: Unknown fs: 0xee
 disk2:   BIOS drive E:
   disk2s1: Unknown fs: 0xee
 pxe devices:
 OK set vfs.root.mountfrom=ufs:da0s1a
 OK boot
 kernel text=0x972528 data=...

 ...

 at this point, disk1, disk2 are da0, da1; disk0 is (ISTR) da8.


Since disk0 has the rootdev, try da8s1a,

 set vfs.root.mountfrom=ufs:da8s1a
 boot


Peeter

--

Re: [ANN] Rust ported to DragonFlyBSD

[karu.pruun]

Thanks, great work!

Peeter

--



On Tue, Jul 29, 2014 at 9:59 PM, Michael Neumann mneum...@ntecs.de wrote:
 Hi all,

 I am happy to announce that I succeeded in porting Rust to the DragonFly
 operating system. [This article][1] describes `how` and might also be of
 interest to others porting Rust to NetBSD or OpenBSD. Within the next
 week I submit patches to the LLVM project (segmented stack support for
 DragonFly) and also to the rust repo itself. My `dragonfly` branch of
 the Rust repository can be found [here][2].

 If someone is interested to try out early, I can provide some binaries
 until I set up some automatic builds.

 Regards,

   Michael

 [1]: http://www.ntecs.de/blog/2014/07/29/rust-ported-to-dragonflybsd/
 [2]: https://github.com/mneumann/rust/tree/dragonfly

Re: Status of nginx

[karu.pruun]

How do you enable the so_reuseport options, is it the sysctl ones?

net.inet.tcp.reuseport_ext: 1
net.inet.udp.reuseport_ext: 1


Peeter

--




On Tue, Oct 21, 2014 at 12:26 PM, Sepherosa Ziehau sepher...@gmail.com wrote:
 I'd also suggest to enable so_reuseport global option, which makes
 your server much faster :)

 On Tue, Oct 21, 2014 at 7:18 AM, Benjamin Woolley tauto...@gmail.com wrote:
 Hi Konrad,

 If you don't want to try master, know that I am using that combination now, 
 even with a dragonfly release that has the bug. I disabled sendfile support, 
 and it is working great. The combination runs great even without sendfile 
 support, as long as your requirements can live without it. I use it on an 
 image heavy site which hits sendfile often, but it is low traffic, so the 
 system is usually idle, and the difference is not noticeable.

 Ben

 On Oct 20, 2014, at 11:48 AM, Joris Giovannangeli jo...@giovannangeli.fr 
 wrote:

 Hi
 On 20/10/2014 16:04, Konrad Neuwirth wrote:
 Dear reader,

 what is the current status of nginx/php_fpm on DragonflyBSD? I know that 
 there were some discussion about that combination a while ago -- and we 
 also had been bitten by the bug that scrambled uploaded files. I also 
 remember seing patches; but did they solve the problem? Is nginx/php_fpm 
 now a well-running combination?

 This particular issue with sendfile has been fixed on master. I'm not
 aware of any other issues at the moment.

 Regards,
 joris





 --
 Tomorrow Will Never Die

Re: polkit does not start

[karu.pruun]

I am using 64 bit machines in both cases. It's a puzzling situation
since the relevant ports all seem to be identical on both machines.

Tested with your com.mycompany.Test1 and got the same result.

Thanks for mentioning the dbus webpage, I'll try and see if I can get
more info how the dbus EXTERNAL authentication mechanism is supposed
to work: what the strings 30 (FreeBSD) vs 34323934393637323935
(DragonFly) stand for.

Cheers,

Peeter

--



On Mon, Jan 12, 2015 at 12:33 AM, Stephane Russell
sruss...@prodigeinfo.qc.ca wrote:
 Hi again,

 Well you're maybe right. But for what I could see, the problem don't
 seem to originate from polkit alone.

 I'm not expert of dbus, but for what I know, your logs indicates that
 the EXTERNAL authentication mecanism is not supported
 (http://dbus.freedesktop.org/doc/dbus-specification.html#auth-command-auth).
 You'll see the reason below. Otherwise, you would have another kind of
 response.

 For hald and consolekit. Maybe hald is not using the same auth mecanism
 then polkit. ConsoleKit is probably working fine for users, but the
 system service is probably not, since its dependent on polkit. For what
 I can see its failing as well.

 I compiled glib, dbus, polkit and ConsoleKit with debug symbols on, by
 using copies generated by dports and activating it. When testing, I'm
 also using dbus with logs activated. I compiled it with an extra debug
 flag and then can start it manually by adding an environment parameter
 (DBUS_VERBOSE=1) as indicated in the man page of dbus-daemon.

 I made some traces and found that polkitd can't reserve its service
 name. I made a pseudo service named com.mycompany.Test1 and tried to
 reserve a name with the test tool provided with gio (in glib),
 gbbus-exemple-own-name:

 /usr/local/etc/dbus-1/system.d/com.mycompany.Test1.conf:

 ?xml version=1.0 encoding=UTF-8? !-- -*- XML -*- --

 !DOCTYPE busconfig PUBLIC
  -//freedesktop//DTD D-BUS Bus Configuration 1.0//EN
  http://www.freedesktop.org/standards/dbus/1.0/busconfig.dtd;
 busconfig

   !-- This configuration file specifies the required security policies
for the ColorManager to work. --

   !-- Only user root or user colord can own the colord service --
   policy user=root
 allow own=com.mycompany.Test1/
   /policy
   policy user=myuser
 allow own=com.company.Test1/
   /policy

  !-- Allow anyone to call into the service - we'll reject callers using
 PolicyKit --
   policy context=default
 allow send_destination=com.company.Test1/
 allow receive_sender=com.company.Test1/
   /policy

 /busconfig

 I then modified and runned the test tool with myuser and root users as
 well:

 gdbus-example-own-name -r -a -n com.mycompany.Test1
 Lost the name com.company.Test1 on the session bus

 I got the same results for both cases. I concluded that the fail was
 coming from there, but I can't figure why.

 I noted in the logs you provided that the length of exchanged data are
 twice as long in DragonFly then in FreeBSD. Are you using a 32 bits
 machine for FreeBSD and 64bits for DragonFly? I wonder if it could be
 related. Maybe the glib, dbus or polkit port for 64 bit is broken. Or
 maybe SASL itself, but I think its less likelly.

 Cheers,

 SR


 karu.pruun a écrit :
 That's interesting---I wonder if you could tell me what research you
 did on dbus and polkit. I would be interested in trying to make it
 work. Here's what I've found so far.

 I figured out that consolekit and hald are fine, it's polkit that
 messes up. Dbus cannot spawn it, and the problem is not dbus, it's
 polkit. Since I know almost nothing about dbus and friends my approach
 has been to compare with FreeBSD, where it's working fine, and try to
 isolate the piece of code where polkit on DragonFly deviates from the
 behavior seen on FreeBSD. I ran ktrace on polkit and the deviation
 from FreeBSD happens on the second attempt (type: AUTH EXTERNAL) of
 authentication:

 ---FreeBSD (successful)---
   1579 polkitd  CALL  sendto(0x5,0x803076280,0x12,0x2MSG_NOSIGNAL,0,0)
   1579 polkitd  GIO   fd 5 wrote 18 bytes
AUTH EXTERNAL 30\r

   1579 polkitd  RET   sendto 18/0x12
   . . .

   1579 polkitd  CALL  recvfrom(0x5,0x80310f000,0x1000,0,0,0)
   1579 polkitd  GIO   fd 5 read 37 bytes
OK 42dd72d3f9cfe49615fb1f2154aabce1\r

   1579 polkitd  RET   recvfrom 37/0x25
 --- ---

 vs

 ---DragonFly---
 16116:1 0 polkitd  0.06 CALL
 sendto(0x5,0x800740280,0x24,MSG_NOSIGNAL,0,0)
 16116:1 0 polkitd  0.10 GIO   fd 5 wrote 36 bytes
AUTH EXTERNAL 34323934393637323935\r

 16116:1 0 polkitd  0.06 RET   sendto 36/0x24
 . . .

 16116:1 0 polkitd  0.000520 CALL  recvfrom(0x5,0x800735000,0x1000,0,0,0)
 16116:1 0 polkitd  0.13 GIO   fd 5 read 46 bytes
REJECTED EXTERNAL DBUS_COOKIE_SHA1 ANONYMOUS\r

 16116:1 0 polkitd  0.05 RET   recvfrom 46/0x2e
 --- ---

 The challenge is to find out why DragonFly computes

Re: polkit does not start

[karu.pruun]

Thanks for the tip. Can you tell me if there is a similar (elegant)
way to get GDB symbols in dport ports as well? In come cases there's
an OPTION which allows for gdb symbols, but in other cases there is
nothing. I ended up editing Makefiles by hand, which is probably not
the proper way.

Cheers,

Peeter

--



On Mon, Jan 12, 2015 at 7:55 AM, Sascha Wildner s...@online.de wrote:
 On Fri, 09 Jan 2015 15:14:40 +0100, karu.pruun karu.pr...@gmail.com wrote:

 I compiled polkit and glib with gdb symbols but it seems I might need
 more symbols in perhaps other libraries? In any case, so far the call
 to sendto() seems to occur in a piece of code that I can't step in.


 Likely, libc, at least. Our world is not compiled with debug symbols by
 default. You can easily do so by putting DEBUG_FLAGS=-g in your
 /etc/make.conf and rebuild/install world. It will increase disk space used
 and I guess build time too. I do it on all my boxes.

 Regards,
 Sascha

fixed: Re: polkit does not start

[karu.pruun]

On Fri, Jan 23, 2015 at 3:45 PM, Pierre Abbat
p...@leaf.dragonflybsd.org wrote:
 On Tuesday, January 20, 2015 20:21:32 karu.pruun wrote:
 I am using 64 bit machines in both cases. It's a puzzling situation
 since the relevant ports all seem to be identical on both machines.

 Tested with your com.mycompany.Test1 and got the same result.

 Thanks for mentioning the dbus webpage, I'll try and see if I can get
 more info how the dbus EXTERNAL authentication mechanism is supposed
 to work: what the strings 30 (FreeBSD) vs 34323934393637323935
 (DragonFly) stand for.

 34323934393637323935 is obviously 4294967295 (2^32-1) in ASCII hex. What it
 means in the protocol, I don't know.


Turns out it was supposed to be a UID. In the end, the error was with
credentials passing in devel/glib20 that polkitd uses. The attached
patch fixes it:

- copy the patch to /usr/dports/devel/glib20/dragonfly/
- make install

All the problems listed above go away:

-upowerd crashes
-ConsoleKit can't make a session active
-X login is freezing a minute or two
-The session menu of xfce is freezing also for some time
-any logoff from xfve is freezing temporarilly and leaves some
applications opened.

I haven't tested xdm etc though.

I'll try to get it to dports on github.

Peeter

--
--- gio/gcredentialsprivate.h.orig	2015-01-25 13:11:55.0 +0900
+++ gio/gcredentialsprivate.h	2015-01-25 13:14:07.0 +0900
@@ -39,6 +39,14 @@
 #define G_CREDENTIALS_UNIX_CREDENTIALS_MESSAGE_SUPPORTED 1
 #define G_CREDENTIALS_SPOOFING_SUPPORTED 1
 
+#elif defined(__DragonFly__)
+#define G_CREDENTIALS_SUPPORTED 1
+#define G_CREDENTIALS_USE_FREEBSD_CMSGCRED 1
+#define G_CREDENTIALS_NATIVE_TYPE G_CREDENTIALS_TYPE_FREEBSD_CMSGCRED
+#define G_CREDENTIALS_NATIVE_SIZE (sizeof (struct cmsgcred))
+#define G_CREDENTIALS_UNIX_CREDENTIALS_MESSAGE_SUPPORTED 1
+#define G_CREDENTIALS_SPOOFING_SUPPORTED 1
+
 #elif defined(__NetBSD__)
 #define G_CREDENTIALS_SUPPORTED 1
 #define G_CREDENTIALS_USE_NETBSD_UNPCBID 1

Re: fixed: Re: polkit does not start

[karu.pruun]

On Sun, Jan 25, 2015 at 8:36 AM, John Marino wrote:
 On 1/25/2015 06:15, karu.pruun wrote:
 Turns out it was supposed to be a UID. In the end, the error was with
 credentials passing in devel/glib20 that polkitd uses. The attached
 patch fixes it:

 - copy the patch to /usr/dports/devel/glib20/dragonfly/
 - make install

 All the problems listed above go away:

 -upowerd crashes
 -ConsoleKit can't make a session active
 -X login is freezing a minute or two
 -The session menu of xfce is freezing also for some time
 -any logoff from xfve is freezing temporarilly and leaves some
 applications opened.

 I haven't tested xdm etc though.

 I'll try to get it to dports on github.


 Nice, can you start at https://github.com/DragonFlyBSD/DPorts/issues/26
 and see if that's really the same issue?  It's one of our oldest
 unresolved issues.


Sure, can you tell me what is the port it belongs to? Or is it the
/usr/local/libexec/upowerd of

sysutils/upower

that you mean? It used to core dump but now (or at least so far) it's
running fine with the mate desktop I'm using.

I don't have a kde4 desktop on this machine but will get it next week
so I can see if it's working with kde4 as well (or if someone has it
and can test).

Peeter

--

Re: fixed: Re: polkit does not start

[karu.pruun]

On Sun, Jan 25, 2015 at 9:01 AM, John Marino wrote:
 On 1/25/2015 09:58, karu.pruun wrote:
 On Sun, Jan 25, 2015 at 8:36 AM, John Marino wrote:
 On 1/25/2015 06:15, karu.pruun wrote:
 Turns out it was supposed to be a UID. In the end, the error was with
 credentials passing in devel/glib20 that polkitd uses. The attached
 patch fixes it:

 - copy the patch to /usr/dports/devel/glib20/dragonfly/
 - make install

 All the problems listed above go away:

 -upowerd crashes
 -ConsoleKit can't make a session active
 -X login is freezing a minute or two
 -The session menu of xfce is freezing also for some time
 -any logoff from xfve is freezing temporarilly and leaves some
 applications opened.

 I haven't tested xdm etc though.

 I'll try to get it to dports on github.


 Nice, can you start at https://github.com/DragonFlyBSD/DPorts/issues/26
 and see if that's really the same issue?  It's one of our oldest
 unresolved issues.


 Sure, can you tell me what is the port it belongs to? Or is it the
 /usr/local/libexec/upowerd of

 sysutils/upower

 that you mean? It used to core dump but now (or at least so far) it's
 running fine with the mate desktop I'm using.

 I don't have a kde4 desktop on this machine but will get it next week
 so I can see if it's working with kde4 as well (or if someone has it
 and can test).


 It's a generic issue.
 I think you misunderstood.  I just wanted to know if the issue seen
 there is probably the same one you think you solved.  If it is, you'd
 leave your patch on that issue instead of creating a new one.

 the basic issue is that upowerd coredumps due to a misconfigured polkit
 and I never could figure out how to make it happy.


I see. I have to test on the kde4 desktop and see what happens.

I was going to ask how to submit patches to DPorts on github: there
seems no way to attach a file to an issue.

Also, I'll try to submit the patch to the glib maintainers.

Peeter

--

Re: polkit does not start

[karu.pruun]

That's interesting---I wonder if you could tell me what research you
did on dbus and polkit. I would be interested in trying to make it
work. Here's what I've found so far.

I figured out that consolekit and hald are fine, it's polkit that
messes up. Dbus cannot spawn it, and the problem is not dbus, it's
polkit. Since I know almost nothing about dbus and friends my approach
has been to compare with FreeBSD, where it's working fine, and try to
isolate the piece of code where polkit on DragonFly deviates from the
behavior seen on FreeBSD. I ran ktrace on polkit and the deviation
from FreeBSD happens on the second attempt (type: AUTH EXTERNAL) of
authentication:

---FreeBSD (successful)---
  1579 polkitd  CALL  sendto(0x5,0x803076280,0x12,0x2MSG_NOSIGNAL,0,0)
  1579 polkitd  GIO   fd 5 wrote 18 bytes
   AUTH EXTERNAL 30\r
   
  1579 polkitd  RET   sendto 18/0x12
  . . .

  1579 polkitd  CALL  recvfrom(0x5,0x80310f000,0x1000,0,0,0)
  1579 polkitd  GIO   fd 5 read 37 bytes
   OK 42dd72d3f9cfe49615fb1f2154aabce1\r
   
  1579 polkitd  RET   recvfrom 37/0x25
--- ---

vs

---DragonFly---
16116:1 0 polkitd  0.06 CALL
sendto(0x5,0x800740280,0x24,MSG_NOSIGNAL,0,0)
16116:1 0 polkitd  0.10 GIO   fd 5 wrote 36 bytes
   AUTH EXTERNAL 34323934393637323935\r
   
16116:1 0 polkitd  0.06 RET   sendto 36/0x24
. . .

16116:1 0 polkitd  0.000520 CALL  recvfrom(0x5,0x800735000,0x1000,0,0,0)
16116:1 0 polkitd  0.13 GIO   fd 5 read 46 bytes
   REJECTED EXTERNAL DBUS_COOKIE_SHA1 ANONYMOUS\r
   
16116:1 0 polkitd  0.05 RET   recvfrom 46/0x2e
--- ---

The challenge is to find out why DragonFly computes
'34323934393637323935' for the argument of AUTH EXTERNAL while FreeBSD
does '30'. However, I haven't been able to identify the piece of code
in polkit (and glib-2.0) where that particular call above to sendto()
occurs.

I compiled polkit and glib with gdb symbols but it seems I might need
more symbols in perhaps other libraries? In any case, so far the call
to sendto() seems to occur in a piece of code that I can't step in.
I'll keep looking.

Cheers

Peeter

--


On Thu, Jan 8, 2015 at 7:45 PM,  sruss...@prodigeinfo.qc.ca wrote:
 I had the same problem and had to give up. Dbus logs are saying that no
 valid external protocols were found.

 REJECTED EXTERNAL DBUS_COOKIE_SHA1 ANONYMOUS

 498557: [dbus-auth.c(1704):handle_auth] server: Trying mechanism EXTERNAL
 498557: [dbus-auth.c(1624):process_data] server: data: '4294967295'
 498557: [dbus-auth.c(1026):handle_server_data_external_mech] server: no
 credentials, mechanism EXTERNAL can't authenticate
 498557: [dbus-auth.c(430):shutdown_mech] server: Shutting down mechanism
 EXTERNAL

 It happens only for the system dbus. I made some reseach and it occurs when
 dbus is trying to reserve the name of a service. Gio returns:

 Lost the name [...] on session bus.

 I verified the pam confiruration and it looked ok.

 Overall:
 -upowerd crashes
 -ConsoleKit can't make a session active
 -X login is freezing a minute or two
 -The session menu of xfce is freezing also for some time
 -any logoff from xfve is freezing temporarilly and leaves some applications
 opened.

 I spent a lot of time trying to figure out the problem, but I had to give
 up. My workaround was to disable the global dbus service. Dbus is still
 available for applications, but I think that dports is disabling it for
 applications as well, when possible.

 SR


 Le 2015-01-08 10:50, karu.pruun a écrit :


 Hello

 I seem to have troubles getting polkit working. I have in /etc/rc.conf

 ---/etc/rc.conf---
 hald_enable=YES
 dbus_enable=YES
 --- ---

 and this combination results in dbus, hald and consolekit services
 running, but polkitd fails

 ---/var/log/messages---
 Jan  8 17:42:55 dfly dbus[984]: [system] Activating service
 name='org.freedesktop.ConsoleKit' (using servicehelper)
 Jan  8 17:42:55 dfly console-kit-daemon[1010]: WARNING:
 polkit_authority_get: Error getting authority: Error initializing
 authority: Exhausted all available authentication mechanisms (tried:
 EXTERNAL, DBUS_COOKIE_SHA1, ANONYMOUS) (available: EXTERNAL,
 DBUS_COOKIE_SHA1, ANONYMOUS)
 --- ---

 I ran a quick test on a FreeBSD virtual machine, the above two lines
 in rc.conf start dbus, hald, consolekit and polkitd and all seems
 fine. I wonder if I've missed anything on DragonFly?

 Cheers

 Peeter

 --

polkit does not start

[karu.pruun]

Hello

I seem to have troubles getting polkit working. I have in /etc/rc.conf

---/etc/rc.conf---
hald_enable=YES
dbus_enable=YES
--- ---

and this combination results in dbus, hald and consolekit services
running, but polkitd fails

---/var/log/messages---
Jan  8 17:42:55 dfly dbus[984]: [system] Activating service
name='org.freedesktop.ConsoleKit' (using servicehelper)
Jan  8 17:42:55 dfly console-kit-daemon[1010]: WARNING:
polkit_authority_get: Error getting authority: Error initializing
authority: Exhausted all available authentication mechanisms (tried:
EXTERNAL, DBUS_COOKIE_SHA1, ANONYMOUS) (available: EXTERNAL,
DBUS_COOKIE_SHA1, ANONYMOUS)
--- ---

I ran a quick test on a FreeBSD virtual machine, the above two lines
in rc.conf start dbus, hald, consolekit and polkitd and all seems
fine. I wonder if I've missed anything on DragonFly?

Cheers

Peeter

--

Re: AsRock Vision3D 137D, NVidia GeForce 425 (X server conf)

[karu.pruun]

On Tue, Mar 10, 2015 at 2:11 AM, George wrote:
 So these are just wrappers, my bad:

 gox@x140e:~/Downloads/FreeBSD/NVIDIA-FreeBSD-x86_64-346.47$ find . -name
 *.c
 ./src/nvidia_acpi.c
 ./src/nvidia_subr.c
 ./src/nvidia_pci.c
 ./src/nvidia_linux.c
 ./src/nvidia_ctl.c
 ./src/nvidia_os.c
 ./src/nvidia_sysctl.c
 ./src/nvidia_dev.c
 ./src/nvidia_os_registry.c
 ./src/nvidia_i2c.c
 ./src/nvidia_os_pci.c

 OK I see it is a no go I shall stick to VESA and play around with the config
 a bit more.
 Cheers,
 George


I ran into this a while ago when looking for an nvidia driver.

https://forums.freebsd.org/threads/xorg-vesa-driver-massive-speedup-using-mtrr-write-combine.46723/

It's a thread in FreeBSD forums and I thought it'd be fun to try it on
DragonFly but haven't got a moment.

Cheers

Peeter

--

Re: Gnome Show Blank Screen

[karu.pruun]

Seems you have different problems, both with xorg and with the desktop
environment. Get back to the point above where you got the black-grey grid
with X shaped cursor, this shows that Xorg is working but you lack the
desktop environment. This is because your log file shows you put the wrong
path in the .xinitrc file (the DragonFly handbook is outdated). The correct
line should be something in the /usr/local/bin directory, for instance

---.xinitrc---
exec /usr/local/bin/gnome-session
--- ---

so do as it says on the FreeBSD webpage that John Marino gave you above:

% echo exec /usr/local/bin/gnome-session  ~/.xinitrc

this should at least get rid of the error /root/.xinitrc:
/usr/pkg/bin/gnome-session: not found.

If you need to install xorg packages separately then it seems something's
wrong. Normally you only need to write

% sudo pkg install xorg

and this meta package should give you all the xorg you need. I usually get
the graphical desktop running by doing

% sudo pkg install xorg xfce
% echo exec /usr/local/bin/startxfce4  ~/.xinitrc
% startx

(Sometimes need to tweak the xorg.conf too).


Cheers

Peeter

--


On Fri, May 29, 2015 at 12:23 AM, Mohammad BadieZadegan mbzade...@gmail.com
 wrote:


 I installed gnome-session and then see that (on /var/log/Xorg.0.log) I
 must installed fbdev.
 So I installed that and now Xorg note me to install fbdevhw!
 I search fbdevhw on the ports but nothing found!
 What am I doing to install a desktop in DF-BSD?



 On May 28, 2015, at 2:44 PM, Baurzhan Muftakhidinov 
 baurthefi...@gmail.com wrote:

 pkg install gnome-session since your log says (/root/.xinitrc:
 /usr/pkg/bin/gnome-session: not found)

 I've had issues with intel graphics (Sandybridge) also couple of month ago.

 For now I suggest you to try snapshots.


 On Thu, May 28, 2015 at 3:01 PM, Mohammad BadieZadegan 
 mbzade...@gmail.com wrote:

 Hi,
 I commented second device that detected as vesa SCREEN on the
 xorg.conf.new file.
 It leads me to black and grey grid and an X mouse cursor was appeared
 when I run Xorg -config xorg.conf.new -retro.
 Now Still While I try startx the blank screen was appeared and I force to
 reboot by Ctrl+Alt+Del keys.


 unix_usb# startx
 xauth:  file /root/.serverauth.1171 does not exist
 xauth: (argv):1:  bad display name unix_usb:0 in list command
 xauth: (stdin):1:  bad display name unix_usb:0 in add command


 X.Org X Server 1.14.7
 Release Date: 2014-06-05
 X Protocol Version 11, Revision 0
 Build Operating System: DragonFly 4.1-DEVELOPMENT x86_64
 Current Operating System: DragonFly unix_usb 4.0-RELEASE DragonFly
 v4.0.5-RELEASE #29: S x86_64
 Build Date: 03 May 2015  10:32:54AM

 Current version of pixman: 0.32.6
 Before reporting problems, check http://wiki.x.org
 to make sure that you have the latest version.
 Markers: (--) probed, (**) from config file, (==) default setting,
 (++) from command line, (!!) notice, (II) informational,
 (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
 (==) Log file: /var/log/Xorg.0.log, Time: Thu May 28 14:26:31 2015
 (==) Using default built-in configuration (30 lines)
 Initializing built-in extension Generic Event Extension
 Initializing built-in extension SHAPE
 Initializing built-in extension MIT-SHM
 Initializing built-in extension XInputExtension
 Initializing built-in extension XTEST
 Initializing built-in extension BIG-REQUESTS
 Initializing built-in extension SYNC
 Initializing built-in extension XKEYBOARD
 Initializing built-in extension XC-MISC
 Initializing built-in extension XINERAMA
 Initializing built-in extension XFIXES
 Initializing built-in extension RENDER
 Initializing built-in extension RANDR
 Initializing built-in extension COMPOSITE
 Initializing built-in extension DAMAGE
 Initializing built-in extension MIT-SCREEN-SAVER
 Initializing built-in extension DOUBLE-BUFFER
 Initializing built-in extension RECORD
 Initializing built-in extension DPMS
 Initializing built-in extension X-Resource
 Initializing built-in extension XVideo
 Initializing built-in extension XVideo-MotionCompensation
 Initializing built-in extension XFree86-VidModeExtension
 Initializing built-in extension XFree86-DGA
 Initializing built-in extension XFree86-DRI
 Initializing built-in extension DRI2
 Loading extension GLX
 /root/.xinitrc: /usr/pkg/bin/gnome-session: not found
 xinit: connection to X server lost

 waiting for X server to shut down (EE) Server terminated successfully
 (0). Closing log file.

 xauth: (argv):1:  bad display name unix_usb:0 in remove command


 And drop me to Blank screen!
 How can I resolve that?
 Can anyone install a Desktop Environment On the Intel Graphic Device?

