Sepherosa Ziehau has been doing a lot of work with various processors states to save power on DragonFly. He’s published a summary of how well the various P-state/C-state/mwait settings work. He found that setting a lower C-state can perversely improve performance.
For those saying “but how do I set these lower power states?”:
sysctl machdep.mwait.CX.idle: AUTODEEP
sysctl machdep.cpu_idle_hlt: 1 (or higher)
Do you have a ValleyView GPU? It now works much better in DragonFly, and there’s a new accelerated rendering branch to try out, too, if you follow that link.
Here’s how you test the console frame buffer on DragonFly, even though X is the way to go.
Some time ago, I acquired a Chromebook with the help of all you kind readers. Here’s a mini-report on how DragonFly works as a desktop.
The hardware: what I have is an Acer c720 Chromebook. The C720p is the touchscreen model, and is equally well-supported by DragonFly. A larger-capacity M.2 SSD (which is relatively easy to install) is the only real need, as the installed one is only 16G. It’s easy enough to see what the laptops look like; it’s nothing fancy but it’s suitably light.
The software: There’s a wide-ranging and complete install/tweak guide for the c720 and c720p on the DragonFly site. Note that it goes down to the point of even changing the keymap for the special keys on the keyboard.
Things I don’t like:
- The mousepad needs a physical click, not a tap, which decreases accuracy.
- There’s only 2G of RAM, and not expandable. You will notice this if you tend to open a lot of tabs when web browsing.
- I’ve had mousepad trouble, but I’m the only one reporting it, so I think it’s just bad hardware luck on my part.
Things I do like:
- pkg is a godsend, making installation and upgrades almost effortless. I’ve gone binary-only so far.
- Many things Just Work – for example, the xfce4 battery plugin.
- xscreensaver works great; even the 3D modules. I don’t know why it entertains me so.
- I haven’t run the battery out to make sure, but it looks like it would last a few hours. Suspend/hibernate are not supported, but low power modes are.
- There’s a lot of multi-touch shortcuts built into the touchpad.
It’s an excellent BSD laptop, for light use, at low cost. The next step up would be into Thinkpad territory, which raises the cost or increases the age – and may not be as consistently supported.
Sascha Wildner has been removing the no-longer-needed bits of i386 support in DragonFly. One of the things going away is APM, the 32-bit power management superseded by acpiconf. If you still type ‘apm’ out of habit, it’s aliased so you won’t be surprised.
Now that DragonFly can (in most cases) offer video outside of X with KMS, not just text, more console options are possible. By default, your accelerated console will scale to 80×25, but you can now tell it how many columns you want and it’ll automatically scale to fit your resolution. Or you can turn it off.
Thanks to Sepherosa Ziehau, powerd will now start the shutdown process if you are down to 2% battery on your DragonFly laptop. It also will delay for 60 seconds if you just booted up and are desperately searching for a power cable.
Matthew Dillon’s already using a 4K monitor on DragonFly, and he’s written notes on the various performance tweaks that went with it.
Right now, if you have a USB port and a need to get networked, axe(4) is your best bet.
In addition to all the Intel video updates that have gone into DragonFly, there’s been work on radeon support from Michael Neumann. This will show in the next release, coming soon. (Just a few patches more…)
There’s a new ‘ifconsole’ option for /etc/ttys on DragonFly that may help you if your serial output device is a bit strange.
A recent commit from Matthew Dillon means users of Intel Haswell or later CPUs will see reduced power usage, if I’m reading this commit correctly.
If you want to use a scanner on DragonFly, install SANE. That is apparently all you need to do.
This may not be a huge surprise, but the Minnowboard MAX can run DragonFly just fine, modulo some dmesg complaints.
Francois Tigeot has committed his Broadwell work, which has a longer-than-I-realized list of benefits. Does anyone have a 4k screen to try?
I’m late posting about these, but they go together: Sepherosa Ziehau has added the ability to read CPU temperature through various sysctls, and the same for DIMM temperature readings.