 On Thu, May 28, 2015 at 9:48 AM, Mohammad BadieZadegan 
 mbzade...@gmail.com wrote:

 On the last line of Xorg.0.log I see an Error about 2 device finding!

 [   158.875] Number of created screens does not match number of detected
 devices.
   Configuration failed.
 [   158.875] (EE) Server terminated 

Re: Trouble running KDE or Gnome

[karu.pruun]

On Sun, Jun 26, 2016 at 12:11 PM, Bernard Mentink  wrote:
> Hi John,
>
> Appreciate the explanation.
> However, the description of the FreeBSD setup for gnome and the DFly one are
> quite different.
>
> The FreeBSD one says to start dbus and hald  via rc.conf. Dfly does not say
> to do that.
>
> If I follow FreeBSD instructions KDE won't work at all, because KDE starts
> dbus itself and bombs out if it is already running. So at least Dfly works
> with KDE by NOT starting dbus in rc.conf ..


There are two standard dbus session instances normally: one is the
systemwide bus that is started via /etc/rc.conf, the other is a
session bus which is started by KDE, GNOME, XFCE, MATE when they
launch. So you should include

dbus_enable="YES"

in rc.conf. If you do not do this, you'll see some complaining in the
log; but may still be able to run the DE. This is how I run KDE and I
can confirm that it runs fine on my machine DragonFly
v4.5.0.1247.g122b6-DEVELOPMENT. MATE and XFCE run fine as well. I
don't run hald for any of the DEs. Not sure why KDE does not run when
you start the system dbus instance.

GNOME (gnome3) does not work, or at least to my experience. There's
some complaining by gnome-settings-daemon about not being able to
communicate via dbus, and it may well be the reason but I haven't
followed up.


> PS: Man pages don't help a new user get up and running when they are used to
> the Linux way of doing things  for example, the Lack of the lspci
> instruction means I can't see what module I need to load  for my audio
> hardware ...


pciconf -l
pciconf -lv

A lot of this applies to DragonFly as well:

https://www.freebsd.org/doc/en/articles/linux-users/article.html


Cheers

Peeter

--




>> On 6/26/2016 3:38 AM, Bernard Mentink wrote:
>>>
>>> Hi John,
>>>
>>> So why should I use the FreeBSD instructions? I thought this was
>>> DragonFlyBSD forum?
>>>
>>> I used the instructions from
>>> here: https://www.dragonflybsd.org/docs/handbook/X/#index2h2
>>>
>>> Are you saying I should not use the DFly instructions to setup DFly?
>>>
>>> That is just weird ..
>>
>>
>> Hi Bernard,
>> Of course it would be ideal if DragonFly documentation was at the same
>> level as FreeBSD but the fact is that this is not the case.  FreeBSD has an
>> entire team whose primary responsibility is documentation.
>>
>> Now you may be aware that DragonFly DPorts is basically FreeBSD ports made
>> to work on DragonFly.  As such, instructions to install third-party software
>> on FreeBSD are good for DragonFly as well.  So it's not weird to reference
>> FreeBSD documentation when DPorts are involved as is the case here.
>>
>> As a last thought, our handbook is a living wiki.  Some people have
>> improved KDE instructions recently, but if you think it needs more then you
>> have the ability to write these improvements so everyone can benefit.
>>
>> DragonFly MAN pages are top notch, but the handbook / wiki is not at the
>> same quality level (and in some cases are poor quality).
>>
>> John
>>
>>
>> ---
>> This email has been checked for viruses by Avast antivirus software.
>> https://www.avast.com/antivirus
>>
>
>
>
> --
> -
> Life is a sexually transmitted disease with a 100% mortality rate.
> -

Re: Trouble running KDE or Gnome

[karu.pruun]

On Sun, Jun 26, 2016 at 9:37 PM, Bernard Mentink <bment...@gmail.com> wrote:
> Now the only other issue I have running KDE is that when I enable sound on
> my platform by loading the snd_hda module, I get  pulseaudio taking up 100%
> of one of my cpu cores ...
> Sound works ok though  any idea what goes on there? do I have to adjust
> some pulseaudio parameters?


Yes I've noticed that too, but sorry no idea what's going on or how to
fix that; I haven't followed that up. Let us know if you find out.

Cheers

Peeter

--




>
> Cheers,
> Bernie
>
> On Mon, Jun 27, 2016 at 2:06 AM, karu.pruun <karu.pr...@gmail.com> wrote:
>>
>> On Sun, Jun 26, 2016 at 12:11 PM, Bernard Mentink <bment...@gmail.com>
>> wrote:
>> > Hi John,
>> >
>> > Appreciate the explanation.
>> > However, the description of the FreeBSD setup for gnome and the DFly one
>> > are
>> > quite different.
>> >
>> > The FreeBSD one says to start dbus and hald  via rc.conf. Dfly does not
>> > say
>> > to do that.
>> >
>> > If I follow FreeBSD instructions KDE won't work at all, because KDE
>> > starts
>> > dbus itself and bombs out if it is already running. So at least Dfly
>> > works
>> > with KDE by NOT starting dbus in rc.conf ..
>>
>>
>> There are two standard dbus session instances normally: one is the
>> systemwide bus that is started via /etc/rc.conf, the other is a
>> session bus which is started by KDE, GNOME, XFCE, MATE when they
>> launch. So you should include
>>
>> dbus_enable="YES"
>>
>> in rc.conf. If you do not do this, you'll see some complaining in the
>> log; but may still be able to run the DE. This is how I run KDE and I
>> can confirm that it runs fine on my machine DragonFly
>> v4.5.0.1247.g122b6-DEVELOPMENT. MATE and XFCE run fine as well. I
>> don't run hald for any of the DEs. Not sure why KDE does not run when
>> you start the system dbus instance.
>>
>> GNOME (gnome3) does not work, or at least to my experience. There's
>> some complaining by gnome-settings-daemon about not being able to
>> communicate via dbus, and it may well be the reason but I haven't
>> followed up.
>>
>>
>> > PS: Man pages don't help a new user get up and running when they are
>> > used to
>> > the Linux way of doing things  for example, the Lack of the lspci
>> > instruction means I can't see what module I need to load  for my audio
>> > hardware ...
>>
>>
>> pciconf -l
>> pciconf -lv
>>
>> A lot of this applies to DragonFly as well:
>>
>> https://www.freebsd.org/doc/en/articles/linux-users/article.html
>>
>>
>> Cheers
>>
>> Peeter
>>
>> --
>>
>>
>>
>>
>> >> On 6/26/2016 3:38 AM, Bernard Mentink wrote:
>> >>>
>> >>> Hi John,
>> >>>
>> >>> So why should I use the FreeBSD instructions? I thought this was
>> >>> DragonFlyBSD forum?
>> >>>
>> >>> I used the instructions from
>> >>> here: https://www.dragonflybsd.org/docs/handbook/X/#index2h2
>> >>>
>> >>> Are you saying I should not use the DFly instructions to setup DFly?
>> >>>
>> >>> That is just weird ..
>> >>
>> >>
>> >> Hi Bernard,
>> >> Of course it would be ideal if DragonFly documentation was at the same
>> >> level as FreeBSD but the fact is that this is not the case.  FreeBSD
>> >> has an
>> >> entire team whose primary responsibility is documentation.
>> >>
>> >> Now you may be aware that DragonFly DPorts is basically FreeBSD ports
>> >> made
>> >> to work on DragonFly.  As such, instructions to install third-party
>> >> software
>> >> on FreeBSD are good for DragonFly as well.  So it's not weird to
>> >> reference
>> >> FreeBSD documentation when DPorts are involved as is the case here.
>> >>
>> >> As a last thought, our handbook is a living wiki.  Some people have
>> >> improved KDE instructions recently, but if you think it needs more then
>> >> you
>> >> have the ability to write these improvements so everyone can benefit.
>> >>
>> >> DragonFly MAN pages are top notch, but the handbook / wiki is not at
>> >> the
>> >> same quality level (and in some cases are poor quality).
>> >>
>> >> John
>> >>
>> >>
>> >> ---
>> >> This email has been checked for viruses by Avast antivirus software.
>> >> https://www.avast.com/antivirus
>> >>
>> >
>> >
>> >
>> > --
>> >
>> > -
>> > Life is a sexually transmitted disease with a 100% mortality rate.
>> >
>> > -
>
>
>
>
> --
> -
> Life is a sexually transmitted disease with a 100% mortality rate.
> -

Re: Xorg problem on Skylake

[karu.pruun]

Thank you, I'll try this tonight.

On the bright side, Xorg is kind of stable in the sense that the
corruption is systematic and the server does not crash.

Peeter

--


On Mon, Feb 15, 2016 at 9:39 AM, Matthew Dillon <dil...@backplane.com> wrote:
> On skylake, not yet... all you can do is turn off acceleration in
> /etc/X11/xorg.conf (if you are running X without an /etc/X11/xorg.conf you
> have to mess around and set one up, then turn off acceleration in it).
> SkyLake support only just recently started to go in and it is guaranteed to
> be unstable.  We haven't synchronized the code to the point where it is
> supported well yet.
>
> -Matt
>
> On Sun, Feb 14, 2016 at 2:02 PM, karu.pruun <karu.pr...@gmail.com> wrote:
>>
>> Hello
>>
>> I am having trouble getting Xorg working on a Skylake machine with
>> Intel integrated graphics (dmesg and Xorg log attached; a few errors
>> on drm at the end of dmesg too), on this snapshot:
>>
>> ---
>> DragonFly mars 4.5-DEVELOPMENT DragonFly 4.5-DEVELOPMENT #0: Wed Feb
>> 10 11:57:30 UTC 2016
>> r...@pkgbox64.dragonflybsd.org:/usr/obj/usr/src/sys/X86_64_GENERIC
>> x86_64
>> ---
>>
>> I am not using xorg.conf and start Xorg with startx and .xinitrc that
>> starts a mate-session. Several issues:
>>
>> - Xorg starts up and for some reason outputs to two devices, eDP1 and
>> HDMI2. I only have one monitor at HDMI2; it can do 1680x1050 but since
>> Xorg finds that eDP1 can have max resolution up to 1024x760 it forces
>> the latter resolution on my monitor. Can I force it have only HDMI2,
>> i.e. the one that actually has a monitor attached?
>>
>> - But even at 1024x768 the screen starts corrupting when I move
>> windows, text becomes occasionally blurred etc. When I log out of mate
>> to return to text screen, I get a black screen. The system is not
>> frozen, all I can do is reboot.
>>
>> - In the 1024x768 mode, when I try to change the resolution with xrandr:
>>
>> # xrandr --output HDMI2 --auto
>>
>> Xorg does change the mode to 1680x1050, i.e. the native resolution of
>> the monitor, but the screen is half-corrupted and unusable. When I
>> exit to text screen I succeed the first time; but the text mode is now
>> something like 128x48 or similar and in the top left quarter of the
>> display. When I start Xorg again I get a half-corrupted screen; upon
>> return to text mode it gives a blank or a frozen screen---all I can do
>> is reboot.
>>
>> Is there anything one can do?
>>
>> Thanks
>>
>> Peeter
>>
>> --
>
>

Re: Xorg problem on Skylake

[karu.pruun]

I can report some success. I took advice from this thread

https://bbs.archlinux.org/viewtopic.php?id=140315

in xorg.conf I disabled eDP1 and allowed only HDMI2 with mode
"1680x1050". This gives me the desired resolution and a working and
stable display. I did not turn off acceleration. There's occasional
minor blurring of text in e.g. Firefox which goes away when you scroll
down etc. But all in all it's stable so far.

The problem is obviously changing between different modes and back to
text---that does not work. But since I won't need it anyway I'm happy
so far.

Cheers

Peeter

--







On Mon, Feb 15, 2016 at 12:55 PM, karu.pruun <karu.pr...@gmail.com> wrote:
> Thank you, I'll try this tonight.
>
> On the bright side, Xorg is kind of stable in the sense that the
> corruption is systematic and the server does not crash.
>
> Peeter
>
> --
>
>
> On Mon, Feb 15, 2016 at 9:39 AM, Matthew Dillon <dil...@backplane.com> wrote:
>> On skylake, not yet... all you can do is turn off acceleration in
>> /etc/X11/xorg.conf (if you are running X without an /etc/X11/xorg.conf you
>> have to mess around and set one up, then turn off acceleration in it).
>> SkyLake support only just recently started to go in and it is guaranteed to
>> be unstable.  We haven't synchronized the code to the point where it is
>> supported well yet.
>>
>> -Matt
>>
>> On Sun, Feb 14, 2016 at 2:02 PM, karu.pruun <karu.pr...@gmail.com> wrote:
>>>
>>> Hello
>>>
>>> I am having trouble getting Xorg working on a Skylake machine with
>>> Intel integrated graphics (dmesg and Xorg log attached; a few errors
>>> on drm at the end of dmesg too), on this snapshot:
>>>
>>> ---
>>> DragonFly mars 4.5-DEVELOPMENT DragonFly 4.5-DEVELOPMENT #0: Wed Feb
>>> 10 11:57:30 UTC 2016
>>> r...@pkgbox64.dragonflybsd.org:/usr/obj/usr/src/sys/X86_64_GENERIC
>>> x86_64
>>> ---
>>>
>>> I am not using xorg.conf and start Xorg with startx and .xinitrc that
>>> starts a mate-session. Several issues:
>>>
>>> - Xorg starts up and for some reason outputs to two devices, eDP1 and
>>> HDMI2. I only have one monitor at HDMI2; it can do 1680x1050 but since
>>> Xorg finds that eDP1 can have max resolution up to 1024x760 it forces
>>> the latter resolution on my monitor. Can I force it have only HDMI2,
>>> i.e. the one that actually has a monitor attached?
>>>
>>> - But even at 1024x768 the screen starts corrupting when I move
>>> windows, text becomes occasionally blurred etc. When I log out of mate
>>> to return to text screen, I get a black screen. The system is not
>>> frozen, all I can do is reboot.
>>>
>>> - In the 1024x768 mode, when I try to change the resolution with xrandr:
>>>
>>> # xrandr --output HDMI2 --auto
>>>
>>> Xorg does change the mode to 1680x1050, i.e. the native resolution of
>>> the monitor, but the screen is half-corrupted and unusable. When I
>>> exit to text screen I succeed the first time; but the text mode is now
>>> something like 128x48 or similar and in the top left quarter of the
>>> display. When I start Xorg again I get a half-corrupted screen; upon
>>> return to text mode it gives a blank or a frozen screen---all I can do
>>> is reboot.
>>>
>>> Is there anything one can do?
>>>
>>> Thanks
>>>
>>> Peeter
>>>
>>> --
>>
>>

Re: compiling dports with clang

[karu.pruun]

On Tue, Mar 22, 2016 at 4:20 PM, John Marino wrote:
> On 3/22/2016 3:15 PM, karu.pruun wrote:
>>
>> Hello
>>
>> I was trying to compile www/node with clang:
>>
>> # env CCVER=clang36 make install
>>
>> but the result says it's gcc 5.3.1:
>>
>> # objdump -s -j .comment /usr/local/bin/node
>>
>> .GCC: 5.3.1 [DragonFly] Release/2015-12-04
>>
>> while a clang compiled file should say:
>>
>> .GCC: 5.3.1 [DragonFly] Release/2015-12-04.clang version 3.6.2
>> (tags/RELEASE_362/final)
>>
>> What did I miss?
>>
>
> It's intentionally not supported.  Don't do it.
> (Why would you even want to?)
>
> The ports framework will override any setting of CCVER and as I just
> mentioned, that's intentional.


Thanks. It was just an idle thought that since FreeBSD uses clang,
possibly some of the problems with node might be solved by using clang
for compilation. Currently, around 100+ tests (out of 800+) do not
pass under DragonFly. Many of those pass under FreeBSD. I realize that
most of the issues are not compiler related; eg ps in DragonFly does
not support args etc (submitted a patch recently). But as I said, just
an idle thought.

Peeter

--

Wayland/Weston with XWayland works on DragonFly

[karu.pruun]

Hello

So I compiled Xorg with XWayland support and got it working: many
applications that need Xorg work now with wayland/weston. Basically,
Xwayland is the "Xorg" for applications that need Xorg; it's a
compatibility option as long as an application does not work with
wayland/weston directly.

I followed directions on this page:

https://wayland.freedesktop.org/xserver.html

and recompiled x11-servers/xorg-server with options given on that
page. It does not compile cleanly but complains about a missing
file/library; so before compiling, don't uninstall your xorg-server
yet. When the error occurs, I manually copied the file to the build
directory in /usr/obj/. . . (see what is the file and the directory in
error message) and then remove the xorg-server package. Then "make
install" worked fine. (Someone know how to fix this issue?) I then
used the attached weston.ini (save to ~/.config/ and edit paths of
files like background etc) and manually created

> mkdir /tmp/.X11-unix

and then (from webapage https://github.com/DragonFlyBSD/DeltaPorts/pull/123)

> sudo kldload i915
> mkdir /tmp/wayland_xdg
> chmod 0700 /tmp/wayland_xdg
> env XDG_RUNTIME_DIR=/tmp/wayland_xdg weston-launch -- --use-pixman

Applications that seem to work and are stable (there might more but
haven't tried):

gtk3:
  - gedit
  - nautilus
  - evince

xfce4:
  - xfce4-terminal
  - atril

- firefox
- spyder
- scilab
- kile (crashes first but works when launched again)

Doesn't work:
  - chrome (segfault)

Weston is quite minimalistic but does provide a stable working
environment. No tearing and runs smoothly:

weston:
  - has workspaces (switch: super + F1, F2 etc)
  - a panel with launchers
  - switch between apps: super (or ctrl) + tab
  - but: no tray for minimized apps: they can be brought back by
switching between apps
  - Copy-paste works in X and between X and weston
- X cursor could be better configured?

In summary: I am very much impressed. On this machine, wayland/weston
feels faster than X. It's stable, or at least so far. With Xorg I
can't switch between VT and graphical screen more than twice; the
screen hangs after two switches. With Wayland it just works.

I wonder if one might get a minimal DE like Maynard running. It seems
to need wayland support in gtk30.

Cheers

Peeter

--


weston.ini
Description: Binary data

Re: Wayland/Weston with XWayland works on DragonFly

[karu.pruun]

I've tested KDE in the sense of running a KDE application (kile) using
XWayland. It's like using GNOME or XFCE or MATE to run a KDE
application.

Running a full fledged KDE with XWayland (startkde4) did not go well;
basically you already have a minimal native DE, i.e. Weston, and then
you run another DE, i.e. KDE on top of XWayaland --- KDE was
unusuable, even kile did not start; whereas running just kile as
above, i.e. just as an application with Weston gave a functional app.

The KDE effort to rewrite KDE for Wayland is documented here

https://community.kde.org/KWin/Wayland

As far as I understand all of the major desktop environments are
currently in transition of adding native support to Wayland. Until
this is done, that is, until they do not run natively under Wayland
one has to use a compatibility backend like XWayland to run
applications that rely on Xorg.

Cheers

Peeter

--



On Fri, May 20, 2016 at 4:08 PM, PeerCorps Trust Fund
<i...@peercorpstrust.org> wrote:
> Hi Peter,
>
> This is quite interesting. Have you tested KDE at all with that
> configuration?
>
>
> On 05/20/2016 04:05 PM, karu.pruun wrote:
>>
>> Hello
>>
>> So I compiled Xorg with XWayland support and got it working: many
>> applications that need Xorg work now with wayland/weston. Basically,
>> Xwayland is the "Xorg" for applications that need Xorg; it's a
>> compatibility option as long as an application does not work with
>> wayland/weston directly.
>>
>> I followed directions on this page:
>>
>> https://wayland.freedesktop.org/xserver.html
>>
>> and recompiled x11-servers/xorg-server with options given on that
>> page. It does not compile cleanly but complains about a missing
>> file/library; so before compiling, don't uninstall your xorg-server
>> yet. When the error occurs, I manually copied the file to the build
>> directory in /usr/obj/. . . (see what is the file and the directory in
>> error message) and then remove the xorg-server package. Then "make
>> install" worked fine. (Someone know how to fix this issue?) I then
>> used the attached weston.ini (save to ~/.config/ and edit paths of
>> files like background etc) and manually created
>>
>>> mkdir /tmp/.X11-unix
>>
>>
>> and then (from webapage
>> https://github.com/DragonFlyBSD/DeltaPorts/pull/123)
>>
>>> sudo kldload i915
>>> mkdir /tmp/wayland_xdg
>>> chmod 0700 /tmp/wayland_xdg
>>> env XDG_RUNTIME_DIR=/tmp/wayland_xdg weston-launch -- --use-pixman
>>
>>
>> Applications that seem to work and are stable (there might more but
>> haven't tried):
>>
>> gtk3:
>>- gedit
>>- nautilus
>>- evince
>>
>> xfce4:
>>- xfce4-terminal
>>- atril
>>
>> - firefox
>> - spyder
>> - scilab
>> - kile (crashes first but works when launched again)
>>
>> Doesn't work:
>>- chrome (segfault)
>>
>> Weston is quite minimalistic but does provide a stable working
>> environment. No tearing and runs smoothly:
>>
>> weston:
>>- has workspaces (switch: super + F1, F2 etc)
>>- a panel with launchers
>>- switch between apps: super (or ctrl) + tab
>>- but: no tray for minimized apps: they can be brought back by
>> switching between apps
>>- Copy-paste works in X and between X and weston
>>  - X cursor could be better configured?
>>
>> In summary: I am very much impressed. On this machine, wayland/weston
>> feels faster than X. It's stable, or at least so far. With Xorg I
>> can't switch between VT and graphical screen more than twice; the
>> screen hangs after two switches. With Wayland it just works.
>>
>> I wonder if one might get a minimal DE like Maynard running. It seems
>> to need wayland support in gtk30.
>>
>> Cheers
>>
>> Peeter
>>
>> --
>>
>

Re: Wayland/Weston with XWayland works on DragonFly

[karu.pruun]

I had a look but it seems the missing dependency is wayland-egl, which
I don't know how to build. It should be built by graphics/libEGL, but
a naively adding 'wayland' in the list of 'CONFIGURE_ARGS' in
graphics/libGL/Makefile.common does not work.

It'd be nice to have all the various toolkits (EFL, gtk30, Qt) build
on DragonFly. I'll dig further if I have time next week.

Peeter

--



On Sat, May 21, 2016 at 1:57 PM, Carsten Mattner
<carstenmatt...@gmail.com> wrote:
> If you have time, can you test Enlightenment with their Wayland
> compositor and report back?
>
> You may have to build from GIT, but they have instructions/scripts
> for that.
>
> On Fri, May 20, 2016 at 8:47 PM, karu.pruun <karu.pr...@gmail.com> wrote:
>> I've tested KDE in the sense of running a KDE application (kile) using
>> XWayland. It's like using GNOME or XFCE or MATE to run a KDE
>> application.
>>
>> Running a full fledged KDE with XWayland (startkde4) did not go well;
>> basically you already have a minimal native DE, i.e. Weston, and then
>> you run another DE, i.e. KDE on top of XWayaland --- KDE was
>> unusuable, even kile did not start; whereas running just kile as
>> above, i.e. just as an application with Weston gave a functional app.
>>
>> The KDE effort to rewrite KDE for Wayland is documented here
>>
>> https://community.kde.org/KWin/Wayland
>>
>> As far as I understand all of the major desktop environments are
>> currently in transition of adding native support to Wayland. Until
>> this is done, that is, until they do not run natively under Wayland
>> one has to use a compatibility backend like XWayland to run
>> applications that rely on Xorg.
>>
>> Cheers
>>
>> Peeter
>>
>> --
>>
>>
>>
>> On Fri, May 20, 2016 at 4:08 PM, PeerCorps Trust Fund
>> <i...@peercorpstrust.org> wrote:
>>> Hi Peter,
>>>
>>> This is quite interesting. Have you tested KDE at all with that
>>> configuration?
>>>
>>>
>>> On 05/20/2016 04:05 PM, karu.pruun wrote:
>>>>
>>>> Hello
>>>>
>>>> So I compiled Xorg with XWayland support and got it working: many
>>>> applications that need Xorg work now with wayland/weston. Basically,
>>>> Xwayland is the "Xorg" for applications that need Xorg; it's a
>>>> compatibility option as long as an application does not work with
>>>> wayland/weston directly.
>>>>
>>>> I followed directions on this page:
>>>>
>>>> https://wayland.freedesktop.org/xserver.html
>>>>
>>>> and recompiled x11-servers/xorg-server with options given on that
>>>> page. It does not compile cleanly but complains about a missing
>>>> file/library; so before compiling, don't uninstall your xorg-server
>>>> yet. When the error occurs, I manually copied the file to the build
>>>> directory in /usr/obj/. . . (see what is the file and the directory in
>>>> error message) and then remove the xorg-server package. Then "make
>>>> install" worked fine. (Someone know how to fix this issue?) I then
>>>> used the attached weston.ini (save to ~/.config/ and edit paths of
>>>> files like background etc) and manually created
>>>>
>>>>> mkdir /tmp/.X11-unix
>>>>
>>>>
>>>> and then (from webapage
>>>> https://github.com/DragonFlyBSD/DeltaPorts/pull/123)
>>>>
>>>>> sudo kldload i915
>>>>> mkdir /tmp/wayland_xdg
>>>>> chmod 0700 /tmp/wayland_xdg
>>>>> env XDG_RUNTIME_DIR=/tmp/wayland_xdg weston-launch -- --use-pixman
>>>>
>>>>
>>>> Applications that seem to work and are stable (there might more but
>>>> haven't tried):
>>>>
>>>> gtk3:
>>>>- gedit
>>>>- nautilus
>>>>- evince
>>>>
>>>> xfce4:
>>>>- xfce4-terminal
>>>>- atril
>>>>
>>>> - firefox
>>>> - spyder
>>>> - scilab
>>>> - kile (crashes first but works when launched again)
>>>>
>>>> Doesn't work:
>>>>- chrome (segfault)
>>>>
>>>> Weston is quite minimalistic but does provide a stable working
>>>> environment. No tearing and runs smoothly:
>>>>
>>>> weston:
>>>>- has workspaces (switch: super + F1, F2 etc)
>>>>- a panel with launchers
>>>>- switch between apps: super (or ctrl) + tab
>>>>- but: no tray for minimized apps: they can be brought back by
>>>> switching between apps
>>>>- Copy-paste works in X and between X and weston
>>>>  - X cursor could be better configured?
>>>>
>>>> In summary: I am very much impressed. On this machine, wayland/weston
>>>> feels faster than X. It's stable, or at least so far. With Xorg I
>>>> can't switch between VT and graphical screen more than twice; the
>>>> screen hangs after two switches. With Wayland it just works.
>>>>
>>>> I wonder if one might get a minimal DE like Maynard running. It seems
>>>> to need wayland support in gtk30.
>>>>
>>>> Cheers
>>>>
>>>> Peeter
>>>>
>>>> --
>>>>
>>>

Re: skylake video

[karu.pruun]

Can you post your xorg.conf to force vsync etc? How did you start compton?

Thanks

Peeter

--


On Wed, May 11, 2016 at 3:32 PM, Carsten Mattner
 wrote:
> Have you tried compton to force vsync for avoiding tearing? It worked for me.
> Also when using mpv to play videos with vaapi output set to opengl works
> tear-free too. Moreover you should enable the tearfree options in xorg.conf.

Re: skylake video

[karu.pruun]

Yes it's integrated intel on skylake.

I also included these

---
Option  "ReprobeOutputs" "TRUE"
Option  "TripleBuffer" "FALSE"
Option  "SwapbuffersWait" "FALSE"
---

from Matt Dillon's notes on 4k monitors, but don't know if they are
strictly necessary.

Enlightenment 0.20 is said to have wayland/weston support. I wonder if
one can build it on DragonFly and if it might have a terminal
emulator.

Peeter

--



On Tue, May 17, 2016 at 12:13 AM, Carsten Mattner
<carstenmatt...@gmail.com> wrote:
> On 5/16/16, karu.pruun <karu.pr...@gmail.com> wrote:
>> I have to correct a mistake: I did get compton + uxa with xfce4
>> working tearfree on a 4k monitor. I used this conf file
>
> On Intel?
>
>> http://duncanlock.net/blog/2013/06/07/how-to-switch-to-compton-for-beautiful-tear-free-compositing-in-xfce/
>>
>> with the glx backend.
>
> Glad to hear it finally worked for you.
>
>> On a different note: I was very much impressed by wayland/weston.
>> Following these instructions
>>
>> https://github.com/DragonFlyBSD/DeltaPorts/pull/123
>>
>> one gets a teafree smooth graphical environment. No apps apart from a
>> terminal but looks promising. I wonder if the xwayland module would
>> allow run the usual xorg applications? The xwayland module did not
>> load; complained about the missing X11 socket.
>
> The bundled wayland terminal emulator isn't very good and it's a pain
> that there's no xterm or rxvt-unicode yet since Konsole and all VTE-based
> emulators are incompatible or bad in some way.
>
> Were it not for that, I might consider using Weston.

Re: skylake video

[karu.pruun]

I have to correct a mistake: I did get compton + uxa with xfce4
working tearfree on a 4k monitor. I used this conf file

http://duncanlock.net/blog/2013/06/07/how-to-switch-to-compton-for-beautiful-tear-free-compositing-in-xfce/

with the glx backend.

On a different note: I was very much impressed by wayland/weston.
Following these instructions

https://github.com/DragonFlyBSD/DeltaPorts/pull/123

one gets a teafree smooth graphical environment. No apps apart from a
terminal but looks promising. I wonder if the xwayland module would
allow run the usual xorg applications? The xwayland module did not
load; complained about the missing X11 socket.

Cheers

Peeter

--





On Thu, May 12, 2016 at 3:29 PM, Carsten Mattner
 wrote:
> On Thu, May 12, 2016 at 1:47 PM, Andrew Slaughter wrote:
>> Using compton with backend "glx" and vsync "opengl-swc" ("sna" acceleration
>> and "tearfree" enabled in the driver) reduced the tearing significantly for
>> me (on both XFCE and spectrwm), although it didn't eliminate it (mainly left
>> with some minor corruption of fonts in firefox). It seemed usable, though.
>
> After testing it for a day, I can say that DRI 3 and Accel glamor don't solve
> tearing and cause more tearing actually. Back to default DRI 2 and SNA.
>
> Does anybody have an AMD gpu to report how much tearing happens there?
> I'm fed up with the frequent regressions and missing QA in Intel GPUs
> during the last year. It only got worse since the linux-4.3 atomic modesetting
> changes.

Re: libGL error: failed to open drm device: Permission denied

[karu.pruun]

Thank you! This fixed it: got a 4k monitor with AccelMethod "uxa" +
compton working.

Cheers

Peeter

--


On Mon, May 16, 2016 at 3:57 PM, Andrew Slaughter
<andrew.slaugh...@gmail.com> wrote:
> Have you added yourself to the video group?
>
> On Mon, May 16, 2016 at 3:35 AM, karu.pruun <karu.pr...@gmail.com> wrote:
>>
>> Hello
>>
>> Starting the compton compositor gives this error:
>>
>> ---
>> libGL error: failed to open drm device: Permission denied
>> libGL error: failed to load driver: i965
>> ---
>>
>> How can the permissions be fixed? There's no device /dev/drm. . .
>>
>> Cheers
>>
>> Peeter
>>
>> --
>
>

libGL error: failed to open drm device: Permission denied

[karu.pruun]

Hello

Starting the compton compositor gives this error:

---
libGL error: failed to open drm device: Permission denied
libGL error: failed to load driver: i965
---

How can the permissions be fixed? There's no device /dev/drm. . .

Cheers

Peeter

--

Re: skylake video

[karu.pruun]

Thanks for this --- indeed, corruption gone!

This error I see in the log:

--
May 11 09:19:06 dfly kernel: error:
[drm:pid44841:intel_pipe_update_start] *ERROR* Potential atomic update
failure on pipe B
--


Cheers

Peeter

--



On Tue, May 10, 2016 at 11:42 PM, Andrew Slaughter
 wrote:
> Just to throw out a single data point for the curious, video support for
> Skylake on master seems to be working pretty well for me.
>
> At first there was a lot of tearing and corruption (usable, sort of), but
> using "uxa" instead of "sna" acceleration seems to have fixed things so far:
>
> /etc/X11/xorg.conf.d/20-intel.conf (taken from the Arch wiki):
>
> Section "Device"
> Identifier "Intel Graphics"
> Driver "intel"
> Option "AccelMethod" "uxa"
> EndSection
>
> Trying to watch youtube was pretty painful, but it works well for what I do
> (lots of text, light web browsing).
>
> FWIW, using "sna" and the "tearfree" option didn't improve the corruption.

Re: Wayland Weston

[karu.pruun]

Did you have any luck running wayland with i915? What wm/desktop did
you use? I read enlightenment 0.20 supports wayland. . .

Cheers

Peeter

--


On Fri, May 6, 2016 at 11:49 AM, Bernard Mentink  wrote:
> Thanks,
>
> I was looking for the kernel module called i915kms, as I thought that was
> the DRM version  guess I was looking in the wrong place :)
>
> I will try the standard i915 module. Cheers.
>
> On Fri, May 6, 2016 at 8:42 PM, Imre Vadasz  wrote:
>>
>> Hi,
>>
>> Maybe you just need to manually run "kldload i915" or "kldload radeonkms"
>> (depending on whether your graphics-chipset is from AMD or Intel), before
>> trying to launch weston.  At the moment, weston doesn't load the
>> kms-driver
>> automatically like Xorg does.
>>
>> With an nvidia GPU you are out of luck, since the nouveau driver port to
>> DragonFly isn't finished yet.
>>
>> Imre
>>
>> On 15:39 Fri 06 May , Bernard Mentink wrote:
>> > Hi,
>> >
>> > I tried that, and wayland tried to run, then failed saying it can't find
>> > drm ..
>> > ( weston run with --backend=drm-backend.so)
>> >
>> > What do I have to install to get the drm-backend?
>> >
>> > Cheers,
>> >
>> > On Thu, May 5, 2016 at 6:37 PM, Imre Vadasz  wrote:
>> >
>> > > Hi,
>> > >
>> > > DragonFly actually has a udevd daemon, which needs to be activated
>> > > with
>> > > udevd_enable="YES" in /etc/rc.conf for the weston port to work with
>> > > the
>> > > DRM backend.
>> > >
>> > > The DragonFly udev is somewhat similar in naming and functionality to
>> > > Linux's udev, but e.g. it uses a slightly different API in the library
>> > > (which is called libdevattr instead of linux's libudev).
>> > >
>> > > Imre
>> > >
>> > > On 15:13 Thu 05 May , Bernard Mentink wrote:
>> > > > Hi Guys,
>> > > >
>> > > > Just ran up weston/wayland to see how the port is going, link here:
>> > > > https://github.com/DragonFlyBSD/DeltaPorts/pull/123
>> > > >
>> > > > It works fine in an X11 window, but having trouble via DRM. I
>> > > > followed
>> > > the
>> > > > guide, but when run in a tty, get the following error:
>> > > > *wayland-launch:
>> > > > udev-new failed* ..
>> > > >
>> > > > I did not know dragonfly was using udev, thought it was only devd ..
>> > > > Forgive my ignorance .. and I know it is a work in progress ..
>> > > >
>> > > > Be good to see it up and running though ...
>> > > >
>> > > > Bernie
>> > > >
>> > > > --
>> > > >
>> > >
>> > > -
>> > > > *BSD is what you get when a bunch of UNIX hackers sit down to try to
>> > > port a
>> > > > UNIX system to the PC. Linux is what you get when a bunch of PC
>> > > > hackers
>> > > sit
>> > > > down and try to write a UNIX system for the PC*
>> > > >
>> > >
>> > > 
>> > >
>> >
>> >
>> >
>> > --
>> >
>> > -
>> > *BSD is what you get when a bunch of UNIX hackers sit down to try to
>> > port a
>> > UNIX system to the PC. Linux is what you get when a bunch of PC hackers
>> > sit
>> > down and try to write a UNIX system for the PC*
>> >
>> > -
>
>
>
>
> --
> -
> BSD is what you get when a bunch of UNIX hackers sit down to try to port a
> UNIX system to the PC. Linux is what you get when a bunch of PC hackers sit
> down and try to write a UNIX system for the PC
> 

Re: skylake video

[karu.pruun]

So far no luck. Combining these (on skylake, i915)

(1) compton --backend xr_glx_hybrid
(2) compton --backend glx
(3) Tearfree + SwapbuffersWait

gives tearing:

(1) + (3): tearing and hang
(2) + (3): tearing
(2): some tearing and slow

in addition, keyboard and mouse become sluggish on occasion.

I am using MATE but that should not matter. In sum, so far the
fastest, tearfree desktop I had with

Option "AccelMethod" "none"

Strangely, now this also gives sluggish mouse and unusable keyboard.


Peeter

--


On Wed, May 11, 2016 at 10:35 PM, Carsten Mattner
 wrote:
> On Wed, May 11, 2016 at 9:07 PM, Bernard Mentink  wrote:
>> Hi Carsten,
>>
>> Will starting compton also allow my bar to work in 3D instead of 2D ( Cairo
>> Dock)?
>> Do I just execute compton while I am in an X session 
>
> I only use compton to force vsync and otherwise rely on Xmonad,
> so I don't have any experience with trying Cairo dock, but I can
> confirm that OpenGL still works when compton is running, if that's
> what you're asking.
>
> But I'd suggest to try DRI3 and TearFree or first just TearFree
> to see how well it solves the tearing issues. I'm testing Accelaration 
> "glamor"
> with DRI "3" and TearFree "true" right now and no compton at all.
> Let's see if this works without video corruption and tearing for
> a long time...

Re: loading i915 with two graphics cards

[karu.pruun]

On Sun, Jul 17, 2016 at 12:50 PM, Francois Tigeot 
wrote:

> >
> > the system (DragonFly 4.5) gets stuck, the screen is frozen at console. I
> > can only connect an external keyboard and press Ctrl-Alt-Del which gives
> a
> > reboot. Am I wrong to suppose that loading i915 should work on a machine
> > that has two graphics devices?
>
> It should and it does: that's almost exactly the same setup I have on my
> desktop box (albeit the nvidia card is setup as the boot device).
>
> The Geforce GT 330M is apparently a mobile device.
> What is your hardware exactly ? If this is a laptop, it is possible the two
> GPUs are not independent but use something marketed as "optimus technology"
> which is currently unsupported by the DragonFly graphics stack.
>
>

Thanks, yes, you're right! It's a laptop, 15" macbookpro (MacBookPro6,2)

http://www.everymac.com/systems/apple/macbook_pro/specs/macbook-pro-core-i5-2.53-aluminum-15-mid-2010-unibody-specs.html

And yes, turns out apple have gpu switching:

http://www.techworld.com/review/laptops/apple-macbook-pro-15-inch-mid-2010-266ghz-core-i7-review-3223803/

I ran FreeBSD for a test; it gets stuck too. Xorg with linux behaves fine
though. Xorg log shows two drivers are loaded, nouveau and intel.

Until DragonFly's grahpics stack gets this feature, I wonder if there's a
way to switch off one gpu so Xorg will run?

Cheers

Peeter

--

console in UEFI boot

[karu.pruun]

Hello

I am booting DragonFly via UEFI on a macbookpro and running into
issues with console. UEFI puts console into a resolution 1680x1050,
but as soon as kernel has loaded and starts it switches to 40x25 mode
(or something similar, i.e. the default mode one sees at say CD boot),
but does not quite succeed: there's a 40x25 black-white quadrant in
the top-left corner and the rest is gray. This is compounded with
another issue: when booting is almost finished, the console starts
printing line feeds very fast although no key is pressed. In the end
the console is unusable as the prompt gets line feeds 10x a second and
nothing can be entered. Not sure the keyboard works at all. Aside from
that the system is running fine when used via ssh.

When booted using a CD the console and keyboard are fine.

Can I set a variable in loader.conf or elsewhere to switch the mode
back to 1680x1050 or something else?

Cheers

Peeter

--

loading i915 with two graphics cards

[karu.pruun]

Hello

I have an inbuilt graphics device (Intel HD Graphics: vgapci1 and it's the
boot video device) and another one attached to PCIe (NVIDIA GeForce GT
330M). When I load i915

kldload i915

the system (DragonFly 4.5) gets stuck, the screen is frozen at console. I
can only connect an external keyboard and press Ctrl-Alt-Del which gives a
reboot. Am I wrong to suppose that loading i915 should work on a machine
that has two graphics devices?

Cheers

Peeter

--

Re: console in UEFI boot

[karu.pruun]

Apparently this is due to the 32 bitness of the original operating system
that came with the machine: the UEFI spec requires that UEFI and OS bitness
match. The original OS was probably a 32 bit Windows.

Peeter

--

On Thu, Jul 7, 2016 at 11:14 PM, Bernard Mentink <bment...@gmail.com> wrote:

> In my case, I have a small Atom based micro-PC which is 64bits with a
> 32-bit UEFI ...
> I use it as a media center PC ...
>
> On Fri, Jul 8, 2016 at 6:58 AM, Samuel Paik <s...@paiks.org> wrote:
>
>> Some Windows 8 PCs (particularly Atom based tablets with 1-2 GB
>> non-expandable RAM) have 64-bit capable CPUs but come with 32-bit UEFI boot
>> roms.
>>
>> On Thu, Jul 7, 2016 at 5:43 AM, karu.pruun <karu.pr...@gmail.com> wrote:
>>
>>> Hi Bernie
>>>
>>> I don't know. Do you mean your machine is 64bit but has a 32bit EFI?
>>> This is a good overview of UEFI
>>>
>>>
>>> https://www.happyassassin.net/2014/01/25/uefi-boot-how-does-that-actually-work-then/
>>>
>>> I think the terminology is that the type of your firmware is either BIOS
>>> or UEFI, and the latter can emulate BIOS boot if needed.
>>>
>>> I should also retract my original complaint: booting now works fine on
>>> macbookpro even without interfering, ie setting 'mode 0' at loader prompt.
>>> No idea why the error is gone. All I did is correct the vfs.root.mountfrom
>>> variable so I don't need to enter it by hand in the middle of the boot
>>> sequence; and delete /etc/issue left over from the boot CD. In any case,
>>> the problem now is that i915 does not load; the machine gets stuck.
>>>
>>> Cheers
>>>
>>> Peeter
>>>
>>> --
>>>
>>>
>>> On Thu, Jul 7, 2016 at 12:03 AM, Bernard Mentink <bment...@gmail.com>
>>> wrote:
>>>
>>>> Hi Peeter,
>>>>
>>>> So is there a 32 bit EFI loader available for machines with a 32bit EFI
>>>> bios?
>>>>
>>>> Cheers,
>>>> Bernie
>>>>
>>>> On Wed, Jul 6, 2016 at 9:51 PM, karu.pruun <karu.pr...@gmail.com>
>>>> wrote:
>>>>
>>>>> Update: the console works fine if one escapes to loader prompt and
>>>>> sets mode to either 0 (80x25) or 1 (80x50), and then boots the system up.
>>>>> The original mode set by EFI is 2 (210x55). When the system has booted up,
>>>>> ttyv0 will still be smaller in the top left corner, but switching to any
>>>>> other ttyv gives a full display (105x32).
>>>>>
>>>>> So all in all: using DragonFly 4.5 and UEFI I have a multiboot machine
>>>>> (macbookpro 15) with DragonFly, MacOSX, and I also installed FreeBSD for
>>>>> testing purposes.
>>>>>
>>>>> For those interested: DragonFly can be installed manually as described
>>>>> on the manual page gpt(8). The difference is in creating partitions. For
>>>>> booting, create a 1 GB partition and format without labelling:
>>>>>
>>>>> gpt add -s 200 -t "DragonFly UFS1" adX
>>>>> newfs /dev/adXsY
>>>>>
>>>>> where X and Y are the disk and slice respectively. Other partitions
>>>>> can be labelled as usual. cpdup boot stuff to the boot partition as
>>>>> explained in gpt(8), and the rest to other partitions. When you're
>>>>> finished, mount the EFI system partition
>>>>>
>>>>> mount -t msdos /dev/adsX /mnt
>>>>>
>>>>> and copy DragonFly's EFI bootloader to it
>>>>>
>>>>> mkdir -p /mnt/EFI/dragonfly
>>>>> cp /boot/boot1.efi /mnt/EFI/dragonfly/dragonfly_x64.efi
>>>>>
>>>>> If you have an EFI boot manager like refind, you can boot DragonFly or
>>>>> any other system by choosing the respective EFI bootloader at startup. If
>>>>> you don't have a boot manager, then EFI will look for
>>>>>
>>>>> \EFI\BOOT\BOOTx64.EFI
>>>>>
>>>>> so copy DragonFly's EFI bootloader to this file,
>>>>>
>>>>> cp /boot/boot1.efi /mnt/EFI/boot/bootx64.efi
>>>>>
>>>>>
>>>>> Cheers
>>>>>
>>>>> Peeter
>>>>>
>>>>> --
>>>>>
>>>>>
>>>>>
>>>

kernel (4.7) panic at buildworld: assertion "gd->gd_spinlocks == 0" failed

[karu.pruun]

Hello

I am on v4.7.0.109.g609bcd-DEVELOPMENT and got a kernel panic after "make
buildworld". Crash dump info says:

---
Panic String: assertion "gd->gd_spinlocks == 0" failed in
dfly_schedulerclock at /usr/src/sys/kern/usched_dfly.c:757
---

I attached a kgdb output to the bug report.


Peeter

--

Re: loading i915 with two graphics cards

[karu.pruun]

Thanks for all the information and links. I had a look at the linux driver
apple-gmux.c, and tried to write a simple DragonFly kernel module that
switches off the gmux. However, I'm stuck with the attachment of the gmux
device: probing via acpi works fine, but I'm not able to allocate IO ports.
The gmux controller is at the LPC bus:

---
LPC
---
pciconf -lv:

isab0@pci0:0:31:0: . . .
 vendor = 'Intel Corporation'
 device = 'HM55 Chipset LPC Interface Controller'
 class  = bridge
 subclass = PCI-ISA
---


GMUX

devinfo -rv | grep GMUX:

unknown pnpinfo _HID=APP000B _UID=0 at handle=\_SB_.PCI0.LPCB.GMUX
---

It seems devices attached to LPC bus can be treated as isa devices:

https://lists.freebsd.org/pipermail/freebsd-hackers/2004-October/008733.html

I tried to follow the logic of similar drivers, e.g. acpi_ec, which seems
to sit at the same bus (if I understand it correctly that the handle says
it's located at the LPC bus?):

acpi_ec0 pnpinfo . . . handle=\_SB_.PCI0.LPCB.EC__

The acpi_ec probe part is done with roughly:

---
sys/dev/acpica/acpi_ec.c:
---

acpi_ec_probe(. . .)
{
. . .

ACPI_ID_PROBE(device_get_parent(dev), dev, ec_ids)

. . .
}
---

and that works fine for gmux too. Then attach() calls bus_alloc_resource():

---
sys/dev/acpica/acpi_ec.c:
---

acpi_ec_attach(. . .)
{

. . .

res = bus_alloc_resource_any(sc->ec_dev, SYS_RES_IOPORT,
>ec_data_rid, RF_ACTIVE);

. . .
}
---

But that same call returns a NULL pointer for gmux. Where have I gone
wrong?


Cheers

Peeter

--



On Thu, Aug 4, 2016 at 5:28 PM, Pedro Andrés Aranda Gutiérrez <
paag...@gmail.com> wrote:

> BTW, there was some code around in the switcher and in mackintosh lists on
> this. File was something like gmux.c
>
>
>
> On 28 Jul 2016, at 11:45, Francois Tigeot  wrote:
>
> > On Thu, Jul 28, 2016 at 10:49:26AM +0200, Pedro Andres Aranda Gutierrez
> wrote:
> >> Similar problem here with an ASUS with nVidia Optimus... So it's not Mac
> >> specific...
> >
> > Well, in a way it is.
> >
> > If I'm not mistaken, most Optimus devices have their video outputs
> directly
> > connected to the Intel GPU so it's just a matter of disabling the NVidia
> one.
> >
> > On at least some Macbook models, the video output are connected to a
> third
> > chip, and this chip must also be told to talk exclusively to the Intel
> GPU at
> > boot time first.
> >
> > --
> > Francois Tigeot
>
>

gmux and switcheroo for DragonFly

[karu.pruun]

Hello

As a followup to this thread

http://lists.dragonflybsd.org/pipermail/users/2016-July/312927.html

on why Xorg won't start on apple macbookpro, I had a look at linux's
apple-gmux driver. It is part of the switcheroo subsystem for apple laptops
with hybrid graphics. Over the last weeks I've been working on bringing the
apple-gmux driver to DragonFly.

https://github.com/peeterm007/gmux-apple

This is my first encounter with kernel and driver programming, and it's
still work in progress, but the gmux driver now un/loads fine:

---
# kldload ./gmux_apple_acpi.ko

gmux_apple_acpi0:  port 0x700-0x7fe on acpi0
gmux_apple_acpi0: Found gmux version 1.9.33 [classic]

# kldunload gmux_apple_acpi

gmux_apple_acpi0: detached
---

(Thanks to Sephe for fixing the acpi IO address range!) It still needs more
work regarding initialization and power handling.

As far as I can see, Linux's switcheroo consists of other parts that need
to be brought over to DragonFly. I wonder if anyone is working on this? I'd
like to avoid duplicating work. If there's work done on switcheroo I could
test if it works with the apple gmux driver. Alternatively, if there's
still work to be done I'd be happy to contribute if I can.

Cheers

Peeter

--

Re: Trying to install it on a Thinkpad

[karu.pruun]

Hello

I should clarify that I'm not a developer, I'm a user who's interested in
the said feature. I guess you should get in contact with the development
team before embarking on the project, perhaps try the irc channel?

If I was you I'd start by getting familiar with the installer

src/usr.sbin/installer

Also, there may have been emails on this in the past, try googling the
email list with 'gpt' etc.

If I may suggest another project to work on: get GL acceleration working
under Weston/Wayland.