Matthew Dillon bought a system with a Broadwell series CPU, installed DragonFly, and wrote up his experience. Read it if you plan on purchasing this hardware any time soon.
Francois Tigeot has a new update to the drm/i915 driver for testing. It matches, feature-wise, what’s in Linux 3.12. Try it if you’ve got the hardware. (and dragonfly-master)
Matthew Dillon pulled in a new USB update from FreeBSD to DragonFly. What does it change? I’m not completely sure, but he did it to get apcupsd working, so that may be a hint.
If you’ve been sitting with a Radeon-based video card and wishing you had all the nice updates i915 users are getting, today is your lucky day. Michael Neumann has brought Radeon support equivalent to Linux 3.9 into DragonFly, and he has a 3.10 branch for testing if you feel adventurous.
I admit I never thought about it much, but I’ve also never had enough RAM to matter: there’s a memtemp(4) tool that monitors temperature sensors for your system’s memory. Sepherosa Ziehau has updated it on DragonFly to support some newer processor setups.
If you have a em(4)/emx(4) card, AKA ‘Intel(R) PRO/1000′, Michael Neumann has an update for you. It’s from Intel’s 7.2.4 release of the code. This is to support the new I218 cards. Initial reports are positive.
Matthew Dillon brought in some wireless networking updates – the ath(4), iwn(4), and wpi(4) drivers are updated. There’s porting notes if you need them, too. In related news, rum(4) is also improved. The updates apparently benefited DragonFly and FreeBSD at the same time.
Francois Tigeot has updated the i915 drivers in DragonFly (remember the call for testing) to match what’s in Linux 3.10, which means you should get excellent accelerated video performance on most any recent Intel video chipset, on DragonFly.
If you have very recent alc(4) hardware, it may be supported now. If you are booting over NFS, it may be faster now. These changes are unrelated other than both being recent – NFS is improved for any chipset.
powerd now can be adjusted on DragonFly, for quicker returns to high CPU frequencies, or slower … slowdowns? It’s quickly quick or slowly slow. That’s not the best explanation, but I like rhymes. For a less stupid description, look at the man page, which now includes usage examples.
Francois Tigeot has updated the drm/i915 code again, matching Linux 3.10 for feature level… but it’s a big update. If you are
- Running DragonFly-master
- Using a i915 chipset
- (optional) On a chipset that is not Haswell or Ivy Bridge
… He could use your testing and feedback.
Matthew Dillon purchased some Haswell-based motherboards, and documented his hardware setup, for anyone who is looking to build a decent, new DragonFly system.
Thanks to Sascha Wildner porting from FreeBSD, mixer(8) now remembers state. This is something I’ve wanted for a long time.
For whatever reason, I’ve seen several people in the last week or so have mouse problems on install, and they were often solved by running moused. So, there’s your little reminder.
The short answer is ath(4) and iwn(4), via this post. There’s an update coming for the wireless infrastructure in DragonFly; Matthew Dillon and Adrian Chadd (on the FreeBSD side) are working together for improvements.
While I’m mentioning recommendations, the Silicon Image 3132 chipset is apparently excellent for eSATA drives on DragonFly.
Francois Tigeot has performed a major upgrade of DragonFly’s sound system. If you had sound problems or unsupported hardware before, this may fix them. It will require a full buildworld+buildkernel.
Romick posted some more tips on setting up various special keys on an Acer c720 Chromebook, running DragonFly of course, and Matthew Dillon also has backlight key configuration. I wish I had a spare $200 right now for one of these.
Do you remember the BSDNow story a while ago about a Tanzanian community effort using FreeBSD to build a library? They’re looking at DragonFly, too, because of the low resource requirements. From that discussion: a hardware reason for an ‘indefinite wait buffer’ error, and a note on how to most efficiently download packages for multiple machines.
I had to type it that way because it rhymes. Sascha Wildner has committed an IPMI driver port, tested/watchdogged by Markus Pfeiffer. What’s it do? It’s a machine management standard.
If you’re using one of those Acer C720 or C720p Chromebooks with DragonFly, remember to set:
To automatically enter the right power-saving states on the CPU. You used to have to do it manually, and now you don’t.