In general though I suppose it'd be a good idea to update the list that
Tomohiro Kusumi mentioned:

https://www.dragonflybsd.org/docs/developer/ProjectsPage/

Cheers

Peeter

--


On Sat, Sep 10, 2016 at 5:17 PM, Satyajit Ranjeev <s...@ranjeev.in> wrote:

> I'm definitely interested. Any particular advice on getting started?
>
> On a side note now I have DragonFly BSD on my laptop.
>
> On Sep 10, 2016 10:59 AM, "karu.pruun" <karu.pr...@gmail.com> wrote:
>
>> DragonFly has UEFI boot support, but it's not integrated in the installer
>> yet. This could be a project to work on if you're interested.
>>
>> Cheers
>>
>> Peeter
>>
>> --
>>
>>
>> On Fri, Sep 9, 2016 at 11:51 PM, Satyajit Ranjeev <s...@ranjeev.in> wrote:
>>
>>> Thanks a lot, I stumbled on it as well. I'm still in the installation
>>> process. I need to use EFI boot method to get it working on NVME. But
>>> excited that the disk is loading.
>>>
>>>
>>> On Fri, Sep 9, 2016 at 10:25 PM, PeerCorps Trust Fund
>>> <i...@peercorpstrust.org> wrote:
>>> > This might be helpful:
>>> >
>>> > "We now have a NVMe driver (PCIe SSDs). It currently must be kldloaded
>>> with
>>> > nvme_load="YES" in /boot/loader.conf."
>>> >
>>> > https://www.dragonflybsd.org/release46/
>>> >
>>> > My guess would be it is not visible at install time because of that.
>>> >
>>> >
>>> >
>>> > On 09/09/2016 11:08 PM, Satyajit Ranjeev wrote:
>>> >>
>>> >> I am trying to install the latest snapshot on my work laptop. Its a
>>> >> Thinkpad T460s. Its got a NVME SSD. When I come to the disk selection
>>> >> screen the SSD is not listed. Has anyone had similar problems?
>>> >>
>>> >> satran
>>> >>
>>> >
>>>
>>
>>

Re: Trying to install it on a Thinkpad

[karu.pruun]

DragonFly has UEFI boot support, but it's not integrated in the installer
yet. This could be a project to work on if you're interested.

Cheers

Peeter

--


On Fri, Sep 9, 2016 at 11:51 PM, Satyajit Ranjeev  wrote:

> Thanks a lot, I stumbled on it as well. I'm still in the installation
> process. I need to use EFI boot method to get it working on NVME. But
> excited that the disk is loading.
>
>
> On Fri, Sep 9, 2016 at 10:25 PM, PeerCorps Trust Fund
>  wrote:
> > This might be helpful:
> >
> > "We now have a NVMe driver (PCIe SSDs). It currently must be kldloaded
> with
> > nvme_load="YES" in /boot/loader.conf."
> >
> > https://www.dragonflybsd.org/release46/
> >
> > My guess would be it is not visible at install time because of that.
> >
> >
> >
> > On 09/09/2016 11:08 PM, Satyajit Ranjeev wrote:
> >>
> >> I am trying to install the latest snapshot on my work laptop. Its a
> >> Thinkpad T460s. Its got a NVME SSD. When I come to the disk selection
> >> screen the SSD is not listed. Has anyone had similar problems?
> >>
> >> satran
> >>
> >
>

python: spyder error when connecting to ipython

[karu.pruun]

Hello

The python IDE spyder (py27-spyder) cannot connect to the ipython kernel in
order to start the ipython (py27-ipython) console. When spyder starts up,
the ipython console displays the message "Connecting to kernel. . ." and
that's it: it will loop there forever. When quitting spyder, this error
shows:

---
Exception in thread Thread-2:
Traceback (most recent call last):
  File "/usr/local/lib/python2.7/threading.py", line 801, in
__bootstrap_inner
self.run()
  File "/usr/local/lib/python2.7/threading.py", line 754, in run
self.__target(*self.__args, **self.__kwargs)
  File "/usr/local/lib/python2.7/site-packages/spyderlib/spyder.py",
line 2093, in start_open_files_server
req, dummy = self.open_files_server.accept()
  File "/usr/local/lib/python2.7/socket.py", line 206, in accept
sock, addr = self._sock.accept()
  File "/usr/local/lib/python2.7/socket.py", line 174, in _dummy
raise error(EBADF, 'Bad file descriptor')
error: [Errno 9] Bad file descriptor


[1]Done  spyder
---

py27-ipython alone runs fine on DragonFly (e.g. when started with 'ipython
notebook').

I ran python2.7 unit tests that have 'thread' in their name, i.e.

% ls -l /usr/local/lib/python2.7/test | grep thread
% python -m test.regrtest test_threading
etc

and they all pass. The test_socket also passes when DragonFly is added to
the conditional which checks the platform. The test_socketserver is
skipped.

I checked that on FreeBSD spyder and ipython run fine.

I wonder if anybody has ideas about what is going on?

Thanks,

Peeter

--

question about hammer mirror-stream

[karu.pruun]

Hello

A quick question regarding hammer mirror-streaming. The master and slave
are mounted as

# mount_null /pfs/master /home/master
# mount_null /pfs/slave /home/slave

and streaming is ongoing with

# hammer mirror-stream /home/master /home/slave

When there's a change in master, e.g.

% sudo touch /home/master/fileX

then fileX will show up in /pfs/slave but not in /home/slave. It will show
up in the latter when I stop streaming, unmount the slave and mount it
again. Is this the intended behavior?

Thanks

Peeter

--

Re: question about hammer mirror-stream

[karu.pruun]

On Mon, Dec 19, 2016 at 6:50 PM, Tim Darby <t+df...@timdarby.net> wrote:

> On Mon, Dec 19, 2016 at 6:21 AM, karu.pruun <karu.pr...@gmail.com> wrote:
>
>> then fileX will show up in /pfs/slave but not in /home/slave. It will
>> show up in the latter when I stop streaming, unmount the slave and mount it
>> again. Is this the intended behavior?
>>
>
> Yes, that's how it works. You won't see the updates to the mounted slave
> until you unmount and re-mount.
>


Thanks!

Peeter

--

Re: Trying to multiboot with FreeBSD

[karu.pruun]

On Thu, Jun 29, 2017 at 8:48 PM, Bryan C. Everly
<br...@bceassociates.com> wrote:
> On Tue, Jun 27, 2017 at 2:30 PM, karu.pruun <karu.pr...@gmail.com> wrote:
>> Can you put the line
>>
>> ---
>> i915_load="YES"
>> ---
>>
>> in /etc/rc.conf, not in /boot/loader.conf? Not sure this makes a
>> difference but might be worth a try. When you get to loader prompt and
>> give 'gop get', what is the output?
>>
>> Cheers
>>
>> Peeter
>>
>
> Peeter,
>
> Sorry it took so long for me to reply.  #dayjob
>
> Anyhow, the gop get output is:
> mode 3: 1366x768x32, stride=1376
> frame buffer: address=0, size=0
> color mask: R=00xx, G=ff00, B=00ff

Hmm, this does not look good---I'm not a specialist but I don't
understand how you can see anything at all, the framebuffer address
and size are nonexistent. This is where it seems to break down.

When you give 'gop list', do you see any other modes? Can you change
to any other mode (sth like 'mode X' with X the mode number might
work?) and try boot?

Do you see the same numbers in FreeBSD and OpenBSD? If yes then we'd
need to investigate how Free/OpenBSD fix this.

If you can connect your machine to a wired network you can ssh into
the machine. You'd need to configure you NIC in /etc/rc.conf (same as
in FreeBSD) and note that by default DragonFly ssh configuration does
not allow passwds; you need either keys or then change the
configuration to allow passwds.

Peeter

--

Re: Trying to multiboot with FreeBSD

[karu.pruun]

On Thu, Jun 29, 2017 at 11:09 PM, Bryan C. Everly
 wrote:
>> Hmm, this does not look good---I'm not a specialist but I don't
>> understand how you can see anything at all, the framebuffer address
>> and size are nonexistent. This is where it seems to break down.
>>
>> When you give 'gop list', do you see any other modes? Can you change
>> to any other mode (sth like 'mode X' with X the mode number might
>> work?) and try boot?
>>
>> Do you see the same numbers in FreeBSD and OpenBSD? If yes then we'd
>> need to investigate how Free/OpenBSD fix this.
>>
>> If you can connect your machine to a wired network you can ssh into
>> the machine. You'd need to configure you NIC in /etc/rc.conf (same as
>> in FreeBSD) and note that by default DragonFly ssh configuration does
>> not allow passwds; you need either keys or then change the
>> configuration to allow passwds.
>>
>> Peeter
>>
>> --
>
> Peeter,
>
> When I do a "gop list", I get:
>
> mode 0: 1024x768x32, stride=1024
> mode 1: 640x480x32, stride=640
> mode 2: 800x600x32, stride=800
> mode 3: 1366x768x32, stride=1376
>
> So first off, that's weird that whatever "stride" is lines up 100% to
> the horizontal resolution in modes 0,1 and 2 but not mode 3 (which
> appears to be the default).
>
> Is there a way to change the default mode by putting something in
> loader.conf do you know?


Please try this: check that you are in the video group. If not, add
yourself in the video group in /etc/group, and then ensure that

i915_load="YES"

is in /etc/rc.conf. Then boot again.

When this fails, back to modes at loader prompt. First we'd need to
figure out if there's a mode that works. At the loader prompt, try set
a mode that is not 3, I believe giving just 'mode 0' for instance to
get 1024x768. Can you then check with 'gop get' what are the values
address and size, if there's a mode where they are not zero. If yes,
then give 'boot'.

Also: when i915 loads, it should probe your screen and allocate a new
framebuffer with a suitable resolution; or 1024x768 if it fails to
find a resolution. If this is correct, then I'm not sure why putting
'i915_load="YES"' in /etc/rc.conf still resulted in corrupt screen. If
you can ssh into the machine you can see the debug msgs from i915
doing this:

% sudo kldload drm
% sudo sysctl hw.dri.debug=0x777
% sudo kldload i915

You'll get a lot of debug output in /var/log/messages, but grep for
the lines containing 'intelfb_create'; it should look like this:

---
[drm:pid1334:intelfb_create] no BIOS fb, allocating a new one
[drm:pid1334:intelfb_create] allocated 1680x1050 fb: 0x00047000, bo
0xff82e8342dc0
kernel: kms console: xpixels 1680 ypixels 1050
---


Peeter

--

Re: Trying to multiboot with FreeBSD

[karu.pruun]

Can you put the line

---
i915_load="YES"
---

in /etc/rc.conf, not in /boot/loader.conf? Not sure this makes a
difference but might be worth a try. When you get to loader prompt and
give 'gop get', what is the output?

Cheers

Peeter

--




On Tue, Jun 27, 2017 at 6:05 PM, Bryan C. Everly
 wrote:
> On Fri, Jun 23, 2017 at 12:07 PM, Bryan C. Everly
>  wrote:
>> Hello everyone,
>>
>> I'm working on a blog post about how I was able to set up a Thinkpad
>> x230 to UEFI multiboot Arch Linux, OpenBSD, FreeBSD and DragonFlyBSD.
>> I have installed DragonFlyBSD (I believe) successfully and copied
>> /boot/boot1.efi to my EFI partition and created an entry pointing to
>> it for rEFInd (just like I did for the others).  Unfortunately the
>> behavior I'm seeing is that the first stage bootloader for
>> DragonFlyBSD is seeing the UFS partition for FreeBSD and is trying to
>> boot that.
>>
>> Any ideas what I can do to get around this issue?
>>
>> Thanks,
>> Bryan
>
> All,
>
> I did the painful but logical thing and wiped the drive to start over
> and put DragonFlyBSD in front of FreeBSD in terms of partition
> sequence and was able to boot.  However, on this x230 in UEFI mode,
> I'm getting a similar problem to what I saw in FreeBSD:
>
> 1.  Boot
> 2.  See the DragonFlyBSD menu
> 3.  Boots fine until the EFI framebuffer loads
> 4.  Now everything is a bunch of diagonally striped chatter in the
> first line of the framebuffer
> 5.  It actually is booting and running, but I can't see anything
> because the display is off
>
> To fix this same problem in FreeBSD on this same laptop, I loaded the
> i915 kernel module and used the vt(4) "Graphics Mode Console" and all
> was good.
>
> When I horsed around in DragonFlyBSD to do this, I added the following
> to my /boot/loader.conf file:
>
> kern.kms_console=1
> i915_load="YES"
>
> Unfortunately, even though I see i915 and DRM loading, I still get the
> EFI Framebuffer and the same problem.
>
> Any ideas?  I'd really like to get into DragonFlyBSD and figured what
> could be safer than an x230 but I think the hangup is that I'm UEFI
> booting and I'm guessing most of the other x230 users of DragonFlyBSD
> are not.
>
> I'd appreciate any help guys as I want to jump in and see what I can
> do to help this project out!

Re: Packages for 4.9

[karu.pruun]

On Thu, Jun 29, 2017 at 1:00 PM, Pierre Abbat
 wrote:
> On Thursday, June 29, 2017 2:08:47 AM EDT Pierre Abbat wrote:
>> I got "illegal option -- f". I'm letting it run, but then I'll try the
>> command again with -f.
>
> It still says "illegal option". What is the correct option?

The option "-f" is at the end:

% pkg upgrade -f


Peeter

--

Re: DragonflyBSD 4.8 on Mac Mini 2010 - fails to install

[karu.pruun]

Hello

I may have ran into similar issue with a mac laptop from 2009. By way
of testing, can you check if dragonfly 3.6.2 boots without acpi --- at
boot, get to prompt and give

> lunset acpi_load
> set hint.acpi.0.disabled=1
> boot -v

If it boots, then the behavior is similar. There are possibly two
problems that need to be resolved so that 4.8 can boot on mac. The
first one has a patch, the second one needs more debugging.

Cheers

Peeter

--



On Sun, Apr 23, 2017 at 11:47 PM, ron georgia  wrote:
> I tried to install DragonflyBSD 4.8 on my Mac Mini 2010 model. It hangs
> after “printing” this message:
>
> ACPI: 4 ACPI AML tables successfully acquired and loaded.
>
> I was able to install both OpenBSD and FreeBSD current (not at the same
> time). OpenBSD worked out of the box and I needed a little Google foo to get
> FreeBSD to use the bge interface. I did try all the “boot without” options
> and ended up in the same place. Are there any boot flags I can set or some
> black magic incantation to get Dragonfly on my Mac Mini?
>
> Ron
>
>
>
>

Re: random freezes, xterm not displaying everything correctly what to do?

[karu.pruun]

I don't know about xterm, I'm using xfce4-terminal with Droid Sans
Mono Bold. I would try force installing the xterm and the said fonts
again

> sudo pkg upgrade -f xorg-fonts xterm

note the "-f" is important. If you still don't see the fonts, then
start investigating what fonts the xterm uses and that the font paths
in xorg.conf are correct.

Peeter

--



On Thu, May 4, 2017 at 10:39 PM, Antonio Olivares
 wrote:
> One problem appears to be solved.  However one remains.  The Chromium
> browser works and I dont get the crashes that I did before.  However,
> I have the other problem with xterm.  I cannot see the prompt.  The
> display does not show the characters or fonts correctly.  I have
> installed some fonts, but xorg-fonts metapackage is installed.  I
> wonder how can I get xterm to display the letters/chars correctly.
>
> Best Regards,
>
>
> Antonio
>
> On Wed, May 3, 2017 at 4:00 PM, Michael Neumann  wrote:
>>
>>
>> On 05/03/17 21:31, Antonio Olivares wrote:
>>> Dear folks,
>>>
>>> I have successfully installed DragonFlyBSD4.8 on an ASUS eight core
>>> machine with AMD Graphics card, nvidia hardware still use nv driver
>>> and screen output do not help much.  When I open firefox, and I see a
>>> youtube video, or try to check mail, I get random freezes and sceen
>>> blacks out several times.   As I was going to send this message , the
>>> machine has frozen out two times on me :(
>>
>> Can you try "chrome" instead? I see those freezes as well on my machine
>> (Radeon too), but with the packages from about 2 weeks ago, chrome did
>> work like a charm (while I had some glitches with firefox). With the
>> latest packages, chrome don't even start up.
>>
>> Regards,
>>
>>   Michael
>>
>>>
>>> Copyright (c) 2003-2017 The DragonFly Project.
>>>
>>> Copyright (c) 1992-2003 The FreeBSD Project.
>>>
>>> Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
>>>
>>> The Regents of the University of California. All rights reserved.
>>>
>>> DragonFly v4.8.0-RELEASE #4: Sun Mar 26 20:55:22 EDT 2017
>>>
>>> r...@www.shiningsilence.com:/usr/obj/home/justin/release/4_8/sys/X86_64_GENERIC
>>>
>>> i8254 clock: 1193193 Hz
>>>
>>> TSC invariant clock: 3511723754 Hz
>>>
>>> CPU: AMD FX(tm)-8320 Eight-Core Processor (3511.72-MHz K8-class CPU)
>>>
>>> Origin = "AuthenticAMD" Id = 0x600f20 Stepping = 0
>>>
>>> Features=0x178bfbff
>>>
>>> Features2=0x3698320b
>>>
>>> AMD Features=0x2e500800
>>>
>>> AMD 
>>> Features2=0x1ebbfff
>>>
>>> Structured Extended Features=0x8
>>>
>>> MONITOR/MWAIT Features=0x2
>>>
>>> real memory = 8482402304 (8089 MB)
>>>
>>> avail memory = 7869161472 (7504 MB)
>>>
>>> LAPIC: AMD LINTEN on
>>>
>>> lapic: divisor index 0, frequency 100334578 Hz
>>>
>>> Compute unit iDS:
>>>
>>> 0-0;
>>>
>>> srat_probe: can't locate SRAT
>>>
>>> Initialize MI interrupts
>>>
>>> wdog: In-kernel automatic watchdog reset enabled
>>>
>>> kbd1 at kbdmux0
>>>
>>> md0: Malloc disk
>>>
>>> interrupt uses mplock: swi_taskq
>>>
>>> ACPI: RSDP 0xBA865000 24 (v02 ALASKA)
>>>
>>> ACPI: XSDT 0xBA865070 5C (v01 ALASKA A M I 01072009 AMI 
>>> 00010013)
>>>
>>> ACPI: FACP 0xBA86C200 00010C (v05 ALASKA A M I 01072009 AMI 
>>> 00010013)
>>>
>>> ACPI BIOS Warning (bug): Optional FADT field Pm2ControlBlock has valid
>>> Length but zero Address: 0x/0x1 (20170303/tbfadt-801)
>>>
>>> ACPI: DSDT 0xBA865168 007095 (v02 ALASKA A M I  INTL 
>>> 20051117)
>>>
>>> ACPI: FACS 0xBB953F80 40
>>>
>>> ACPI: APIC 0xBA86C310 9E (v03 ALASKA A M I 01072009 AMI 
>>> 00010013)
>>>
>>> ACPI: FPDT 0xBA86C3B0 44 (v01 ALASKA A M I 01072009 AMI 
>>> 00010013)
>>>
>>> ACPI: MCFG 0xBA86C3F8 3C (v01 ALASKA A M I 01072009 MSFT 
>>> 00010013)
>>>
>>> ACPI: HPET 0xBA86C438 38 (v01 ALASKA A M I 01072009 AMI 
>>> 0005)
>>>
>>> ACPI: SSDT 0xBA86C500 001714 (v01 AMD POWERNOW 0001 AMD 
>>> 0001)
>>>
>>> ACPI: BGRT 0xBA86C4C8 38 (v00 ALASKA A M I 01072009 AMI 
>>> 00010013)
>>>
>>> cryptosoft0:  on motherboard
>>>
>>> aesni0:  on motherboard
>>>
>>> padlock0: No ACE support.
>>>
>>> rdrand0: No RdRand support.
>>>
>>> acpi0:  on motherboard
>>>
>>> ACPI: Executed 2 blocks of module-level executable AML code
>>>
>>> ACPI: 2 ACPI AML tables successfully acquired and loaded
>>>
>>> ACPI FADT: SCI testing interrupt mode ...
>>>
>>> ACPI FADT: SCI select level/low
>>>
>>> acpi0: Power Button (fixed)
>>>
>>> acpi_timer0 on acpi0
>>>
>>> acpi_hpet0:  iomem 0xfed0-0xfed003ff on 
>>> acpi0
>>>
>>> acpi_hpet0: 

Re: random freezes, xterm not displaying everything correctly what to do?

[karu.pruun]

It seems you have a Radeon card and use the radeon kms driver. I can't
see any NVidia devices in your dmesg. Maybe it helps narrow down the
issue when you try running weston/wayland instead of your DE/Xorg. If
weston does not get the freezes then chances are Xorg is at fault.
Quick instructions are here

http://lists.dragonflybsd.org/pipermail/users/2016-May/249620.html

You can add weston with pkg, also add xwayland to run Xorg apps:

% sudo pkg install weston xwayland

You should "kldload radeonkms", and add yourself to the "video" group;
also create the group "weston-launch" and add yourself to it. And
enable/start udevd in rc.conf:

---
udevd_enable="YES"
---

Then launch weston on the console

% weston-launch -- --use-pixman --modules=xwayland.so

Firefox runs under xwayland, but chrome does not.

Cheers

Peeter

--



On Wed, May 3, 2017 at 10:31 PM, Antonio Olivares
 wrote:
> Dear folks,
>
> I have successfully installed DragonFlyBSD4.8 on an ASUS eight core
> machine with AMD Graphics card, nvidia hardware still use nv driver
> and screen output do not help much.  When I open firefox, and I see a
> youtube video, or try to check mail, I get random freezes and sceen
> blacks out several times.   As I was going to send this message , the
> machine has frozen out two times on me :(
>
> Copyright (c) 2003-2017 The DragonFly Project.
>
> Copyright (c) 1992-2003 The FreeBSD Project.
>
> Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
>
> The Regents of the University of California. All rights reserved.
>
> DragonFly v4.8.0-RELEASE #4: Sun Mar 26 20:55:22 EDT 2017
>
> r...@www.shiningsilence.com:/usr/obj/home/justin/release/4_8/sys/X86_64_GENERIC
>
> i8254 clock: 1193193 Hz
>
> TSC invariant clock: 3511723754 Hz
>
> CPU: AMD FX(tm)-8320 Eight-Core Processor (3511.72-MHz K8-class CPU)
>
> Origin = "AuthenticAMD" Id = 0x600f20 Stepping = 0
>
> Features=0x178bfbff
>
> Features2=0x3698320b
>
> AMD Features=0x2e500800
>
> AMD 
> Features2=0x1ebbfff
>
> Structured Extended Features=0x8
>
> MONITOR/MWAIT Features=0x2
>
> real memory = 8482402304 (8089 MB)
>
> avail memory = 7869161472 (7504 MB)
>
> LAPIC: AMD LINTEN on
>
> lapic: divisor index 0, frequency 100334578 Hz
>
> Compute unit iDS:
>
> 0-0;
>
> srat_probe: can't locate SRAT
>
> Initialize MI interrupts
>
> wdog: In-kernel automatic watchdog reset enabled
>
> kbd1 at kbdmux0
>
> md0: Malloc disk
>
> interrupt uses mplock: swi_taskq
>
> ACPI: RSDP 0xBA865000 24 (v02 ALASKA)
>
> ACPI: XSDT 0xBA865070 5C (v01 ALASKA A M I 01072009 AMI 00010013)
>
> ACPI: FACP 0xBA86C200 00010C (v05 ALASKA A M I 01072009 AMI 00010013)
>
> ACPI BIOS Warning (bug): Optional FADT field Pm2ControlBlock has valid
> Length but zero Address: 0x/0x1 (20170303/tbfadt-801)
>
> ACPI: DSDT 0xBA865168 007095 (v02 ALASKA A M I  INTL 20051117)
>
> ACPI: FACS 0xBB953F80 40
>
> ACPI: APIC 0xBA86C310 9E (v03 ALASKA A M I 01072009 AMI 00010013)
>
> ACPI: FPDT 0xBA86C3B0 44 (v01 ALASKA A M I 01072009 AMI 00010013)
>
> ACPI: MCFG 0xBA86C3F8 3C (v01 ALASKA A M I 01072009 MSFT 00010013)
>
> ACPI: HPET 0xBA86C438 38 (v01 ALASKA A M I 01072009 AMI 0005)
>
> ACPI: SSDT 0xBA86C500 001714 (v01 AMD POWERNOW 0001 AMD 0001)
>
> ACPI: BGRT 0xBA86C4C8 38 (v00 ALASKA A M I 01072009 AMI 00010013)
>
> cryptosoft0:  on motherboard
>
> aesni0:  on motherboard
>
> padlock0: No ACE support.
>
> rdrand0: No RdRand support.
>
> acpi0:  on motherboard
>
> ACPI: Executed 2 blocks of module-level executable AML code
>
> ACPI: 2 ACPI AML tables successfully acquired and loaded
>
> ACPI FADT: SCI testing interrupt mode ...
>
> ACPI FADT: SCI select level/low
>
> acpi0: Power Button (fixed)
>
> acpi_timer0 on acpi0
>
> acpi_hpet0:  iomem 0xfed0-0xfed003ff on acpi0
>
> acpi_hpet0: frequency 14318180
>
> acpi_ec0:  port 0x66,0x62 on acpi0
>
> pcib0:  port 0xcf8-0xcff on acpi0
>
> pci0:  on pcib0
>
> pcib1:  irq 52 at device 2.0 on pci0
>
> pci1:  on pcib1
>
> vgapci0:  port 0xe000-0xe0ff mem
> 0xfea0-0xfea3,0xc000-0xcfff irq 24 at device 0.0 on
> pci1
>
> vgapci0: Boot video device
>
> hdac0:  mem 0xfea6-0xfea63fff irq 25
> at device 0.1 on pci1
>
> hdac0: link ctrl 0x2800
>
> hdac0: disable nosnoop
>
> hdac0: hdac_get_capabilities: Invalid corb size (0)
>
> device_probe_and_attach: hdac0 attach returned 6
>
> pcib2:  irq 52 at device 4.0 on pci0
>
> pci2:  on pcib2
>
> re0:  port 0xd000-0xd0ff mem
> 0xd000-0xd0003fff,0xd0004000-0xd0004fff 

Re: DragonFly 5.0 released!

[karu.pruun]

DragonFly 5 supports hybrid graphics in the form of 'basic switching'
as explained here

https://wiki.archlinux.org/index.php/hybrid_graphics

where you have a chip, a mux, that switches between two GPUs. The
drivers vga_switcheroo and apple_gmux take care of this; it has been
tested and works on some models of Apple laptops that have dual
graphics, Intel/NVidia or Intel/Radeon. Note that since DragonFly does
not have a drm driver for NVidia chips, neither the proprietry nor
(yet) the open source one (nouveau), it switches to the Intel GPU by
default. The latter is powerful enough if one does not use graphics
intensive apps; I'm using it for daily work on my laptop with an
1920x1080 external monitor.

DragonFly does not (yet) support the second form of switching,
'dynamic switching' (Optimus for NVIDIA, AMD Dynamic Switchable
Graphics for Radeon). This would require PRIME (in addition to nouveau
for NVidia) to be ported to DragonFly.

https://wiki.archlinux.org/index.php/PRIME

I think it would be an interesting project for the future.

Best

Peeter

--



On Tue, Oct 17, 2017 at 10:52 AM, Bret Busby  wrote:
> On 17/10/2017, Predrag Punosevac  wrote:
>> On Mon, Oct 16, 2017 at 4:09 PM, Bret Busby 
>> wrote:
>>> On 16/10/2017, Justin Sherrill  wrote:
 DragonFly 5.0 is released - here's the release page:

 https://www.dragonflybsd.org/release50/

>>>
>>> Does it support the nVidia Optimus architecture?
>>>
>>
>> Only nVidia supports its own product with binary blob drivers. Somebody
>> have mentioned Nouveau driver in one of the posts. That is beating a
>> dead horse. Even Red Hat removed Nouveau driver from 7.4 release. If you
>> want to use nVidia gear you need to use their proprietary binary blobs
>> and to stick to a list of approved OSs. I have 7 large GPU servers (4+
>> Titan Xp cards per server 32 cores and 256 GB of RAM) used for Deep
>> learning and image recognition.  They all run Red Hat like a champ.
>>
>
> "
> :~$ lshw
> WARNING: you should run this program as super-user.
> bret-aspire-v3-772-ubuntumate
> description: Computer
> width: 64 bits
> capabilities: vsyscall32
>   *-core
>description: Motherboard
>physical id: 0
>  *-memory
>   description: System memory
>   physical id: 0
>   size: 31GiB
>  *-cpu
>   product: Intel(R) Core(TM) i7-4702MQ CPU @ 2.20GHz
>   vendor: Intel Corp.
>   physical id: 1
>   bus info: cpu@0
>   size: 3096MHz
>   capacity: 3200MHz
>   width: 64 bits
>   capabilities: fpu fpu_exception wp vme de pse tsc msr pae
> mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr
> sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp x86-64 constant_tsc
> arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf
> eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sdbg
> fma cx16 xtpr pdcm pcid sse4_1 sse4_2 movbe popcnt tsc_deadline_timer
> aes xsave avx f16c rdrand lahf_lm abm epb tpr_shadow vnmi flexpriority
> ept vpid fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid xsaveopt
> dtherm ida arat pln pts cpufreq
>  *-pci
>   description: Host bridge
>   product: Xeon E3-1200 v3/4th Gen Core Processor DRAM Controller
>   vendor: Intel Corporation
>   physical id: 100
>   bus info: pci@:00:00.0
>   version: 06
>   width: 32 bits
>   clock: 33MHz
> *-pci:0
>  description: PCI bridge
>  product: Xeon E3-1200 v3/4th Gen Core Processor PCI
> Express x16 Controller
>  vendor: Intel Corporation
>  physical id: 1
>  bus info: pci@:00:01.0
>  version: 06
>  width: 32 bits
>  clock: 33MHz
>  capabilities: pci normal_decode bus_master cap_list
>  configuration: driver=pcieport
>  resources: irq:24 ioport:4000(size=4096)
> memory:d200-d2ff ioport:a000(size=536870912)
>*-display
> description: 3D controller
> product: GK107M [GeForce GT 750M]
> vendor: NVIDIA Corporation
> physical id: 0
> bus info: pci@:01:00.0
> version: a1
> width: 64 bits
> clock: 33MHz
> capabilities: bus_master cap_list rom
> configuration: driver=nouveau latency=0
> resources: irq:32 memory:d200-d2ff
> memory:a000-afff memory:b000-b1ff
> ioport:4000(size=128) memory:b200-b207
> "
>
> Note the display designation and the driver.
>
> "
> :~$ lsb_release -a
> No LSB modules are available.
> Distributor ID: Ubuntu
> Description:Ubuntu 16.04.3 LTS
> Release:16.04
> Codename:   xenial
> "
>
>
> --

Re: TFES RFIS-34 Message repeats

[karu.pruun]

I remember now that I got the same repetitive behavior at some point
couple of months ago on my laptop (macbookpro). When I updated the
kernel (I'm on master), the messages disappeared; it only shows once
at boot. I'm running a master at commit 05e0d (22 Aug, see gitweb or
'git log' in /usr/src, ), and it's stable. You can try boot a snapshot
usb img from 22 Aug and see if you still get the messages. To be on
the safe side, be sure that the kernel is not before 05e0d, there was
a brief instability around that date that crashed my machine, but
05e0d is stable again.

Cheers

Peeter

--


On Wed, Aug 23, 2017 at 11:21 PM, Rod Person  wrote:
> Ok. Just got home and can confirm, with a CD in the drive the messages
> stop.
>
>
> On Wed, 23 Aug 2017 13:16:12 -0700
> Matthew Dillon  wrote:
>
>> Ok.  It wasn't quite the intended behavior but I don't think it hurts
>> anything.  The system is probing the drive every so often looking for
>> media.  The AHCI driver got a little more verbose on reporting error
>> conditions, but this particular error condition is actually expected
>> when there is no CD/DVD in the drive.  I'll look into 'quieting' the
>> message for the CD/DVD probe when I have a bit of free time.
>>
>> -Matt
>
>
>
> --
> Rod
>
> http://www.rodperson.com
>
> For we are both ready to believe the thing we want to believe, and also
> hope other people feel what we feel ourselves.
>
>   Julius Caesar
> The Civil War, book II, 27
>  Expedition of Curio to Africa
>

Re: Network improvement 4.8 -> 5.0 for short-lived HTTP/1.1 workload.

[karu.pruun]

Nice! Thanks for the overview.

Cheers

Peeter

--



On Tue, Oct 17, 2017 at 5:57 PM, Sepherosa Ziehau  wrote:
> In this release cycle, several items are committed to improve
> performance and reduce/stablize latency.
>
> 30K concurrent connections, 1 request/connection, 1KB web object.
> Server have 24 HT.
>
> Baseline (32 nginx workers w/ 16 netisrs):
> performance 215907.25tps, latency-avg 33.11ms, latency-stdev 41.76ms,
> latency-99% 192.36ms.
>
> The performance for 16 nginx workers is too low to be used as baseline
> (16 nginx workers w/ 16 netisrs):
> performance 191920.81tps, latency-avg 32.04ms, latency-stdev 25.15ms,
> latency-99% 101.37ms.
>
> ===
>
> Make # of netisr tunable.
>
> If # of netisrs is set to ncpus, this allows two optmized settings in nginx:
> 1) Make # of nginx workers same as # of netisrs.
> 2) CPU-bind nginx workers.
>
> 24 nginx workers w/ 24 netisrs:
> performance 212556.02tps, latency-avg 56.18ms, latency-stdev 7.90ms,
> latency-99% 70.31ms.
>
> 24 nginx workers w/ 24 netisrs, cpu-bound nginx workers:
> performance 210658.80tps, latency-avg 58.01ms, latency-stdev 5.20ms,
> latency-99% 68.73ms.
>
> As you can see, performance dropped a bit.  Though average latency is
> increased, latency is significantly stablized.
>
> ===
>
> Limit the # of acceptable sockets returned by kevent(2).
>
> 24 nginx workers w/ 24 netisrs, cpu-bound nginx workers:
> performance 217599.01tps, latency-avg 32.00ms, latency-stdev 2.35ms,
> latency-99% 35.59ms.
>
> Compared w/ baseline, performance improved a bit and latency is
> reduced a bit.  However, latency is significantly stabled.
>
> ===
>
> Summary of the comparison of different web object size:
>
> 1KB web object
>
>  | perf (tps) | lat-avg | lat-stdev | lat-99%
> -++-+---+-
> baseline |  215907.25 |   33.11 | 41.76 |  192.36
> -++-+---+-
> netisr_ncpus |  210658.80 |   58.01 |  5.20 |   68.73
> -++-+---+-
> kevent.data  |  217599.01 |   32.00 |  2.35 |   35.59
>
> 8KB web object
>
>  | perf (tps) | lat-avg | lat-stdev | lat-99%
> -++-+---+-
> baseline |  182719.03 |   42.62 | 58.70 |  250.51
> -++-+---+-
> netisr_ncpus |  181201.11 |   68.78 |  6.43 |   80.68
> -++-+---+-
> kevent.data  |  186324.41 |   37.41 |  4.81 |   48.69
>
> 16KB web object
>
>  | perf (tps) | lat-avg | lat-stdev | lat-99%
> -++-+---+-
> baseline |  138625.67 |   72.01 | 65.78 |  304.78
> -++-+---+-
> netisr_ncpus |  138323.40 |   93.61 | 16.30 |  137.12
> -++-+---+-
> kevent.data  |  138778.11 |   60.90 | 11.80 |   92.07
>
> So performance is improved a bit, latency-avg is reduced by 3%~15%,
> latency-stdev is reduced by 82%~94%, latency-99% is reduced by
> 69%~81%!
>
> +++
>
> And as a bonus, forwarding performance is also improved!  We now can
> do 13.2Mpps (dual direction forwarding, output packets count) w/
> fastforwarding, and 11Mpps w/ normal forwarding.
>
> Thanks,
> sephe
>
> --
> Tomorrow Will Never Die

Re: HAMMER2 stability

[karu.pruun]

Hello

Good to hear you got DragonFly running. I suppose it was HAMMER2 that
caused problems, not HAMMER---the latter is stable.

Apart from HAMMER2, could you possibly tell more about what are you
missing (list or privately)? This may help improve things in the
future.

Cheers

Peeter

--



On Mon, Oct 23, 2017 at 11:02 AM, Pedro Andres Aranda Gutierrez
 wrote:
> Hi folks,
>
> After I got DFly running on my old Optimus laptop, I got an exception from
> the hammer module and wasn't able to recover it. I gave  FreeBSD 11.1 a try
> and have almost everything I need running much more out of the box than with
> DFly. You are getting better and I will put my notes on getting it running
> on Optimus in the WIKI, but it's not yet where I need it to be.
>
> Sorry, /PA
>
> El 23 oct. 2017 1:59, "Antonio Olivares"  escribió:
>>
>>
>>
>> On Sunday, October 22, 2017, Carsten Mattner 
>> wrote:
>>>
>>> Is HAMMER2 considered experimental and not for production data?
>>
>>
>> That is what the 5.0 release page states:
>>
>>  https://www.dragonflybsd.org/release50/
>>
>>> Do we have debug_fs file restoration tools, in case things goes south?
>>
>>
>> This is being worked on.  A new iso with the restoration tools is in the
>> works, I have to find out where I read this to point it out.
>>
>> Best Regards,
>>
>>
>> Antonio
>>
>>
>>

Re: X2APIC support is in master; further testing is needed

[karu.pruun]

Hello

Shows up fine on my i7 skylake box. The i5 arrandale is probably too
old, can't see it in the list of features.

Cheers

Peeter

--

On Sat, Jun 9, 2018 at 8:07 PM Tim Darby  wrote:
>
> I haven't seen any issues with this.
>
> Tim
>
>
> On Fri, Jun 8, 2018 at 7:58 AM Sepherosa Ziehau  wrote:
>>
>> Hi all,
>>
>> X2APIC support is in master now.  Please help testing it by adding the
>> following line to /boot/loader.conf or set it at boot prompt:
>> hw.x2apic_enable=1
>>
>> You can check whether it is on or not by:
>> sysctl hw.x2apic_enable
>>
>> X2APIC mode will be enabled only if the following command has output
>> (should be available on modern Intel CPUs):
>> dmesg | grep x2APIC
>>
>> It should be safe to test.  I have run with it on my testing machines.
>>
>> I'd like to thank ivadasz@ for his initial work.
>>
>> Thanks,
>> sephe
>>
>> --
>> Tomorrow Will Never Die

Re: KDE 5

[karu.pruun]

On Wed, Jun 20, 2018 at 9:08 PM John Marino  wrote:
>
> On 5/28/2018 08:35, Gerald Henriksen wrote:
> > On Thu, 24 May 2018 12:12:37 -0500, you wrote:
> >
> >>
> >> Just about every KDE5 except the top-level x11/kde5 metaport has been in
> >> dports for a few weeks already.
> >
> > Thanks for the reply and your work on dports
> >
>
> To follow up, the full KDE5 is blocked by security/plasma5-kscreenlocker
> which has a hard dependency on the wayland QT5 support which currently
> DragonFly does not have.
>
> Our wayland port appeared before the FreeBSD ports version and has not
> been updated.  Somebody needs to get FreeBSD and DragonFly wayland
> support together so we can remove our custom port and get all this
> wayland stuff working.
>
> I'm not going to take this is, so  volunteers?


The build log of security/plasma5-kscreenlocker looks as if it fails
because x11/kf5-kwayland fails. The latter fails because of

"error: undefined reference to 'wl_event_loop_get_fd'"

i.e. our wayland dport does not implement the function
'wl_event_loop_get_fd'. It's not easy to implement it using libevent2
which we are using to replace epoll in wayland's event loop.

There's a quick and dirty hack to fix this if make
wl_event_loop_get_fd a dummy in our wayland dport, that would unblock
the build. I don't think anyone will be using kde5 with wayland anyway
at the moment. As for weston, our only other official consumer of
wayland, it does not use wl_event_loop_get_fd. So the hack does not
seem to cause much trouble in short term. I can make a deltaports pull
request if you think it's a good idea.

In the longer term we should get our wayland fixed and we can discuss
what's the best way to do it.

Cheers

Peeter

--

Re: X on dell D630 / intel GM965 only shows black screen

[karu.pruun]

Hello

When xorg starts it first loads the respective drm driver and then
starts xorg. Doing these steps manually show the sequence fails at
setting up the drm/i915 driver.

Now when you have two gpus, then the situation may get problematic. If
there are two different devices, as in intel and nvidia, then one
problem I've seen is that although the driver attaches correctly, we
generally do not have the infrastructure to deal with this situation
100% correctly. The intel driver is not aware of the other gpu, and it
will grab the output, leaving the screen frozen because the actual
output to the screen is rendered from the buffers of the first device,
the nvidia chip. In my case the solution was to flip a switch (on a
gmux device) so output comes from the intel device. The nvidia chip
was left on; it does not matter. From the point of power saving it
might be better to turn it off though.

However, in your case the system seems to have found two intel
devices. I'm not sure how the drm/i915 is supposed to handle this.

By way of debugging, if you get to the db> prompt, then give 'call
dumpsys' and then 'reboot'. The system will save the core dump at next
boot which you can investigate. Can you do this with the latest
snapshot (it's stable and you can install with the snapshot .img) to
be sure the issue is still there? It's possibly still there, but
better to have the latest snapshot in terms of debugging.

Cheers

Peeter

--






On Fri, Mar 16, 2018 at 11:58 AM, Pedro Andres Aranda Gutierrez
<paag...@gmail.com> wrote:
> Hi
>
> mileage may vary but I was quite successful shutting off the Nvidia card
> with my FreeBSD.
>
> /PA
>
> On 15 March 2018 at 10:44, karu.pruun <karu.pr...@gmail.com> wrote:
>>
>> Hello
>>
>> Can you try booting up and then manually loading the i915 driver as
>> root or with sudo
>>
>> % sudo kldload i915
>>
>> does this load fine? You should see informational output in
>> /var/log/messages. What does
>>
>> % pciconf -lv
>>
>> say about the vga devices, see the blocks starting with 'vga'. For
>> instance, I have
>>
>> ---
>> vgapci1@pci0:0:2:0:class=0x03 card=0x chip=0x00468086
>> rev=0x18 hdr=0x00
>> vendor = 'Intel Corporation'
>> device = 'Core Processor Integrated Graphics Controller'
>> class  = display
>> subclass   = VGA
>> vgapci0@pci0:1:0:0:class=0x03 card=0x00c7106b chip=0x0a2910de
>> rev=0xa2 hdr=0x00
>> vendor = 'NVIDIA Corporation'
>> device = 'GT216M [GeForce GT 330M]'
>> class  = display
>> subclass   = VGA
>> ---
>>
>> Can you try manually create a simple xorg.conf and set the intel
>> driver as the only one with the pci address you get from pciconf (eg
>> PCI:0:2:0)? In my xorg log I'm not seeing any other drivers loaded
>> except intel and it works fine.
>>
>> Cheers
>>
>> Peeter
>>
>> --
>>
>>
>>
>>
>>
>>
>> On Wed, Mar 14, 2018 at 6:32 PM,  <spuelmasch...@email.de> wrote:
>> > MoinMoin,
>> > having problems to get X running on a dell D630 / GM965 graphics laptop.
>> > It just stays black and it's not killable but by switching it off.
>> > I've tried intel and vesa drivers by specifying them explicitly
>> > in /usr/local/etc/X11/xorg.conf.d
>> > Vesa works, but is slooow.
>> > Intel just hangs with  BusID "PCI:0:2:0" as well as with "PCI:0:2:1".
>> > In Linux and OpenBSD modesetting drivers just work without any
>> > modifications to the default xorg configuration.
>> > Does anybody has a hint what to do ?
>> >
>> > Thanks, Tschuess,
>> >  Sascha
>> >
>> > --> Xorg.0.log
>> > [ 13790.928]
>> > X.Org X Server 1.18.4
>> > Release Date: 2016-07-19
>> > [ 13790.928] X Protocol Version 11, Revision 0
>> > [ 13790.928] Build Operating System: DragonFly 5.0-SYNTH x86_64
>> > [ 13790.928] Current Operating System: DragonFly  5.0-RELEASE DragonFly
>> > v5.0.2-RELEASE #4: Su x86_64
>> > [ 13790.931] Build Date: 31 December 2017  01:53:00AM
>> > [ 13790.931]
>> > [ 13790.931] Current version of pixman: 0.34.0
>> > [ 13790.931]Before reporting problems, check http://wiki.x.org
>> > to make sure that you have the latest version.
>> > [ 13790.931] Markers: (--) probed, (**) from config file, (==) default
>> > setting,
>> > (++) from command line, (!!) notice, (II) informational,
>> > (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
>

Re: X on dell D630 / intel GM965 only shows black screen

[karu.pruun]

Hello

Can you try booting up and then manually loading the i915 driver as
root or with sudo

% sudo kldload i915

does this load fine? You should see informational output in
/var/log/messages. What does

% pciconf -lv

say about the vga devices, see the blocks starting with 'vga'. For
instance, I have

---
vgapci1@pci0:0:2:0:class=0x03 card=0x chip=0x00468086
rev=0x18 hdr=0x00
vendor = 'Intel Corporation'
device = 'Core Processor Integrated Graphics Controller'
class  = display
subclass   = VGA
vgapci0@pci0:1:0:0:class=0x03 card=0x00c7106b chip=0x0a2910de
rev=0xa2 hdr=0x00
vendor = 'NVIDIA Corporation'
device = 'GT216M [GeForce GT 330M]'
class  = display
subclass   = VGA
---

Can you try manually create a simple xorg.conf and set the intel
driver as the only one with the pci address you get from pciconf (eg
PCI:0:2:0)? In my xorg log I'm not seeing any other drivers loaded
except intel and it works fine.

Cheers

Peeter

--






On Wed, Mar 14, 2018 at 6:32 PM,   wrote:
> MoinMoin,
> having problems to get X running on a dell D630 / GM965 graphics laptop.
> It just stays black and it's not killable but by switching it off.
> I've tried intel and vesa drivers by specifying them explicitly
> in /usr/local/etc/X11/xorg.conf.d
> Vesa works, but is slooow.
> Intel just hangs with  BusID "PCI:0:2:0" as well as with "PCI:0:2:1".
> In Linux and OpenBSD modesetting drivers just work without any
> modifications to the default xorg configuration.
> Does anybody has a hint what to do ?
>
> Thanks, Tschuess,
>  Sascha
>
> --> Xorg.0.log
> [ 13790.928]
> X.Org X Server 1.18.4
> Release Date: 2016-07-19
> [ 13790.928] X Protocol Version 11, Revision 0
> [ 13790.928] Build Operating System: DragonFly 5.0-SYNTH x86_64
> [ 13790.928] Current Operating System: DragonFly  5.0-RELEASE DragonFly 
> v5.0.2-RELEASE #4: Su x86_64
> [ 13790.931] Build Date: 31 December 2017  01:53:00AM
> [ 13790.931]
> [ 13790.931] Current version of pixman: 0.34.0
> [ 13790.931]Before reporting problems, check http://wiki.x.org
> to make sure that you have the latest version.
> [ 13790.931] Markers: (--) probed, (**) from config file, (==) default 
> setting,
> (++) from command line, (!!) notice, (II) informational,
> (WW) warning, (EE) error, (NI) not implemented, (??) unknown.
> [ 13790.932] (==) Log file: "/var/log/Xorg.0.log", Time: Tue Mar 13 17:42:35 
> 2018
> [ 13790.932] (II) Loader magic: 0x615460
> [ 13790.932] (II) Module ABI versions:
> [ 13790.932]X.Org ANSI C Emulation: 0.4
> [ 13790.932]X.Org Video Driver: 20.0
> [ 13790.932]X.Org XInput driver : 22.1
> [ 13790.932]X.Org Server Extension : 9.0
> [ 13790.932] (--) PCI:*(0:0:2:0) 8086:2a02:1028:01f9 rev 12, Mem @ 
> 0xf6e0/1048576, 0xe000/268435456, I/O @ 0xeff0/8, BIOS @ 
> 0x/65536
> [ 13790.932] (--) PCI: (0:0:2:1) 8086:2a03:1028:01f9 rev 12, Mem @ 
> 0xf6f0/1048576
> [ 13790.932] (==) Using default built-in configuration (30 lines)
> [ 13790.932] (==) --- Start of built-in configuration ---
> [ 13790.932]Section "Device"
> [ 13790.932]Identifier  "Builtin Default intel Device 0"
> [ 13790.932]Driver  "intel"
> [ 13790.932]EndSection
> [ 13790.932]Section "Screen"
> [ 13790.932]Identifier  "Builtin Default intel Screen 0"
> [ 13790.932]Device  "Builtin Default intel Device 0"
> [ 13790.932]EndSection
> [ 13790.932]Section "Device"
> [ 13790.932]Identifier  "Builtin Default modesetting Device 0"
> [ 13790.932]Driver  "modesetting"
> [ 13790.932]EndSection
> [ 13790.932]Section "Screen"
> [ 13790.932]Identifier  "Builtin Default modesetting Screen 0"
> [ 13790.932]Device  "Builtin Default modesetting Device 0"
> [ 13790.932]EndSection
> [ 13790.932]Section "Device"
> [ 13790.932]Identifier  "Builtin Default vesa Device 0"
> [ 13790.932]Driver  "vesa"
> [ 13790.932]EndSection
> [ 13790.932]Section "Screen"
> [ 13790.932]Identifier  "Builtin Default vesa Screen 0"
> [ 13790.932]Device  "Builtin Default vesa Device 0"
> [ 13790.932]EndSection
> [ 13790.932]Section "ServerLayout"
> [ 13790.932]Identifier  "Builtin Default Layout"
> [ 13790.932]Screen  "Builtin Default intel Screen 0"
> [ 13790.932]Screen  "Builtin Default modesetting Screen 0"
> [ 13790.932]Screen  "Builtin Default vesa Screen 0"
> [ 13790.932]EndSection
> [ 13790.932] (==) --- End of built-in configuration ---
> [ 13790.932] (==) ServerLayout "Builtin Default Layout"
> [ 13790.932] (**) |-->Screen "Builtin Default intel Screen 0" (0)
> [ 13790.932] (**) |   |-->Monitor ""
> [ 13790.933] (**) |   |-->Device "Builtin Default intel Device 0"
> [ 13790.933] (==) No monitor 

Re: DragonFly on VirtualBox

[karu.pruun]

On Fri, Mar 9, 2018 at 6:58 AM, Pierre Abbat  wrote:
>
> The file is there, but it's /EFI/BOOT/BOOTX64.EFI, not /EFI/boot/bootx64.efi.
>
> Is there a way in VirtualBox to install DragonFly without EFI?


Yes, BIOS boot seems to be the default setting. If you want EFI boot
you need to tick the box

Settings -> System -> Motherboard -> Extended features -> Enable EFI

or else your system won't boot with your EFI/GPT installation.

As a quick test, I installed DragonFly on Virtualbox, everything works fine.

You don't need EFI on Virtualbox though, it's handy when you have
multiple systems on the same disk. On virtual installations you
typically want one system per disk and then legacy BIOS boot is fine.

Cheers

Peeter

--

Re: Reduce algebra system successfully built on DragonFlyBSD 5.2

[karu.pruun]

Hello

Cool you could build it! However, reduce is available as a port and a
binary package on DragonFly 5.0, 5.2 and master.

% pkg search reduce
reduce-20180117REDUCE Codemist Standard Lisp
general-purpose computer system
% sudo pkg install reduce-20180117

installs the snapshot as of Jan 2018. DragonFly follows FreeBSD's
ports when possible and reduce does not seem to need anything
DragonFly specific at the moment. Regarding how to create a port, have
a look at math/reduce in /usr/dports on DragonFly.