Sascha Wilder ported over the urio(4) driver to DragonFly. It’s for the USB-based Rio mp3 players. Does anyone have one of these and is running DragonFly? That would be a startling coincidence.
Sascha Wildner has removed the old USB system from DragonFly; you’re getting USB4BSD no matter what now, after the 4.0 release. While we’re at it, xhci is now automatically loaded in the installer, so installer USB drives attached to USB3 ports will work.
Markus Pfeiffer has made usb_pf work on DragonFly, which means it’s possible to dump USB traffic and filter it, similar to tcpdump. This can be handy when debugging a USB device, and that’s like 90% of all devices anyway.
Imre Vadaz’s recent change to dev/drm, adding kqueue support, has (from anecdotal reports in IRC) made video performance much better. It’s committed to DragonFly 4.0, so it’ll be in the next release.
I said “USB ethernet drivers should work now” yesterday, but didn’t specify what works with the new USB structure in DragonFly. Sascha has fixed that by explicitly porting aue(4), cue(4), ipheth(4) and kue(4) from FreeBSD. As his commit notes, there’s still a few more devices to go.
If you’ve got a USB Ethernet device on DragonFly, it should work. Also, some cell phones tether correctly now, when they may have had trouble before.
Sascha Wildner brought in led(4) from FreeBSD. It’s a driver for flashing LEDs, as you might have guessed. I’d like to see someone make Blinkenlights, whether BeBox-style or just generally mysterious.
The powersaving page on dragonflybsd.org has seen a bunch of updates; this should be handy even if you aren’t on battery power that often.
Markus Pfeiffer has made it possible to control your laptop’s backlight using ACPI – if you have a i915 chipset and DragonFly. xbacklight does not work, but setting hw.acpi.video.lcd0.brightness does.
Thanks to Francois Tigeot, the drm/i915 driver now supports Haswell processors, and the special FDI and DDI ports they use. I am late posting this.
If you are using the ATI Mach64 drm driver on DragonFly, Francois Tigeot would like to know. He’s done something that breaks it, but he’s making the educated guess that this more-than-10-years-old card is no longer in use.
If you remember the earlier work to support DragonFly on An Acer c720 Chromebook, it’s been repeated for the c720p. The “p” means it’s a touchscreen.
If you have a DragonFly system with an iwn wireless chipset, and you are having trouble connecting and running in the 5Ghz part of the spectrum only, here’s a tip: the -ht switch may fix it.
There’s been good progress in Francois Tigeot’s work on Haswell graphics support in DragonFly. If you have one of those newer units, you should be able to use the i915 driver with it now – as long as you keep acceleration off. (You won’t notice any difference in 2D anyway.)
If you have a i915 video chipset (which appears to be most every recent laptop), Francois Tigeot would like you to try his huge patch. It doesn’t support Haswell chips yet, though it lays some of the groundwork for it.
A frequent question people ask when trying Hammer is “How can I do software RAID to cover a disk failure?” Hammer provides for streaming one volume to another, so you can duplicate drives, but there isn’t an automatic failover mechanism as there is with a RAID setup. The first answer is usually “get hardware RAID“; my preferred solution. The remaining software solutions are vinum, ccd, and lvm for DragonFly.
If you were looking for something to do, finishing Francois Tigeot’s sound update would help a lot of people. He’s currently tied up with i915 support work. The patches need device cloning to work with devfs, and midi removal.
As mentioned before, the mrsas(4) driver works best for ‘Thunderbolt’ RAID controllers. Now, the switch has happened.
Tethering now works via the urndis(4) device, from a patch contributed by Sascha Wildner/tested by Yellow Rabbit.
(Updated for correct attribution)
Francois Tigeot has been working on making i915 video support work better; with his latest update, it’s worth trying the Intel-specific driver instead of vesa if you have both the 915 chipset and are running X.
Matthew Dillon changed powerd on DragonFly so that the system is set to max performance if powerd is killed. Now you’ll know why your fans turned on!