Cheers

Peeter

--




On Sat, Apr 21, 2018 at 11:35 PM, Nelson H. F. Beebe
 wrote:
> Recent correspondence on the Reduce algebra system developers list at
>
> reduce-algebra-develop...@lists.sourceforge.net
>
> raised the issue of porting that system to OpenBSD and other BSD
> members.  Reduce is the second oldest computer algebra system, and was
> developed at Utah, shortly after MIT's Macsyma; both first appeared in
> 1968, a half century ago this year.
>
> Older versions of Reduce have been ported to FreeBSD (e.g., binary
> package reduce-20170124 in FreeBSD 10.x, 11.x and 12.x, and in PCBSD
> and its succcessor TrueOS), but packaging systems of all other BSD
> family members, including DragonFlyBSD, appear to lack Reduce.
>
> I'm happy to report that a recent version of Recent has been
> successfully built on the new DragonFlyBSD 5.2 (11-Apr-2018) with this
> recipe:
>
> cd $prefix/ashare/reduce
> gtar xf ~/Reduce-svn4372-src.tar.gz
> cd Reduce-svn4372-src
> mkdir /tmp/bin
> ln -s /usr/local/bin/guname /tmp/bin/uname
> set path = ( /tmp/bin /bin /usr/bin /usr/local/bin )
> env CC='gcc6 -I/usr/local/include/freetype2 
> -I/usr/local/include/ncurses'   \
>CXX='g++6 -I/usr/local/include/freetype2 
> -I/usr/local/include/ncurses'   \
>./configure --with-csl && gmake all check
> ln -s `pwd`/bin/redcsl $prefix/bin/reduce-20180228-csl
> ln -s reduce-20180228-csl $prefix/bin/reduce-csl
> ln -s reduce-csl $prefix/bin/reduce
>
> Here is a short demonstration:
>
> % set path=( $path $prefix/ashare/reduce/Reduce-svn4372-src/scripts )
>
> Reduce (Free CSL version, revision 4330), 21-Apr-18 ...
>
> 1: int(sin(x), x);
>
>  - cos(x)
>
> 2: int(exp(-x) * sin(x), x);
>
>   - (cos(x) + sin(x))
> --
> x
>  2*e
> 3: on rounded;
>
> 4: precision 75;
>
> 12
>
> 5: sin(22);
>
>  
> -0.00885130929040387592169025681577233246328920395133256644233083529808955201463
>
> More information about publications about Reduce can be found in
> bibliographies at
>
> http://www.math.utah.edu/tex/bib/
>
> in the files
>
> red-a-f.bib red-g-l.bib red-m-z.bib redbooks.bib redextra.bib 
> reduce.bib
>
> and at the FAQ at
>
> http://www.math.utah.edu/faq/reduce/
>
> The official SourceForce site for Reduce source code is
>
> http://reduce-algebra.sourceforge.net/
>
> Its release as Free Software was announced in its original author's
> paper
>
> Anthony C. Hearn
> REDUCE is free software as of January 2009
> ACM Communications in Computer Algebra 43(1--2) 15--16 March/June 2009
> https://doi.org/10.1145/1610296.1610299
>
> and its first 40 years are described at
>
> http://reduce-algebra.com/reduce40.pdf
>
> Three further points should be noted:
>
> (1) Reduce can be hosted on either Portable Standard Lisp (PSL) or on
> Codemist Standard Lisp (CSL).  My DragonFlyBSD build selected the
> latter, because the PSL version has proved more difficult to port,
> and because PSL will not currently work on systems with W^X
> (write, or execute, but not both) protection for stack memory.
>
> (2) Reduce must currently be built where it is to be installed; that
> restriction may be lifted in the future, because it makes Reduce
> unusual compared to other packages.
>
> (3) I had to extend PATH above to run it outside its build tree; that
> has not been necessary on other O/Ses, but a 3-line wrapper shell
> script could easily hide that misfeature on DragonFlyBSD.
>
> Perhaps someone on this list who is familiar with package construction
> for DragonFlyBSD might be willing to build and repackage the latest
> version of Reduce for inclusion in the DragonFlyBSD package
> repository.
>
> I'm now turning to build attempts for Reduce on older DragonFlyBSD
> releases.
>
> ---
> - Nelson H. F. BeebeTel: +1 801 581 5254  
> -
> - University of UtahFAX: +1 801 581 4148  
> -
> - Department of Mathematics, 110 LCBInternet e-mail: be...@math.utah.edu  
> -
> - 155 S 1400 E RM 233   

Re: Reduce algebra system successfully built on DragonFlyBSD 5.2

[karu.pruun]

Hello

Thanks for pointing out and I am thankful for all the work Dr Beebe
has done. I genuinely meant it was great that Reduce builds on
DragonFly! I haven't used it myself but colleagues of mine do use it
for their work.

If there is anything that can be done to improve the support on
DragonFly, then I suppose people are happy to send patches upstream or
add further tweaks to the already existing port in dports. I am not
sure how much effort it would be to create a port in ravenports

http://ravenports.ironwolf.systems

Cheers

Peeter

--



On Tue, Apr 24, 2018 at 4:18 AM, Predrag Punosevac
 wrote:
>> Hello
>>
>> Cool you could build it! However, reduce is available as a port and a
>> binary package on DragonFly 5.0, 5.2 and master.
>>
>
> You must be new in the UNIX world or completely oblivious to whom you
> were responding. The person behind original email is Dr. Nelson H. F.
> Beebe himself
>
> https://www.math.utah.edu/~beebe/
>
> or his UNIX books on Amazon
>
> https://www.amazon.com/Classic-Shell-Scripting-Hidden-Commands-ebook/dp/B0043GXMPW
>
>
> Reduce would not have being available in ports as quite a few other
> basic tool from UNIX toolbox if not for the hard work of Dr. Beebe. He
> is the gatekeeper. He is the upstream. He is the guy who tests many
> things on all those UNIX-es you have heard of and you have never heard
> of. He is testing all those packages on those exotic architectures.
> Without people like Dr. Beebe all of us would be running Wintel crap and
> Windows XP right now including super computing guys like me.
>
> Cheers,
> Predrag
>
>
>
>> % pkg search reduce
>> reduce-20180117REDUCE Codemist Standard Lisp
>> general-purpose computer system
>> % sudo pkg install reduce-20180117
>>
>> installs the snapshot as of Jan 2018. DragonFly follows FreeBSD's
>> ports when possible and reduce does not seem to need anything
>> DragonFly specific at the moment. Regarding how to create a port, have
>> a look at math/reduce in /usr/dports on DragonFly.
>>
>> Cheers
>>
>> Peeter
>>
>> --
>>
>>
>>
>>
>> On Sat, Apr 21, 2018 at 11:35 PM, Nelson H. F. Beebe
>>  wrote:
>> > Recent correspondence on the Reduce algebra system developers list at
>> >
>> > reduce-algebra-develop...@lists.sourceforge.net
>> >
>> > raised the issue of porting that system to OpenBSD and other BSD
>> > members.  Reduce is the second oldest computer algebra system, and was
>> > developed at Utah, shortly after MIT's Macsyma; both first appeared in
>> > 1968, a half century ago this year.
>> >
>> > Older versions of Reduce have been ported to FreeBSD (e.g., binary
>> > package reduce-20170124 in FreeBSD 10.x, 11.x and 12.x, and in PCBSD
>> > and its succcessor TrueOS), but packaging systems of all other BSD
>> > family members, including DragonFlyBSD, appear to lack Reduce.
>> >
>> > I'm happy to report that a recent version of Recent has been
>> > successfully built on the new DragonFlyBSD 5.2 (11-Apr-2018) with this
>> > recipe:
>> >
>> > cd $prefix/ashare/reduce
>> > gtar xf ~/Reduce-svn4372-src.tar.gz
>> > cd Reduce-svn4372-src
>> > mkdir /tmp/bin
>> > ln -s /usr/local/bin/guname /tmp/bin/uname
>> > set path = ( /tmp/bin /bin /usr/bin /usr/local/bin )
>> > env CC='gcc6 -I/usr/local/include/freetype2 
>> > -I/usr/local/include/ncurses'   \
>> >CXX='g++6 -I/usr/local/include/freetype2 
>> > -I/usr/local/include/ncurses'   \
>> >./configure --with-csl && gmake all check
>> > ln -s `pwd`/bin/redcsl $prefix/bin/reduce-20180228-csl
>> > ln -s reduce-20180228-csl $prefix/bin/reduce-csl
>> > ln -s reduce-csl $prefix/bin/reduce
>> >
>> > Here is a short demonstration:
>> >
>> > % set path=( $path 
>> > $prefix/ashare/reduce/Reduce-svn4372-src/scripts )
>> >
>> > Reduce (Free CSL version, revision 4330), 21-Apr-18 ...
>> >
>> > 1: int(sin(x), x);
>> >
>> >  - cos(x)
>> >
>> > 2: int(exp(-x) * sin(x), x);
>> >
>> >   - (cos(x) + sin(x))
>> > --
>> > x
>> >  2*e
>> > 3: on rounded;
>> >
>> > 4: precision 75;
>> >
>> > 12
>> >
>> > 5: sin(22);
>> >
>> >  
>> > -0.00885130929040387592169025681577233246328920395133256644233083529808955201463
>> >
>> > More information about publications about Reduce can be found in
>> > bibliographies at
>> >
>> > http://www.math.utah.edu/tex/bib/
>> >
>> > in the files
>> >
>> > red-a-f.bib red-g-l.bib red-m-z.bib redbooks.bib redextra.bib 
>> > reduce.bib
>> >
>> > and at the FAQ at
>> >
>> > http://www.math.utah.edu/faq/reduce/
>> >
>> > The official SourceForce site for Reduce source code is
>> >
>> > http://reduce-algebra.sourceforge.net/
>> >
>> > Its release as Free Software was announced in its original author's
>> > paper
>> >
>> >  

Re: GUI problems on Dragonfly

[karu.pruun]

Can you check that the syntax for starting dbus in /etc/rc.conf is
correct? It should read

dbus_enable="YES"

i.e. no 'd' in 'enable'. As a quick test, I installed the current
lumina package, lumina-1.4.1,3, created the .xinitrc with a single
line

exec /usr/local/bin/start-lumina-desktop

and used startx --- lumina starts up fine. Can you check if this works
for you? If not, can you force update all the packages with

sudo pkg upgrade -f

and try again?

Cheers

Peeter

--




On Tue, Mar 27, 2018 at 1:37 AM, Ravnengel  wrote:
> Hello fellow DragonFly BSD users,
>
> I am a newbie to DragonFly and have experienced some difficulties with
> setting up a graphical interface.
>
> After having installed the system itself;
> having done the obligatory pkg install xorg (and lumina-desktop);
> having added dbus_enabled="YES" to /etc/rc.conf;
> and doing a trick to start lumina - from a small shell script that contains
> exec xinit /usr/local/bin/start-lumina-desktop
> (this was taken from https://github.com/trueos/lumina/issues/280 )
>
>
> - it worked for a while, but in the last few days, I get into the same
> problem: the system boots up correctly, but starting a GUI ends up in "No
> Signal" on the screen and complete darkness.
>
> What else should I check?
> I see other users experiencing something similar, so I think it might be
> useful to set up a detailed troubleshooting.
>
> Does anyone have any further ideas on how to solve this?
> Thanks in advance!
>
> Sincerely,
>
> A.A. Arman a.k.a. Ravnengel
>
>
>

Re: X on dell D630 / intel GM965 only shows black screen

[karu.pruun]

Hello again

To get more background, can you boot to linux and list pci devices
with sth like 'lspci -vv -nn'? Also, do you get one or two drm devices
in linux, ie what do you see with 'ls -l /dev/dri'. Can you get
similar info for OpenBSD that you say works fine?

Cheers

Peeter

--




On Fri, Mar 16, 2018 at 12:44 PM, karu.pruun <karu.pr...@gmail.com> wrote:
> Hello
>
> When xorg starts it first loads the respective drm driver and then
> starts xorg. Doing these steps manually show the sequence fails at
> setting up the drm/i915 driver.
>
> Now when you have two gpus, then the situation may get problematic. If
> there are two different devices, as in intel and nvidia, then one
> problem I've seen is that although the driver attaches correctly, we
> generally do not have the infrastructure to deal with this situation
> 100% correctly. The intel driver is not aware of the other gpu, and it
> will grab the output, leaving the screen frozen because the actual
> output to the screen is rendered from the buffers of the first device,
> the nvidia chip. In my case the solution was to flip a switch (on a
> gmux device) so output comes from the intel device. The nvidia chip
> was left on; it does not matter. From the point of power saving it
> might be better to turn it off though.
>
> However, in your case the system seems to have found two intel
> devices. I'm not sure how the drm/i915 is supposed to handle this.
>
> By way of debugging, if you get to the db> prompt, then give 'call
> dumpsys' and then 'reboot'. The system will save the core dump at next
> boot which you can investigate. Can you do this with the latest
> snapshot (it's stable and you can install with the snapshot .img) to
> be sure the issue is still there? It's possibly still there, but
> better to have the latest snapshot in terms of debugging.
>
> Cheers
>
> Peeter
>
> --
>
>
>
>
>
>
> On Fri, Mar 16, 2018 at 11:58 AM, Pedro Andres Aranda Gutierrez
> <paag...@gmail.com> wrote:
>> Hi
>>
>> mileage may vary but I was quite successful shutting off the Nvidia card
>> with my FreeBSD.
>>
>> /PA
>>
>> On 15 March 2018 at 10:44, karu.pruun <karu.pr...@gmail.com> wrote:
>>>
>>> Hello
>>>
>>> Can you try booting up and then manually loading the i915 driver as
>>> root or with sudo
>>>
>>> % sudo kldload i915
>>>
>>> does this load fine? You should see informational output in
>>> /var/log/messages. What does
>>>
>>> % pciconf -lv
>>>
>>> say about the vga devices, see the blocks starting with 'vga'. For
>>> instance, I have
>>>
>>> ---
>>> vgapci1@pci0:0:2:0:class=0x03 card=0x chip=0x00468086
>>> rev=0x18 hdr=0x00
>>> vendor = 'Intel Corporation'
>>> device = 'Core Processor Integrated Graphics Controller'
>>> class  = display
>>> subclass   = VGA
>>> vgapci0@pci0:1:0:0:class=0x03 card=0x00c7106b chip=0x0a2910de
>>> rev=0xa2 hdr=0x00
>>> vendor = 'NVIDIA Corporation'
>>> device = 'GT216M [GeForce GT 330M]'
>>> class  = display
>>> subclass   = VGA
>>> ---
>>>
>>> Can you try manually create a simple xorg.conf and set the intel
>>> driver as the only one with the pci address you get from pciconf (eg
>>> PCI:0:2:0)? In my xorg log I'm not seeing any other drivers loaded
>>> except intel and it works fine.
>>>
>>> Cheers
>>>
>>> Peeter
>>>
>>> --
>>>
>>>
>>>
>>>
>>>
>>>
>>> On Wed, Mar 14, 2018 at 6:32 PM,  <spuelmasch...@email.de> wrote:
>>> > MoinMoin,
>>> > having problems to get X running on a dell D630 / GM965 graphics laptop.
>>> > It just stays black and it's not killable but by switching it off.
>>> > I've tried intel and vesa drivers by specifying them explicitly
>>> > in /usr/local/etc/X11/xorg.conf.d
>>> > Vesa works, but is slooow.
>>> > Intel just hangs with  BusID "PCI:0:2:0" as well as with "PCI:0:2:1".
>>> > In Linux and OpenBSD modesetting drivers just work without any
>>> > modifications to the default xorg configuration.
>>> > Does anybody has a hint what to do ?
>>> >
>>> > Thanks, Tschuess,
>>> >  Sascha
>>> >
>>> > --> Xorg.0.log
>>> > [ 13790.928]
>>> > X.Org X Server 1.18.4
>>> > Release Date:

Re: Zotero on DragonFly

[karu.pruun]

Hello

I'm using Zotero and I have the exact same concerns. Seems that in the
long run getting Zotero standalone working on DragonFly is the only
option. I haven't looked at building the standalone but according to
the link you posted it can be built on FreeBSD, so I suppose it can be
built on DragonFly as well. Too bad the diffs are not available any
more. If you have time and interest maybe you can plough your way
through the build scripts? I'm definitely interested but won't be able
to invest timewise too much into this before May or so.

Cheers

Peeter

--




On Fri, Mar 2, 2018 at 6:27 PM, M. L. Wilson  wrote:
> Hello!
>
> If there is anyone else on DragonFly using Zotero with firefox-esr 52, it
> seems as though on May 8th firefox-esr 52 becomes depricated in favor of
> firefox-esr 60. The new versions of Firefox beginning with 57 break
> compatibility with a number of useful plugins.
>
> The Zotero developers have decided to switch to a standalone version that
> runs on Linux but doesn't seem to have been ported to *BSD (and likely won't
> be).
>
> There was some discussion started by a FreeBSD user on the Zotero forums:
> https://forums.zotero.org/discussion/64450/freebsd-libreoffice-firefox-zotero-java-error
>
> My question is has anyone tinkered with the standalone code and found it
> buildable/usable on DragonFly? If not what are people planning to use for
> citation manager? Zotero is incredibly useful and user friendly.
>
> --
> Michael
>

Re: LD R710 Notebook an annoying Problems..

[karu.pruun]

On Thu, Jun 28, 2018 at 11:04 AM Michael Neumann wrote:
>
> On Thu, Jun 28, 2018 at 08:36:21AM +0200, Holm Tiffe wrote:
> >
> > Hmm.. sorry for asking but I'm not sure if this is the right Operatings
> > system for me when no wone even wants to talk with me about those
> > problems.
> >
> > Do you really don't care to users problems with Dragonfly BSD?
> >
> > I need a stable platform to get my work done, currently Dragonfly
> > seems to have at least 3 or 4 problems on this machine. One is ACPI
> > hange on every other or next boot, one are those periodic kernel
> > messages about the cdrom and one is that motnths before reported problem
> > with the MSP430-UIF tool from TI and I wanted to fix that.


Hello

To keep track of the issues, it helps if you also post them on
bugs.dragonflybsd.org, and with verbose /var/run/dmesg.boot (hit 'v'
when you boot to get verbose dmesg).

Regarding slow startup of Xorg, it may be that it's not due to Xorg
but the radeonkms driver, which will be loaded before starting Xorg. I
have a radeon card too and I load the radeonkms driver before starting
Xorg. On my machine, the radeonkms driver blanks the screen for 10
seconds which may add to the perception of the Xorg slow startup. So
to test, can you do

% sudo kldload radeonkms

before starting Xorg and see if Xorg starts faster. If this is the
case, then you can put

radeonkms_load="YES"

into /etc/rc.conf and it'll be loaded at boot.

Also: you say you have two graphics devices, an intel and radeon.
Which one are you actually using? See /var/log/Xorg.0.log for
information. If Xorg gets confused at startup, it may help when you
set up an explicit xorg.conf in /etc/X11/ where you explicitly load
the Xorg intel or the radeon driver.

Giving 'kldstat' lists kernel modules so you see which specific drm
drivers have been loaded (i915 for intel, radeonkms for radeon).

Cheers

Peeter

--

Re: graphics card rec?

[karu.pruun]

Intel i915 is for Intel's integrated graphics. For a Ryzen system your
best and only bet is a radeon card. I am running 2600 with R7 360
(Gigabyte) 2GB and although there is no support for gpu acceleration,
desktop work w/o gaming is absolutely fine. I'm using HDMI though, I'm
not sure how good is the displayport support in general; maybe someone
else can comment. It can drive 3840x2160 at 30Hz but I'm using
2560x1440 at 60Hz since I find it easier on my eyes.

Cheers

Peeter

--

On Wed, Oct 17, 2018 at 4:04 AM Justin Sherrill
 wrote:
>
> The Intel i915 driver seems to be well supported and has had a good
> amount of work recently:
>
> https://leaf.dragonflybsd.org/cgi/web-man?command=i915=ANY
>
> On Tue, Oct 16, 2018 at 7:36 PM Andrew Slaughter
>  wrote:
> >
> > So, I was considering buying a Ryzen system (probably a 2600) to replace an 
> > old MB/CPU that finally bit the dust. But if I did, I'd need to buy a basic 
> > graphics card to go with it. Is there a safe bet for something that works 
> > on Dragonfly (master is fine)?
> >
> > I need a displayport, and it has to be able to drive a 3440 x 1440 monitor. 
> > No games. It's strictly coding (stats) and (minimal) 2D/3D plotting. Bonus 
> > points if it will also run on FreeBSD (on the off-chance I ever decide to 
> > use it later on).
> >
> > Thanks!

Re: Recent kernel anomalies

[karu.pruun]

As of 8 Nov (d1dbb0fb), I can't see these issues. This suggests the
need to bisect between 8 and 16 Nov.

Cheers

Peeter

--

On Fri, Nov 16, 2018 at 12:49 PM Antonio Huete Jiménez
 wrote:
>
> Hi Daniel,
>
> Would you mind pin pointing the commit that may be causing the problems?
> You can use 'git bisect' for it.
>
> Regards,
> Antonio Huete
>
> Daniel Bilik  escribió:
>
> > Hi.
> >
> > After updating to current 5.3-DEVELOPMENT (21b2a00+), chromium started to
> > fall on me. Well, not really chromium itself, but "just" extensions are
> > crashing during browser activity, taking down chromium child processes,
> > with kernel reporting something like this:
> >
> > pid 1123 (chrome), uid 1001: exited on signal 10
> > pid 1717 (chrome), uid 1001: exited on signal 4
> > pid 1512 (chrome), uid 1002: exited on signal 4
> > pid 1146 (chrome), uid 1001: exited on signal 6
> > pid 1149 (chrome), uid 1001: exited on signal 10
> > pid 1811 (chrome), uid 1001: exited on signal 11
> > pid 1724 (chrome), uid 1001: exited on signal 4
> > pid 1806 (chrome), uid 1002: exited on signal 11
> > pid 1310 (chrome), uid 1001: exited on signal 10
> > pid 1343 (chrome), uid 1002: exited on signal 10
> > pid 1021 (chrome), uid 1001: exited on signal 11
> >
> > I've also noticed another anomalous behaviour with this kernel...
> >
> > On reboot, ssh-agent(1) leaves its socket laying on the filesystem, which
> > prevents it from starting after next system boot. And, for one time,
> > Windowmaker state file was sort-of reset on reboot, and I had to recover
> > it from a snapshot.
> >
> > Other applications did not seem to have problems, or at least I did not
> > hit any more.
> >
> > Using kernel.old (5af112a+ from Oct 31) - just kernel, userland remains
> > current (21b2a00+) - everything is back to normal, ie. chromium processes
> > are not crashing and ssh-agent cleans its socket on reboot.
> >
> > From Oct 31 to now, there have been some modifications to kernel, and also
> > some vfs related ones. But looking at particular commits, I'm not sure
> > what to try to revert. Any hints?
> >
> > Thanks.
> >
> > --
> >   Daniel
>
>
>

Re: Recent kernel anomalies

[karu.pruun]

On Fri, Nov 16, 2018 at 3:35 PM Daniel Bilik  wrote:
>
> On Fri, 16 Nov 2018 10:49:33 +
> Antonio Huete Jiménez  wrote:
>
> > Would you mind pin pointing the commit that may be causing the problems?
> > You can use 'git bisect' for it.
>
> Done, bisect pointed me to commit c9f83a7+:
>
> Author: Francois Tigeot 
> Date:   Mon Nov 5 22:15:18 2018 +0100
>
> drm/linux: Fix vmap()
>
> Page protection information was not being used.
>
> Prior to this change, chromium is stable for me. This modification makes
> chromium to fall on various signals (seen 4, 6, 10 and 11). I guess the
> update causes memory corruption under some circumstancies, making the
> kernel to misbehave (which can possibly lead to a filesystem not being
> unmounted properly).

Can you specify what graphics hw are you using?

Peeter

--

Re: GRUB2 for EFI in DragonFly?

[karu.pruun]

Hello

My main OS is DragonFly, but I have multiboot settings with Mac OS,
Linux and FreeBSD on different machines and I use refind as the UEFI
boot manager. It works very well. When booting, refind will give you a
menu where you can choose the OS you want to boot. You can install
refind using a USB image or using Linux etc.

The caveat is that the DragonFly installer cannot install the system
in an arbitrary GPT partition, so you have to do a manual installation
if you want multibooting. Just boot up DragonFly's .img or .iso and
follow the example at the end of the gpt man page

https://leaf.dragonflybsd.org/cgi/web-man?command=gpt=8

skip the first (five) lines, and start by creating a gpt partition and
then create the  a, b, d BSD partitions using the disklabel64. If you
want HAMMER2, replace HAMMER with HAMMER2. At the end, copy
DragonFly's /boot/boot1.efi to the EFI system partition into the
directory

/EFI/dfly/dfly_x64.efi

You will end up with such an entry for each OS, e.g. for FreeBSD you
have something like

/EFI/freebsd/fbsd_x64.efi

and linux and MacOS create several ones at installation, I have not
touched those. The filename is not important as long as you keep the
.efi extensions and to be on the safe side I keep names 8 chars.
(Although the latter depends on the exact type of the FAT filesystem
that is your EFI system partition.) Refind will scan for the efi boot
images at startup. Also note that when linux/MacOS update their
system, they tend to reset the default UEFI boot manager, so you have
to enable refind again. I do this by reinstalling refind using the
refind USB image.

Cheers

Peeter

--

On Sun, Nov 18, 2018 at 8:30 PM Dr. Martin Ivanov
 wrote:
>
> Hello,
>
> after successfully installing Dragonfly and OpenBSD in a multiboot setup on 
> the same hard drive on an old BIOS machine, now I am planning to attempt to 
> do the same on a new machine with UEFI Firmware. I googled a lot and it seems 
> at some point I may be in need of the GRUB2 for EFI bootloader or elilo.  I 
> simple
>
> pkg search grub2
>
> or
>
> pkg search elilo
>
> provides no output . Is it possible to install any of these packages on 
> DragonFly?
>
>
> Thank you very much!
>
>
> Best regards,
>
> Martin
>
>
>
>
>
>

Re: what firewall to use ? outdated/misguided/whatever documentation ?

[karu.pruun]

>> > Basically, on OpenBSD, you use PF.  On DFly, you use IPFW.  On FreeBSD, 
>> > you can choose which style of packet filter you prefer (although I'd 
>> > recommend not using IPFilter).
>>
>> I'd suggest pf in DragonFly mostly because I've been using it on
>> DragonFly for years.  Like most things, it depends on what you want to
>> do.
>
>
> Didn't realise DFly had a PF port.  Interesting.

DragonFly's pf has evolved over time and is also SMP aware, have a
look at sys/net/pf

https://github.com/DragonFlyBSD/DragonFlyBSD/commits/e7ab884bd49753f8884eb597d10d6569a08fa0df/sys/net/pf

I'm using it exclusively.

Cheers

Peeter

--

Re: radeon driver on Kabini

[karu.pruun]

Hello

You can try adding this to /boot/loader.conf

vm.dma_reserved=512m

then reboot and see if the error persists.

Cheers

Peeter

--


On Thu, Jun 6, 2019 at 12:20 PM Daniel Bilik  wrote:
>
> Greetings to all drm experts. :)
>
> I'm trying to run Dragonfly desktop on an AMD Kabini machine, but
> radeonkms(4) kernel driver complains about some conflicts, radeon(4x)
> driver fails to initialize and causes Xorg(1) to abort startup due to
> SIGSEGV. See attached logs for details. Note that radeonkms(4) and
> radeon(4x) errors are related, ie. they appear at the same time. I suspect
> Xorg driver asks kernel to perform some operation which fails, it's not
> properly handled in the caller and the result is segfault.
>
> Any hints on where to look and what to adjust to make things work?
>
> Thanks in advance.
>
> --
> Daniel

Re: urtwn0 - driver didn't set altq_maxlen

[karu.pruun]

Just for completeness, this is all I have in /etc/rc.conf for wifi:

---
# dhcpcd
dhcpcd_enable="YES"

# realtek usb wifi
if_urtwn_load="YES"

# wpa_supplicant
wpa_supplicant_program="/usr/local/sbin/wpa_supplicant"

# wlan
wlans_urtwn0="wlan0"
ifconfig_wlan0="WPA"
---

Peeter

--


On Sun, Apr 19, 2020 at 2:58 PM karu.pruun  wrote:
>
> Hello
>
> I have the same usb wifi stick, and I see the same error, but that
> does not prevent it from working. Can you check that the rest of the
> setup in /etc/rc.conf is fine? I have
>
> dhcpcd_enable="YES"
>
> and the rest is the urtwn, wlans and wpa_supplicant as in the
> handbook. Check that the wpa_supplicant connects. Check that the
> dhcpcd is running; or restart it to be sure.
>
> Cheers
>
> Peeter
>
> --
>
>
> On Mon, Apr 13, 2020 at 11:35 PM Ron Georgia  wrote:
> >
> > All,
> >
> > I have a USB Wifi adapter (MAC/BB RTL8188CUS, RF 6052 1T1R) that is
> > having problems initializing. I did do some poking around and updated
> > loader.conf (see below) but that does not seem to help. When I do and
> > ifconfig wlan0 I see the SSID, but not IPv4, instead I see a "driver
> > didn't set altq_maxlen" error. This is not a new USB adapter, maybe 3 or
> > 4 years old. I am able to scan the networks with ifconfig wlan0 scan,
> > but I cannot connect to my home network (yes, the password is correct.)
> > Not sure what might be needed. I can post my full dmesg in paste bin if
> > that helps.
> >
> > Some Info:
> >
> > % ifconfig | grep -B3 -i wireless
> >  country US authmode WPA2/802.11i privacy ON deftxkey UNDEF
> >  AES-CCM 3:128-bit txpower 0 bmiss 7 scanvalid 60 protmode CTS wme
> >  roaming MANUAL
> >  media: IEEE 802.11 Wireless Ethernet OFDM/54Mbps mode 11g
> >
> > % sysctl net.wlan.devices
> > net.wlan.devices: urtwn0
> >
> > filtered dmesg:
> >
> > ess: 74:da:38:41:1c:68
> > rgeorgia@mini01% dmesg | grep urtw
> > urtwn0: 
> > on usbus0
> > urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> > urtwn0: enabling 11n
> > urtwn-rtl8192cfwT: You need to read the LICENSE file in
> > /usr/share/doc/legal/realtek/.
> > urtwn-rtl8192cfwT: If you agree with the license, set
> > legal.realtek.license_ack=1 in /boot/loader.conf.
> > module_register_init: MOD_LOAD (urtwn_rtl8192cfwT_fw, 82b73fa0,
> > 0) error 1
> > urtwn-rtl8192cfwT: firmware image loaded, but did not register
> > urtwn0: failed loadfirmware of file urtwn-rtl8192cfwT
> > urtwn-rtl8192cfwT: could not load firmware image, error 17
> > urtwn0: failed loadfirmware of file urtwn-rtl8192cfwT
> > urtwn-rtl8192cfwT: could not load firmware image, error 17
> > urtwn0: failed loadfirmware of file urtwn-rtl8192cfwT
> > urtwn-rtl8192cfwT: could not load firmware image, error 17
> > urtwn0: failed loadfirmware of file urtwn-rtl8192cfwT
> > urtwn0: at uhub0, port 2, addr 4 (disconnected)
> > urtwn0: detached
> > urtwn0: 
> > on usbus0
> > urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> > urtwn0: enabling 11n
> > urtwn0: at uhub0, port 2, addr 1 (disconnected)
> > urtwn0: detached
> > urtwn0: 
> > on usbus0
> > urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> > urtwn0: enabling 11n
> > urtwn1: 
> > on usbus0
> > urtwn1: MAC/BB RTL8188CUS, RF 6052 1T1R
> > urtwn1: enabling 11n
> > urtwn1: at uhub0, port 1, addr 8 (disconnected)
> > urtwn1: detached
> > urtwn1: 
> > on usbus0
> > urtwn1: MAC/BB RTL8188CUS, RF 6052 1T1R
> > urtwn1: enabling 11n
> > urtwn0: at uhub1, port 3, addr 7 (disconnected)
> > urtwn0: detached
> > urtwn1: at uhub0, port 2, addr 9 (disconnected)
> > urtwn1: detached
> > urtwn0:  > 10> on usbus0
> > urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> > urtwn0: enabling 11n
> > urtwn0: at uhub1, port 3, addr 10 (disconnected)
> > urtwn0: detached
> > urtwn0:  > 11> on usbus0
> > urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> > urtwn0: enabling 11n
> > urtwn0: at uhub1, port 3, addr 11 (disconnected)
> > urtwn0: detached
> > urtwn0: 
> > on usbus0
> > urtwn0: MAC/BB
> > urtwn0: enabling 11n
> > urtwn0: at uhub1, port 3, addr 2 (disconnected)
> > urtwn0: detached
> > urtwn0: 
> > on usbus0
> > urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> > urtwn0: enabling 11n
> > urtwn0: at uhub0, port 2, addr 5 (disconnected)
> > urtwn0: detached
> > urtwn0: 
> > on usbus0
> > urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> > urtwn0: enabling 11n
> > urtwn0: at uhub0, port 2, addr 1 (disconnected)
> > urtwn0: detached
> > urtwn0: 
> > on usbus0
> > urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> > urtwn0: enabling 11n
> >
> > /boot/loader.conf
> >
> > vfs.root.mountfrom="hammer2:serno/50026B726B053990.s1d"
> > legal.realtek.license_ack=1
> > if_urtwn_load="YES"
> > urtwn-rtl8192cfwT_load="YES"
> > urtwn-rtl8192cfwU_load="YES"
> > urtwn-rtl8188eufw_load="YES"
> >
> > --
> > “90% of my problems are due to ignorance, the other 10% is because I just 
> > don’t know any better.”
> >

Re: urtwn0 - driver didn't set altq_maxlen

[karu.pruun]

Hello

I have the same usb wifi stick, and I see the same error, but that
does not prevent it from working. Can you check that the rest of the
setup in /etc/rc.conf is fine? I have

dhcpcd_enable="YES"

and the rest is the urtwn, wlans and wpa_supplicant as in the
handbook. Check that the wpa_supplicant connects. Check that the
dhcpcd is running; or restart it to be sure.

Cheers

Peeter

--


On Mon, Apr 13, 2020 at 11:35 PM Ron Georgia  wrote:
>
> All,
>
> I have a USB Wifi adapter (MAC/BB RTL8188CUS, RF 6052 1T1R) that is
> having problems initializing. I did do some poking around and updated
> loader.conf (see below) but that does not seem to help. When I do and
> ifconfig wlan0 I see the SSID, but not IPv4, instead I see a "driver
> didn't set altq_maxlen" error. This is not a new USB adapter, maybe 3 or
> 4 years old. I am able to scan the networks with ifconfig wlan0 scan,
> but I cannot connect to my home network (yes, the password is correct.)
> Not sure what might be needed. I can post my full dmesg in paste bin if
> that helps.
>
> Some Info:
>
> % ifconfig | grep -B3 -i wireless
>  country US authmode WPA2/802.11i privacy ON deftxkey UNDEF
>  AES-CCM 3:128-bit txpower 0 bmiss 7 scanvalid 60 protmode CTS wme
>  roaming MANUAL
>  media: IEEE 802.11 Wireless Ethernet OFDM/54Mbps mode 11g
>
> % sysctl net.wlan.devices
> net.wlan.devices: urtwn0
>
> filtered dmesg:
>
> ess: 74:da:38:41:1c:68
> rgeorgia@mini01% dmesg | grep urtw
> urtwn0: 
> on usbus0
> urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn0: enabling 11n
> urtwn-rtl8192cfwT: You need to read the LICENSE file in
> /usr/share/doc/legal/realtek/.
> urtwn-rtl8192cfwT: If you agree with the license, set
> legal.realtek.license_ack=1 in /boot/loader.conf.
> module_register_init: MOD_LOAD (urtwn_rtl8192cfwT_fw, 82b73fa0,
> 0) error 1
> urtwn-rtl8192cfwT: firmware image loaded, but did not register
> urtwn0: failed loadfirmware of file urtwn-rtl8192cfwT
> urtwn-rtl8192cfwT: could not load firmware image, error 17
> urtwn0: failed loadfirmware of file urtwn-rtl8192cfwT
> urtwn-rtl8192cfwT: could not load firmware image, error 17
> urtwn0: failed loadfirmware of file urtwn-rtl8192cfwT
> urtwn-rtl8192cfwT: could not load firmware image, error 17
> urtwn0: failed loadfirmware of file urtwn-rtl8192cfwT
> urtwn0: at uhub0, port 2, addr 4 (disconnected)
> urtwn0: detached
> urtwn0: 
> on usbus0
> urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn0: enabling 11n
> urtwn0: at uhub0, port 2, addr 1 (disconnected)
> urtwn0: detached
> urtwn0: 
> on usbus0
> urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn0: enabling 11n
> urtwn1: 
> on usbus0
> urtwn1: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn1: enabling 11n
> urtwn1: at uhub0, port 1, addr 8 (disconnected)
> urtwn1: detached
> urtwn1: 
> on usbus0
> urtwn1: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn1: enabling 11n
> urtwn0: at uhub1, port 3, addr 7 (disconnected)
> urtwn0: detached
> urtwn1: at uhub0, port 2, addr 9 (disconnected)
> urtwn1: detached
> urtwn0:  10> on usbus0
> urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn0: enabling 11n
> urtwn0: at uhub1, port 3, addr 10 (disconnected)
> urtwn0: detached
> urtwn0:  11> on usbus0
> urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn0: enabling 11n
> urtwn0: at uhub1, port 3, addr 11 (disconnected)
> urtwn0: detached
> urtwn0: 
> on usbus0
> urtwn0: MAC/BB
> urtwn0: enabling 11n
> urtwn0: at uhub1, port 3, addr 2 (disconnected)
> urtwn0: detached
> urtwn0: 
> on usbus0
> urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn0: enabling 11n
> urtwn0: at uhub0, port 2, addr 5 (disconnected)
> urtwn0: detached
> urtwn0: 
> on usbus0
> urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn0: enabling 11n
> urtwn0: at uhub0, port 2, addr 1 (disconnected)
> urtwn0: detached
> urtwn0: 
> on usbus0
> urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn0: enabling 11n
>
> /boot/loader.conf
>
> vfs.root.mountfrom="hammer2:serno/50026B726B053990.s1d"
> legal.realtek.license_ack=1
> if_urtwn_load="YES"
> urtwn-rtl8192cfwT_load="YES"
> urtwn-rtl8192cfwU_load="YES"
> urtwn-rtl8188eufw_load="YES"
>
> --
> “90% of my problems are due to ignorance, the other 10% is because I just 
> don’t know any better.”
>

Re: Installing to a GPT partition

[karu.pruun]

You need to put the DragonFly efi boot image /boot/boot1.efi on your
ESP partition (which is just a FAT12, FAT16 or FAT32 partition). If
you are only booting DragonFly, then copy this image to the ESP as
follows:

EFI/BOOT/BOOTX64.EFI

It will be the default boot image for UEFI.

If you want to multiboot, you can save it as something like

EFI/DFLY/DFLYX64.EFI

and use refind as your boot manager. The other partitions you can
install as described on gpt(8). See [1] for an exhaustive and
entertaining overview of how UEFI works.

Cheers

Peeter


[1] 
https://www.happyassassin.net/2014/01/25/uefi-boot-how-does-that-actually-work-then/


On Fri, Mar 27, 2020 at 5:39 PM Gerald Henriksen  wrote:
>
> So a hard drive upgrade means there is no longer an old style drive,
> which it appears installing Dragonfly in the GPT/EFI world is
> currently an adventure.
>
> The relevant partitions (using Linux descriptions) that I would like
> to use are:
>
> /dev/sda8 - EFI
> /dev/sda9 - main Dragonfly
> /dev/sdc5 - a 10GB partition for swap
>
> Any suggestions/links on how to install Dragonfly in such a scenario?

Re: attempting to boot Dell PowerEdge R710 with uEFI still panics with 2020-08-20-LATEST

[karu.pruun]

I have an ASUS Prime TRX40-Pro mobo with Ryzen Threadripper 3970X and
it crashes with the same message when booting with 'pure' UEFI. And it
boots fine when booting with BIOS emulation (enable 'CSM
(Compatibility Support Module)').

However, on this mobo there is also a way to boot with UEFI when parts
of the CMS module are enabled. For this, I enable the CSM module; then
a sub-menu appears which enables further configuration. It gives the
following type of selections for various device classes:

[UEFI and Legacy OPROM] [Legacy OPROM only] [UEFI only]

Tweaking those I managed to boot via UEFI with the video is recognized
with the EFI buffer and a higher resolution than what BIOS would give.
I can't remember exactly which one I enabled, but I can check and
report next week when I get back to office.

So you might be able to make it work if your mobo has similar options.

Peeter

--

On Sat, Aug 22, 2020 at 12:15 AM nacho Lariguet  wrote:
>
> panic: alst_radix_free: freeing free block
>
> Please, check attached image for more details.
> Booted with loader defaults with liveCD 2020-08-20 LATEST.
> Nothing was changed.
>
> I was hoping that commit faeb2db (hack the DMAP size) could finally make my 
> Dell server boot.
>
> If anyone is successfully booting R710s (or the like) via uEFI please tell me 
> so.
> I think after almost two years attempting to run dragonFly on these machines 
> I am giving up.
>
> At this point I do have dragonFly manually installed and configured (with 
> customs scripts)
> running on a Lenovo ThinkCentre M715q via uEFI and it works without any 
> issues so far. This
> started as a test-bed to learn manual installations to later deploy it on the 
> Dell servers.
>
> Besides booting the PowerEdge servers with BIOS (instead of uEFI) I ran out 
> of options :(

Re: can I use the display native resolution on vty0 as on any other vty's using KMS ?

[karu.pruun]

Hello

I should clarify that when you load the drm driver (i915 or radeon, as
root on the console using 'kldload i915' or 'kldload radeon') the
initial buffer restriction does not apply any more. The drm drivers
allocate their own framebuffers, so all ttyv-s will have the same
column x row configuration.

In the pre-drm situation, as explained earlier, the system tries to
work out an optimal column x row configuration based on the
information it gets from BIOS or UEFI. In the case of BIOS, you get
VGA modes; we don't have VESA at the moment, yet, if I'm correct. In
the case of UEFI, the system gets the resolution from UEFI and then
tries to work out the best column x row configuration under the
restriction that it must be max 160 column x 50 rows, font width is 8
and the final configuration must preserve font scaling. For instance,
on my laptop this works out precisely as described:

- UEFI tells the resolution is 1680x1050 (X x Y)
- Font width is 8
- 1680 / 8 = 210, too large
- increase font width x 2
- 1680 / 16 = 105, yes under 160
- 1050 / 32 (font height is 16) = 32.8, round to nearest int, this is
32; yes it's under 50
- final setup: 105x32

The 105x32 is exactly what I get by default on my laptop if I set no
loader variables.

I am not sure what if anything has changed between DragonFly 5.4 and
master. The syscons driver I believe is exactly the same. The BIOS and
UEFI setups would give different results though.

I suggest you try to calculate the column x row configuration as
described above. In UEFI setup, you can see the modes as follows: at
boot, go to boot prompt (press '9'), and then do 'gop list' to see
what modes UEFI gives you. You can set the mode there too. (Give 'gop'
to see the command.) Enter 'boot' to continue booting. I guess you can
set the desired mode by default via a boot loader variable; but I
don't know what it would be.

The graphical target etc stuff is linux speak, and probably systemd;
it has no relevance at all for DragonFly.

Hope this helps.

Peeter

--

On Mon, Aug 10, 2020 at 1:33 AM nacho Lariguet  wrote:
>
> On Sun, 9 Aug 2020 13:53:22 -0700
> Jonathan Engwall  wrote:
>
> > On Sun, Aug 9, 2020, 12:00 PM  wrote:
> >
> > > Send Users mailing list submissions to
> > > users@dragonflybsd.org
> > >
> > > To subscribe or unsubscribe via the World Wide Web, visit
> > > http://lists.dragonflybsd.org/mailman/listinfo/users
> > > or, via email, send a message with subject or body 'help' to
> > > users-requ...@dragonflybsd.org
> > >
> > > You can reach the person managing the list at
> > > users-ow...@dragonflybsd.org
> > >
> > > When replying, please edit your Subject line so it is more specific
> > > than "Re: Contents of Users digest..."
> > >
> > >
> > > Today's Topics:
> > >
> > >1. Re: can I use the display native resolution on vty0 as on any
> > >   other vty's using KMS ? (nacho Lariguet)
> > >2. how do I label (meaning adding a reminder caption) to any
> > >   given BSD slice ? (nacho Lariguet)
> > >
> > >
> > > --
> > >
> > > Message: 1
> > > Date: Sat, 8 Aug 2020 18:06:14 -0300
> > > From: nacho Lariguet 
> > > To: users@dragonflybsd.org
> > > Subject: Re: can I use the display native resolution on vty0 as on any
> > > other vty's using KMS ?
> > > Message-ID: <20200808180614.66f67761@leonov>
> > > Content-Type: text/plain; charset=US-ASCII
> > >
> > > On Sat, 8 Aug 2020 11:26:35 -0700
> > > Jonathan Engwall  wrote:
> > >
> > > > Your picture looks to me like the standard "UNIX multi-user mode" 
> > > > perhaps
> > > > you should select a graphical target somehow. Are making a multi-head
> > > boot,
> > > > or bringing up each tty simultaneously?
> > > > Jonathan Engwall
> > >
> > > ... standard "UNIX multi-user mode" ... sure: do I did try/ask for
> > > anything else ?
> > >
> > > I'm afraid I'm not following
> > >
> >
> > Your photo shows the BSD/UNIX default to boot to a root system
> > administrator account. To change this you will need to boot to a graphical
> > target, or launch a multi head, or something relevant to what you are doing.
> > I don't know what you are doing. Maybe puppet would work for you?
> > Jonathan Engwall
>
> OK. Let me be more clear.
>
> I want to start using dragonFly ... daily.
>
> So the first step is to install it on the hardware that for whatever reasons
> I currently have, get it running without major issues, and last but not least,
> set a working environment that suits my needs, which ends in the following 
> fact:
>
> What I want is dragonFly running ... WITHOUT X11 !
>
> I just want to develop and work with/on the system console, which in our case,
> happens to be syscons since we still don't have vt/wcons/whatever which means
> no native UTF-8 console for the time being.
>
> I feel comfortable working/developing without a major stack like X because:
>
> - I'm learning BSD and the best way to learn 

Re: can I use the display native resolution on vty0 as on any other vty's using KMS ?

[karu.pruun]

Hello

As discussed earlier, the difference between ttyv0 vs others is almost
expected pre-drm. I don't know the syscons code, there might be other
details in play as well and that could be the reason why the algorithm
described earlier does not work out for you. Actually, I tested
another machine with UEFI and a 4K screen, it gives 80x32, which is
not the expected result using the algorithm. So pre-drm, it may work
(eg on my laptop) or it may not work (on the desktop).

However, with drm/radeon/i915 kldloaded, the difference should go
away. That seems unexpected to me.

This said, my laptop shows a difference between cold boot and warm
boot. When I cold boot the laptop and kldload i915, it won't attach or
initialize properly; I'm not sure what is happening. I have to reboot
(with power on), then everything works fine; i915 initializes
properly. Maybe something like this is in play with your machine as
well? You should see the attach etc drm messages in /var/log/messages
if everything is OK.

With UEFI I don't know if there is a utility to change the video mode.
The vidcontrol modes are for VGA, it does not work with UEFI.

The sysctl man says:

---
Some of the variables which cannot be modified during normal system
operation can be initialized via loader(8) tunables.  This can for
example be done by setting them in loader.conf(5).  Please refer to
loader.conf(5) for more information on which tunables are available and
how to set them.
---

Cheers

Peeter

--

On Tue, Aug 11, 2020 at 9:49 PM nacho Lariguet  wrote:
>
> On Tue, 11 Aug 2020 15:26:38 -0300
> nacho Lariguet  wrote:
>
> This is a follow up:
>
> If, instead of, trying to do: [sysctl kern.kms_columns=-1;] post boot
> I do [set kern.kms_columns=-1] in the boot loader options ONLY, I mean
> without initializing the DRM drivers there, I still end up with ttyv0
> @ 80x25 while all the other ttyv>=1 are @ 240x75 :(
>
> > On Tue, 11 Aug 2020 11:18:34 +0300
> > "karu.pruun"  wrote:
> >
> > > Hello
> > >
> > > > Is it possible right now to end up in a pre-DRM situation by default ?
> > > > IIRC kern.kms_console now defaults to 1 (don't know since when it does).
> > >
> > > The terminology here is a little confusing. I used pre-drm to mean
> > > "until a drm driver (i915 or radeon) is loaded". Let me try to clarify
> > > booting.
> >
> > OK. Now it is crystal-clear what pre-DRM is. My guess was pre-DRM meaned
> > before we (I mean dragonFly) started using kms_console by default.
> >
> > > The first screen with the boot menu is displayed by the loader. The
> > > kernel modules to load by the loader are given in
> > >
> > > /boot/loader.conf
> >
> > Right. The only thing I always added there is kern.kms_columns=-1
> >
> > > However, the official advice is that the loader load the absolute
> > > required minimum of kernel modules. For instance, my loader.conf has
> > > no modules. This is because memory is tight at that early stage. If
> > > you include, e.g. a drm driver, i915 or radeon, then the kernel will
> > > most likely crash at boot.
> > >
> > > The official advice is to load modules later by placing them in
> > > /etc/rc.conf, for instance,
> > >
> > > i915_load="YES"
> >
> > Right. Mine (rc.conf) has drm_load="yes" and radeon_load="yes". The
> > first is supposed to be automatically loaded by the second but I leave
> > it there to remind me of the process because it won't hurt at all.
> >
> > > I have all the required wlan, drm etc there.
> >
> > Me too.
> >
> > > And you can also load the drm modules manually on console using
> > > 'kldload i915'.
> >
> > Right.
> >
> > > This means that, for instance, if I choose to load the drm driver
> > > manually then I'm in the pre-drm stage until I will have kldloaded it,
> > > say, on console, manually. The copyright note (the "Regents" screen
> > > you mention), this is the kernel starting to load. This is pre-drm,
> > > there is no drm driver anywhere. This is either VGA or the UEFI
> > > display. The terminology "kms_columns" is misleading; that variable is
> > > read by the system to try to figure out column x row configuration,
> > > but there is no drm driver anywhere. That same "kms_columns" is also
> > > read later by the drm driver to set the column x row configuration.
> >
> > That's what mesmerizing me; suppose the following:
> >
> > I am booting @ 1920x1200 as set by the uEFI framebuffer code and then
> > while booting t

Re: can I use the display native resolution on vty0 as on any other vty's using KMS ?

[karu.pruun]

Hello

The restriction of ttyv0 to max 160 columns is inherent in the
syscons(4) driver. The man page syscons(4) says this:

---
Please note that ttyv0 is restricted to 160 columns and any setting of
kern.kms_columns resulting in a higher number of columns will be
ignored for ttyv0, but applied to other virtual terminals.
---

The reason is that at early boot, memory is tight so ttyv0 is given a
buffer of certain max size. The size of this buffer is not updated
afterwards. The other ttyv-s are initialized later, they can allocate
more memory. Hence the difference between ttyv0 and other ttyv-s.

I suppose allocating initially a buffer of fixed max size is not
uncommon at early boot. If I'm seeing it right, FreeBSD restricts
initial width to max 200 characters, which is possibly updated later.
DragonFly does not update the ttyv0 buffer; so in principle there is
room to extend the code to do this.