The mfi(4) driver has had some data corruption problems on “Thunderbird” series RAID controllers. There’s a newer driver, mrsas(4), that replaces mfi(4) for these controllers and does not have these issues, but switching may mean new drive locations and therefore some work to get booting correctly again. Sascha Wildner has an extensive writeup about what this entails, and how to switch now if you have that hardware (recommended).
DragonFly now supports running on up to 256 CPUs. 256 is the initial limit on the basic interrupt controller, and it can be extended further. It’s been tested on 255 CPUs so far, since that’s the highest number of CPUs you can bring up in qemu.
The max number of CPUs on DragonFly just went from 63 to 64. This is really just a side effect of preparation to move up that limit, but I am entertained by the single-digit bump.
Did you try to install DragonFly relatively recently and it never made it past the bootloader? Apparently there’s a bug in some BIOS when using a smaller USB drive to install. The loader checks multiple places for information, and if it checks somewhere that’s ‘farther’ than the end of the disk (i.e. 6G on a 4G USB key), the machine locks up.
Matthew Dillon and Kyle Davis spent a good number of hours figuring this out today, and Matt committed a fix. So, if you were bit by this problem, try a -LATEST image about 24 hours from now and see if it works.
Sascha Wildner has added the mrsas(4) driver, which works on a variety of LSI Thunderbolt devices – a variety of RAID cards, names for which are listed in the commit message. Note that as of right now, these devices by default get taken by the mfi(4) driver, so you need to take extra steps to get mrsas(4) used.
If you have DragonFly on a laptop, and a docking station for that laptop, it may be better supported now. (no, I don’t know exactly what acpi_dock does.)
Alexandre Perrin contributed an upgrade of wpa_supplicant and hostapd for DragonFly, bringing it from 0.6.10 to version 2.1 – a 4-year jump.
A relatively calm week – probably because there were many people at BSDCan.
Sascha Wildner has updated ral(4) with support for RT2700/2800/3090 chipsets. If you have one of these, please try it out – these are untested.
Matthew Dillon brought in Adrian Chadd’s sleep state changes for the ath(4) driver from FreeBSD to DragonFly; you may see reduced power usage if you have the appropriate hardware.
If you’re using DragonFly in qemu, virtualbox, whatever – but not VMWare – there’s a new virtio-net driver to try out.
You know what always makes me happy? When someone shows up out of the blue and says “Here; I did this cause I needed it; everyone can share.” The latest example of that is Imre Vadasz porting bwn(4), for the Broadcom BCM43xx wireless chipset over from FreeBSD to DragonFly.
In a thread about video cards on DragonFly, Francois Tigeot listed good ATI cards to try, and pointed out the VESA driver is probably your best bet right now with NVidia cards.
The acpi_thinkpad module (section? code?) has been updated. Update if you are on DragonFly 3.7, or be patient if you are on 3.6.
I’ve been away because of some home construction taking up time, but this has actually been happening for a while: plenty of USB device drivers have been getting ported in to work with the new USB4BSD stack. My links for that are not comprehensive.
In part because I asked him, Sepherosa Ziehau benchmarked 10G ix(4) with 2 ports on DragonFly. The results? Good, both for bandwidth and for CPU usage.
DragonFly has moved from the old USB stack to USB4BSD by default. That means:
- If you are already using USB4BSD, you will want to remove WANT_USB4BSD from your kernel config.
- If you have trouble, switch back to the old USB.
- There’s some drivers that are not yet converted; help with them would be appreciated.
- A full kernel/world build and ‘make upgrade’ will be needed in either case.
Sascha Wildner’s announcement email has all the gory details, including the kernel config changes to move back to the old USB setup. This is of course in master; 3.6 users are unaffected.
Sascha Wildner has updated arcmsr(4), which brings in support for the Areca ARC1214, ARC1224, ARC1264, ARC1284, and ARC1883 models, from FreeBSD. Please test if you have the appropriate hardware.
If you’re using the i915 driver for xorg, and xorg dies with a “No monitor specified for screen” error, there’s a config change to fix that, or you can just update.