Cheers

Peeter

--

On Fri, Aug 7, 2020 at 2:15 AM nacho Lariguet  wrote:
>
> On Thu, 6 Aug 2020 19:28:46 -0300
> nacho Lariguet  wrote:
>
> > On Thu, 6 Aug 2020 18:49:39 -0300
> > nacho Lariguet  wrote:
> >
> > > On Wed, 5 Aug 2020 23:17:37 -0300
> > > nacho Lariguet  wrote:
> > >
> > > > On Sun, 2 Aug 2020 22:43:29 -0300
> > > > nacho Lariguet  wrote:
> > > >
> > > > > With the following configuration
> > > > > on a Lenovo ThinkCentre Tiny M715q
> > > > > with an AMD Pro A10/R7 APU installed
> > > > > hooked through DP to a HP Z-24i LCD (1920x1200 native)
> > > > > I successfully got 240x75 characters on all virtual terminals
> > > > > except on vty0 which is stuck at 80x25 (@ 640x480 I guess):
> > > > >
> > > > > ... in /boot/loader.conf:
> > > > >
> > > > > kern.kms_console=1
> > > > > kern.kms_columns=-1
> > > > >
> > > > > ... in /etc/rc.conf:
> > > > >
> > > > > drm_load="yes"
> > > > > radeon_load="yes"
> > > > >
> > > > > allscreens_flags="-t 1800 -h 100 -f 8x16 iso15-thin-8x16 -m off"
> > > > > allscreens_kbdflags="-l us.iso -r 250.34 -b off"
> > > >
> > > > This is a follow-up:
> > > >
> > > > Today I fired-up once again the laptop I last used to check dragonFly 
> > > > (a test-bed) before attempting to install it on the PowerEdge servers.
> > > >
> > > > Although it runs 5.4.1-RELEASE it has the same configuration I am 
> > > > running in the newer Lenovo regarding this post, except for the fact 
> > > > that the laptop is booting through BIOS (no uEFI at all) while the 
> > > > Lenovo is booting through uEFI.
> > > >
> > > > The laptop has a Radeon video card and loads the radeon[kms] driver as 
> > > > well.
> > > >
> > > > - 5.4.1 on the laptop gets me ttyv0 @ 1920x1200 by default
> > > > - 5.9.# on the Lenovo gets me standard VGA only
> > > >
> > > > Obviously something changed between 5.4.1 and 5.9.# because I am now 
> > > > sure not to be configuring anything different.
> > >
> > > This is another follow up:
> > >
> > > I think that vty0 is running @ 1920x1200 but not honoring 
> > > kern.kms_columns=-1 as honored by all other vty's.
> > > In other words: it seems font scaling is going on on vty0 regardless my 
> > > setting giving me the impression that is running @ 640x480 or similar. 
> > > Could it be the case on 5.9.# ?
> >
> > And the last follow up:
> >
> > Definitely.
> > Booting once again with the live image (loading radeon and KMS) gives me 
> > 80x25.
> > But when logged out as root and logged in as intaller I get 240x75.
> > So it is not a resolution problem; it is already @ 1920x1200 on vty0.
> > It is root (or any other user I presume) logging in on vty0 and getting 
> > 80x25 instead of 240x75 no matter what.
>
> I suspect what I said on my last post is wrong:
>
> Examining the installer script (installer.sh):
>
> # Console start sequence:
> # - Backend (and all other logging) goes to console (ttyv0)
> # - curses frontend starts on ttyv1.
> # - Uses vidcontrol -s 2 to switch to ttyv1 once the frontend is up.

Re: can I use the display native resolution on vty0 as on any other vty's using KMS ?

[karu.pruun]

Hello

> Is it possible right now to end up in a pre-DRM situation by default ?
> IIRC kern.kms_console now defaults to 1 (don't know since when it does).

The terminology here is a little confusing. I used pre-drm to mean
"until a drm driver (i915 or radeon) is loaded". Let me try to clarify
booting.

The first screen with the boot menu is displayed by the loader. The
kernel modules to load by the loader are given in

/boot/loader.conf

However, the official advice is that the loader load the absolute
required minimum of kernel modules. For instance, my loader.conf has
no modules. This is because memory is tight at that early stage. If
you include, e.g. a drm driver, i915 or radeon, then the kernel will
most likely crash at boot.

The official advice is to load modules later by placing them in
/etc/rc.conf, for instance,

i915_load="YES"

I have all the required wlan, drm etc there. And you can also load the
drm modules manually on console using 'kldload i915'.

This means that, for instance, if I choose to load the drm driver
manually then I'm in the pre-drm stage until I will have kldloaded it,
say, on console, manually. The copyright note (the "Regents" screen
you mention), this is the kernel starting to load. This is pre-drm,
there is no drm driver anywhere. This is either VGA or the UEFI
display. The terminology "kms_columns" is misleading; that variable is
read by the system to try to figure out column x row configuration,
but there is no drm driver anywhere. That same "kms_columns" is also
read later by the drm driver to set the column x row configuration.

So long story short: it is all pre-drm until the drm driver is
literally kldloaded.

The screen change you see when going from loader to kernel loading
(the copyright, "Regents" note), there is no resolution change. The
resolution is the same as on the loader screen, it has been set by
UEFI. What you notice though is the change of font size. Between these
two screens, the system has done the calculation described in the
previous email, set the new column x row configuration, which is
larger than what you have in the UEFI screen; but there is no change
in resolution. Or perhaps there is slight change since it tries to
allocate a buffer that covers most of the screen.

I may have missed details, maybe someone can clarify more.

Cheers

Peeter

--

On Tue, Aug 11, 2020 at 1:41 AM nacho Lariguet  wrote:
>
> On Mon, 10 Aug 2020 14:37:14 +0300
> "karu.pruun"  wrote:
>
> I'm attaching the output of "gop list" as mentioned before.
>
> You can clearly see that it boots straight to 1920x1200 right from the
> liveCD. So uEFI is informing 1920x1200 capability without any doubt.
>
> When it gets to the REGENTS message, it switches to a lower resolution.
>
> PS: I keep forgetting to CC the users list: sorry to reply twice Peeter
>
> > Hello
> >
> > I should clarify that when you load the drm driver (i915 or radeon, as
> > root on the console using 'kldload i915' or 'kldload radeon') the
> > initial buffer restriction does not apply any more. The drm drivers
> > allocate their own framebuffers, so all ttyv-s will have the same
> > column x row configuration.
> >
> > In the pre-drm situation, as explained earlier, the system tries to
> > work out an optimal column x row configuration based on the
> > information it gets from BIOS or UEFI. In the case of BIOS, you get
> > VGA modes; we don't have VESA at the moment, yet, if I'm correct. In
> > the case of UEFI, the system gets the resolution from UEFI and then
> > tries to work out the best column x row configuration under the
> > restriction that it must be max 160 column x 50 rows, font width is 8
> > and the final configuration must preserve font scaling. For instance,
> > on my laptop this works out precisely as described:
> >
> > - UEFI tells the resolution is 1680x1050 (X x Y)
> > - Font width is 8
> > - 1680 / 8 = 210, too large
> > - increase font width x 2
> > - 1680 / 16 = 105, yes under 160
> > - 1050 / 32 (font height is 16) = 32.8, round to nearest int, this is
> > 32; yes it's under 50
> > - final setup: 105x32
> >
> > The 105x32 is exactly what I get by default on my laptop if I set no
> > loader variables.
> >
> > I am not sure what if anything has changed between DragonFly 5.4 and
> > master. The syscons driver I believe is exactly the same. The BIOS and
> > UEFI setups would give different results though.
> >
> > I suggest you try to calculate the column x row configuration as
> > described above. In UEFI setup, you can see the modes as follows: at
> > boot, go to boot prompt (press '9'), and then do 'gop list' to see
> > what modes UEFI g

Re: Can sklearn be installed and if so in a venv?

[karu.pruun]

Hello

I'm using DragonFly as a jupyterhub server, installed the hub and
compute parts into different venv-s. For fortran, just do

# pkg install gcc8
# ln -s /usr/local/bin/gfortran8 /usr/local/bin/gfortran

or the same for gcc9. Then install sklearn with pip into your (compute) venv.

Cheers

Peeter

--





On Wed, Jul 1, 2020 at 3:05 PM phansi.work  wrote:
>
> Hello,
>
> 1 Would like to shift to DragonflyBSD from Linux for work. As a compute 
> server.
> 2 Able to install really no problems at all.
>
> Problem(s)
>
> 1 I use python, with sklearn. In a virtual environment, not global as require 
> different versions.
> 2 Without pasting the (huge error) pip install sklearn APPEARS to fail 
> because be lack of a fortran compiler.
> 3 sklearn seems to need one, cannot locate it,
> 4 gcc claims to have a fortran compiler, NOT gfortran but with flags which I 
> am unable to decipher.
>
> cheers
> phansi
> 

Re: Macbook Air 2017 (7,2) Support

[karu.pruun]

Hello

For what it's worth, I have a Macbookpro6,2 and /etc/rc.conf has

dbus_enable="YES"
udevd_enable="YES"

no hald or moused enabled, trackpad works fine.

Another test you could run is boot up FreeBSD using a usb stick and
see if the problem persists. If not, we can try figure out how to
patch the relevant drivers in DragonFly.

Cheers

Peeter

--


On Mon, Jun 29, 2020 at 8:47 PM Patrick McDonough <~patrick/d...@awk.is> wrote:
>
> I wouldn't have been terribly surprised if it was, but it doesn't seem to be 
> the case. I dug out some USB drives I had lying around that I presumed would 
> have at least one live environment I could boot into. If only to give me a 
> way to rule out anything DragonFly.
>
> The first was an old version of MacOS Mojave, which didn't display the 
> symptoms in its graphical installer. I didn't think that was a particularly 
> great test, so I also found a drive containing Xubuntu 20.04, which confirmed 
> the trackpad was working normally.
>
> Patrick
>
> On 2020-06-29 17:55, David Walker wrote:
>
> Might the trackpad problem be a hardware problem?
> My 2013 MacBook Air trackpad often randomly throws the cursor back
> to irrelevant positions on earlier lines when I edit text with it.
> There is advice on the web for alleviating the problem by holding
> down diagonally opposite corners of the trackpad rather firmly for
> at least 5 seconds. The problem is not supposed to occur with
> new trackpads. After holding down say lower left and upper right
> corners as described, follow up with upper left and lower right,
> and repeat the whole treatment several times. (For me that
> made a difference, but the effect wears off so I use a mouse
> instead when I can.)
> David
>
>
>
> On Mon, Jun 29, 2020 at 6:32 PM Patrick McDonough <~patrick/d...@awk.is> 
> wrote:
>>
>> Howdy. I'm working on getting Dragonfly 5.8 running on a reasonably
>> recent Macbook Air (2017), with the intent to update the
>> DragonFlyOnLaptops page with something a touch newer. The only real
>> issues I've faced are audio and trackpad related, as I've skirted around
>> dealing with Broadocm wifi using a cheap adapter board.
>>
>> Audio mostly works with snd_hda_load="YES" in loader.conf without
>> further configuration. However this is only on pcm2, which is the
>> headphone output. For speakers (pcm1) to work, I had to fiddle with the
>> gpio configuration.
>>
>> sysctl dev.hdaa.1.gpio_config=0=set
>>
>> Dandy. The trackpad less so.
>>
>> I'm seeing seemingly random middle and right clicks when using the
>> trackpad. It only happens when the trackpad is in use - so using a USB
>> mouse alleviates the problem entirely. However within seconds of moving
>> the cursor using the trackpad, I'm accosted by either middle or right
>> click events.
>>
>> My configuration is a default XFCE4 desktop on Xorg, with everything
>> installed through pkg. dbus, moused, and hald are the only additional
>> services enabled in my rc.conf.
>>
>> If anyone has a clue, that would be great. Failing that - a link to the
>> man/code for the relevant subsystem handling the trackpad
>> (xorg-input-something? moused?) would be welcome.
>>
>> Patrick
>>

Re: Python Pandas speed

[karu.pruun]

Hello

I do not know what causes the difference but one of the likely issues
is that the Linux libraries have been built using optimizations that
DragonFly is missing. I do not mean optimizations as in compiler flags
(although these count as well), I mean different code paths. The
challenge is to identify which part in the whole stack is responsible
for the optimization; and then find the relevant locations in source
code; typically you are looking for #ifdef blocks in the source and
also build part of the software.

It would be very desirable to fix issues like these. Pandas and numpy
are both fundamental in the python scientific computing stack.

Could you fix your example code so it prints out the exact table you
list above so it is easier to test if anyone is willing to pick this
up? Also, can you specify whether larger or smaller numbers are
better? In the current case smaller is better, but can you put this in
the description so it is clear.

So as the first step you can try narrow down which part of the
software is responsible for slow performance; perhaps by writing more
test cases concerning, say, only numpy? And second, you can dig
further and try optimize that dependency.

Cheers

Peeter

--


On Thu, Jul 9, 2020 at 12:10 PM phansi.work  wrote:
>
> Hello,
>
> I wonder whether I am missing some optimisation libs?
>
>
> When I calculate covariance matrix using pandas, the performance is not great.
>
>
> Results: Size of series is always 250, number varies
>
>    
> num init cov   init   cov
> 1000  0.609434 0.335977 0.392807 0.012132
> 3000  2.248877 3.412862 1.375551 0.062324
> 5000  4.797861 9.287197 2.690005 0.161746
> 7000  8.190682 18.66528 4.382373 0.29853
> 1 14.64084 38.76979 7.367079 0.604834
>
> Hope the formatting lasts.
>
> 1. The first number is the create data, dragonfly is slower, but this is not 
> something I am worried about,
> 2. The second is the covariance, this does not look good.
>
> I use a virtual environment, pandas 1.0.5 and numpy 1.19.0 in both cases.
>
> The linux mc has 16GB RAM and dragonfly has 8 GB but top did not show any 
> swap space being used.
>
> Both CPUs are i5, and both around 4 to 5 years old.
>
>
>
> Code below:
>
>
> import numpy as np
> import pandas as pd
> import datetime
> import pickle
>
> def timeme(nvals, nseries):
> t1 = datetime.datetime.now()
> # initialise data
> df = pd.DataFrame()
> for i in range(nseries):
> df[str(i)] = np.random.random_sample(size=nvals)
> t2 = datetime.datetime.now()
> # calculate covariance
> s = df.cov()
> t3 = datetime.datetime.now()
> return (t2 - t1).total_seconds(), (t3 - t2).total_seconds()
>
> def main():
> nvals = 250
> x = {}
> for nseries in [ 1000, 3000, 5000, 7000, 1 ]:
> init_time, calc_time = timeme(nvals, nseries)
> x[(nvals, nseries)] = (init_time, calc_time)
> return x
>
> x = main()
> with open("data.pickle", "wb") as fpw:
> pickle.dump(x, fpw)
>
> cheers
> phansi
> 

Re: Macbook Air 2017 (7,2) Support

[karu.pruun]

Excellent! I second that wsp be integrated into the DragonFly tree.

Peeter

--



On Sun, Jul 5, 2020 at 1:49 PM Patrick McDonough <~patrick/d...@awk.is> wrote:
>
> On 2020-06-29 17:31, Patrick McDonough wrote:
> > Howdy. I'm working on getting Dragonfly 5.8 running on a reasonably
> > recent Macbook Air (2017), with the intent to update the
> > DragonFlyOnLaptops page with something a touch newer. The only real
> > issues I've faced are audio and trackpad related, as I've skirted around
> > dealing with Broadocm wifi using a cheap adapter board.
> >
> > Audio mostly works with snd_hda_load="YES" in loader.conf without
> > further configuration. However this is only on pcm2, which is the
> > headphone output. For speakers (pcm1) to work, I had to fiddle with the
> > gpio configuration.
> >
> > sysctl dev.hdaa.1.gpio_config=0=set
> >
> > Dandy. The trackpad less so.
> >
> > I'm seeing seemingly random middle and right clicks when using the
> > trackpad. It only happens when the trackpad is in use - so using a USB
> > mouse alleviates the problem entirely. However within seconds of moving
> > the cursor using the trackpad, I'm accosted by either middle or right
> > click events.
> >
> > My configuration is a default XFCE4 desktop on Xorg, with everything
> > installed through pkg. dbus, moused, and hald are the only additional
> > services enabled in my rc.conf.
> >
> > If anyone has a clue, that would be great. Failing that - a link to the
> > man/code for the relevant subsystem handling the trackpad
> > (xorg-input-something? moused?) would be welcome.
> >
> > Patrick
> >
>
> I thought I'd make an update, as I'm at the point where I have a working
> trackpad.
>
> I spent some time yesterday trying to pinpoint where the problem was
> occurring. I work almost exclusively in Xfce on this laptop, so I
> naturally saw the problem there most of the time. Wanting to either pin
> it on Xfce, Xorg, or something lower, I fired up moused with debug
> output on the console. Pretty consistently, I'd see button presses in
> the output while only moving the cursor. So I put Xorg aside and started
> looking at drivers.
>
> FreeBSD has two drivers for Macbook trackpads, atp and wsp, and I was
> curious if there had been any work done to bring them over. A few
> searches later I found
> https://www.dragonflybsd.org/~swildner/0001-WIP-atp-4-wsp-4.patch - work
> done by Sascha Wildner a few years ago. Unfortunately, but reasonably
> given the age, the patch didn't apply cleanly. So I pulled out the wsp
> code, slipped it in by hand, and ran a make nativekernel.
>
> After a reboot, I pointed moused at /dev/wsp0, and was delighted to see
> a working trackpad, including two-finger-right-click and
> two-finger-scrolling.
>
> I'd be curious why the patch never made it into the tree. While I can't
> vouch for the atp code as I didn't touch it, the wsp module built
> without changes and worked first time.
>
> Patrick

Re: 6.0 and console on ThinkPad x220

[karu.pruun]

The kern.kms_console is set to 1 by default, you can remove it from
loader.conf. There is a man page for drm (see 'man drm') and also for
the drivers, i915 and radeon.

I checked my intel (i7 skylake) machine, I get the same errors and
messages, but drm/i915 and xorg/xfce work fine.

Just to clarify, has drm been working earlier on this machine, with
the bios booting?

When you manually load i915, is it that you see some messages on
console, and the last one is

'[drm] Initialized i915 1.6.0 20171023 for dev_name on minor 0'

and the next line 'kms console: xpixels 1366 ypixels 768' you do not
see on console, but it's in /var/log/messages?

Peeter

--

On Thu, May 20, 2021 at 11:33 PM Chuck Musser  wrote:
>
> https://man.dragonflybsd.org/?command=syscons=ANY
>
>
>
> Heh, I never managed to stumble across that. Thanks!
>
>

Re: Want to start contributing.....

[karu.pruun]

I haven't read the book but since it's FreeBSD it should be mostly
relevant with caveats since DragonFly has diverged somewhat. So on my
rough understanding, the big picture and topics are relevant, but
beware of DragonFly differences here and there, e.g. kernel locking
routines (use lockmgr(9)); slightly different syntax of driver macros;
conventions of how arguments are passed to drivers; DragonFly only has
kqueue, no need to implement poll() methods. I think that's almost all
I encountered, but there might be more. If I were you, I'd start with
the small dummy module and then build on top of it by adding features
to get used to DragonFly. Also have a look at other modules and
drivers for guidance, grok.dragonflybsd.org is a great tool to browse
the source code of DragonFly and other projects.

Best

Peeter

--

On Wed, May 19, 2021 at 5:50 PM Mike Carroll  wrote:
>
> I was looking at maybe writing a driver that I would share and was wondering 
> how relevant the FreeBSD Device Driver book would for that task.
>
> Book URL: https://nostarch.com/bsddrivers.htm

Re: 6.0 and console on ThinkPad x220

[karu.pruun]

So I am I right to think that you can get the console fine but loading
i915 gives a frozen screen? To test this is true, remove automatic
loading of i915 from /etc/rc.conf and see if you get to the login
prompt. Then try manually loading i915 on console, as root do 'kldload
i915'. You can get more debug messages by setting 'drm.debug=0x777' in
/boot/loader.conf.

It will be a lot of debug output but might help diagnose the issue.

Question: when the screen freezes after kldloading i915, can you still
ssh into the machine? Can you do 'shutdown -r now' by typing on the
console?

Best

Peeter

--

On Wed, May 19, 2021 at 6:56 PM Chuck Musser  wrote:
>
> I tried the following, which did change things but did not result in a 
> working console at the end of the boot:
>
> - added "gop set 2" in /boot/loader.conf. mode 2 was 1024x768 and seemed to 
> display fine in the loader prompt
>
> - moved the i915_load="YES" to /etc/rc.conf and removed drm_load="YES" 
> entirely. It loads anyway, as you said.
>
> - Tried both removing the kern.kms_console=1 from /boot/loader.conf and 
> leaving it in.
>
> With those changes, the top part of the screen is no longer garbled and more 
> output from the boot is shown. But the output stops after:
>
> RPM wakelock ref not held during HW access<4>RPM wakelock ref not held during 
> HW access<6>[drm] taking over the fictitious range 0xe000-0x1000
> [drm] Initialized i915 1.6.0 20171023 for dev_name on minor 0
>
> In the dmsg, right after those lines is:
>
> kms console: xpixels 1366 ypixels 768
>
> That's a different resolution than was set with gop set earlier. Not sure if  
> that is significant.
>
> Chuck

Re: 6.0 and console on ThinkPad x220

[karu.pruun]

Hello

You can try setting a different mode at the loader: get to the prompt
(press 9) and play with 'gop list', 'gop set ' and 'gop get'. Maybe
you can find a resolution that works fine?

If yes, try loading i915 in /etc/rc.conf, that's the recommended
procedure. Loading i915 too early may blow up booting. You don't need
to load drm explicitly, it's a prerequisite of i915 and will be loaded
automatically.

Best

Peeter

--

On Wed, May 19, 2021 at 12:50 AM Chuck Musser  wrote:
>
> Related (maybe) to this are these messages in dmesg:
>
> vgapci0:  port 0x5000-0x503f mem 
> 0xe000-0xefff,0xf000-0xf03f irq 16 at device 2.0 on pci0
> [drm] pdev:  vendor=0x8086  device=0x0126 rev=0x09
> [drm]   svendor=0x17aa sdevice=0x21da irq=17
> WARN_ON(domain->wake_count == 0)error: [drm:pid0:gen6_gmch_probe] *ERROR* 
> Can't set DMA mask/consistent mask (-5)
> [drm] Memory usable by graphics device = 2048M
> [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
> [drm] Driver supports precise vblank timestamp query.
> tunable drm.video.LVDS-1 is not set
> tunable drm.video.VGA-1 is not set
> tunable drm.video.HDMI-A-1 is not set
> tunable drm.video.DP-1 is not set
> tunable drm.video.HDMI-A-2 is not set
> tunable drm.video.HDMI-A-3 is not set
> tunable drm.video.DP-2 is not set
> tunable drm.video.DP-3 is not set
> RPM wakelock ref not held during HW access<4>RPM wakelock ref not held during 
> HW access<6>[drm] taking over the fictitious range 0xe000-0x1000
> [drm] Initialized i915 1.6.0 20171023 for dev_name on minor 0
> kms console: xpixels 1366 ypixels 768
> WARNING !set->fb failed at 
> /usr/src/sys/dev/drm/drm/../drm_atomic_helper.c:2750
> WARNING state->crtc && !state->fb failed at 
> /usr/src/sys/dev/drm/drm/../drm_atomic.c:862
> [drm] DRM_I915_DEBUG_GEM enabled
> drm0 on vgapci0
>
>
>
>
> They are labeled "WARNINGS", but maybe more serious?

Re: 6.0 and console on ThinkPad x220

[karu.pruun]

I could not replicate the problem on my old i915 laptop.

Can you open a bug on the bugs.dragonflybsd.org (if you don't have an
account ask on irc) and attach the drm.debug=0x777 outputs of both 5.8
and 6.0 there? Comparing them might help find the problem.

Peeter

--

On Sat, May 29, 2021 at 1:02 AM Chuck Musser  wrote:
>
> >
> > fb_info = { width = 1366, height = 768, stride = 5504, depth = 32 }
> >
> > So at least the dimensions are the same.
>
> Now that I know to use buildkernel, not quickkernel, I'll try  putting
> in some traces. One thing I noticed is that there are a number of calls
> to set the framebuffer for a plane, most of which end up up settting no
> framebuffer (it shows as [NOFB] in the logs). On 5.8, one plane
> eventually gets a framebuffer assigned to it and there are no errors. On
> 6.0, all the log messages for the planes show [NOFB] and eventually
> there are a couple error messages. One's a warning inside
> __drm_atomic_helper_set_config about the framebuffer pointer being NULL.
> So in some cases, having a NULL framebuffer is worth a warning and other
> it doesn't seem to be.
>
> Chuck

Re: Radeon Acceleration Disabled

[karu.pruun]

It seems not all radeon cards work with GPU acceleration, for instance
I get the following w/o the commit you mention

kernel: error: [drm:pid223:r600_dma_ring_test] *ERROR* radeon: ring 3
test failed (0xCAFEDEAD)
kernel: drm0: error: disabling GPU acceleration

for both Radeon R7 370 and R7 360. I am not sure why the commit was
made, perhaps the cards that generally worked had occasional glitches,
and maybe it's still the case.

Best

Peeter

--

On Sun, Jul 4, 2021 at 8:30 AM Bill Sorenson  wrote:
>
> I recently tested a Radeon R5 240 (Oland) on Dragonfly 6.0 and noticed
> acceleration was disabled. It appears it was disabled with commit
> e8de9e9.
>
> I rebuilt my kernel with acceleration re-enabled and it seems to be
> working fine. I'm using Xfce 4.16, Chrome 91 etc.
>
> Is it possible the underlying issue has been fixed incidentally? Is
> there anything specific I should test, or another GPU I could try that
> is known to exhibit issues?
>
> -Bill