Beta Kernel 19 Dec 10

This should be the last public beta prior to official posting.

Updated BCM4329 wifi drivers to the latest from glacier
Evo/Inc Updated Atmel drivers to the latest from glacier

Evo/Inc have had the 3/5 point multi-touch enabled for some time. I have not modified the board files, since there are different versions of the controllers, instead I choose to use ffolks mod of the atmel driver, same as toast and netarchy.

BETA Kernel Bravo BFS 19 Dec 10 (570 downloads)
BETA Kernel Bravo CFS 19 Dec 10 (429 downloads)
BETA Kernel Evo BFS 19 Dec 10 (525 downloads)
BETA Kernel Evo CFS 19 Dec 10 (822 downloads)
BETA Kernel Inc BFS 19 Dec 10 (661 downloads)
BETA Kernel Inc CFS 19 Dec 10 (480 downloads)




  1. Ping from Tim:

    Switching over from netarchy…running this on my evo now, let’s see how she goes 🙂

  2. Ping from Quick:

    I’ll try this one here soon. I was liking Dec 17th’s EVO release (CFS) however I notice some screen tearing when I’m going between home screens. It wasn’t enough to go to something else but a little annoying. 🙂

    Good performance though (like always). Do you recommend that I just flash over this or do I need to do a cache/davlik-cache wipe?

  3. Ping from

    This version is running like a champ. Will you at least toss a few bugs out there so I feel like I am contributing?

  4. Ping from Tim:

    Been running this for a day — no screen tearing or other issues. Battery life has been pretty damn good too.

  5. Ping from

    OK… So I downloaded this multitouch test app from the Market, and it only registers 2 points. Should it be doing more on my DInc? Not sure if I care… or should care. Just curious.

    • Ping from Ziggy471:

      Don’t worry, I only get 2 also, don’t know why, and I’ve spent way too much time last night looking at the code to figure it out.

      • Ping from

        Ya. I figure it would take some app to actually accept more than 2 points anyway, so I won’t really miss it.

        But for what it’s worth, I know that the chad kernel is attempting to get this to work, and people are running into the same thing. Some devices work, others don’t. I don’t know if the issue has been narrowed down to something like AMOLED vs SLCD or something else. Clearly there is some difference between devices.

        • Ping from Ziggy471:

          Chad isn’t doing anything new, it’s the same mod that the Evo guys have been doing for months. It’s also been in my kernels, at least on the Evo since the beginning of October. The biggest problem with what their trying to do is their modifying the T9[14] right in the board files, which if you have a 0x4F version of the sensor, it may cause it not to work.

          Just my 2 cents.


  6. Ping from PhilE:

    Everything’s been working great so far.


  7. Ping from Quick:

    Still having screen tearing issues on the top of the screen (CFS), about 1/4″ like I was on the 12/17 release.

    Running Myn’s latest, 003 nova screen. Performance is great otherwise.

  8. Ping from jason:

    ne1 having a prob w/ 4g turning on? my 4g radio just turns on, scans for a split second, then turns off….

  9. Ping from IncredibleDude:

    also I’m running uber desire z for inc and saw that after a while the phone starts to loose a bit of snappiness but is back to norm after a restart…. im on cfs have tried your bfs yet

    • Ping from Ziggy471:

      I personally haven’t been happy with the CFS w/autogroup, I use the BFS myself. Maybe I’ll do just a plain CFS build next time also, guess I’ll be up to 9 kernels per release. Oh well, the build script is automated, so I just run it and walk away.

  10. Ping from gmichaelow:

    I’m the developer of Uber Sense. (haha, hi again)

    I’m getting hot restarts after mounting the SD card and copying files to it. I’m about to flash another kernel to see how that fares, but as I haven’t touched any SD stuff, I suspect it may be your kernel.

    Any thoughts?

    • Ping from Ziggy471:

      I’ve done the same, but on ID’s 1.2, and haven’t had any issues. The only issues I’ve had are with the CFS and the Z Roms, could just be something messed up right now, just haven’t flashed anything new recently. However, I was considering yours to try next, since ID hasn’t sent me his newer test he mentioned on Twitter.

  11. Ping from Berzerker:

    I’m getting *crap* quadrant scores, and I mean *crap* Overclocked max/min at 1228mhz, linpack only gives me a 37, and quadrant is giving me around 1100. Are you sure the BFS download is actually the BFS one? The CPU tests in quadrant look more CFS-like than BFS.

    Thanks for the great kernel, running well otherwise.

    • Ping from Ziggy471:

      Yes, it’s definitely the BFS kernel. Both the CFS w/ autogroup and the BFS kernels give me 1500+ at stock speed with smartass configured as default. In other words, I don’t use SetCPU or anything else to change anything.

      Most likely what’s happening is voltage related, either too much, or not enough. If it’s the first, you can change that yourself, since -sqn’s VDD sysfs interface is built in. If it’s the later, I’m not going to raise the voltage any more than it already is. 1.22GHz works for some but not for all.

      I’d recommend either trying a lower VDD or a lower clock rate.


  12. Ping from PrfectOne:

    Using smartass gov oc’d to 1.19ghz on Vaelpak 2.3a, hardware 003 I can regularly get a linpack of 39+ and a quadrant of 1650. I get similar results using the ondemand gov.

    The big difference is that with the smartass, the graphics pulse – hesitate every second or so noticeably. It can be seen in the quadrant tests and when playing Angry Birds. Ondemand is much smoother.

    Just thought I’d mention it to help refine that smartass gov.

  13. Ping from Eric:

    Using smartass gov oc’d to 1190min/1190max BFS kernel, i get 1958 on quadrant. On 1190mhz and performance gov my device is not stable.

  14. Ping from Eric:

    RCMix 0.14 OF

    After reinstall kernel now its ok. Its stable on 1190, performance gov. This kernel is PERFECT.

  15. Ping from Therev:

    Runnin bfs for incredible for a few days, no setcpu. It has been running great, loving the multitouch, its working for me tested on “touch test”
    it seems the cpu is staying on 998 for 10-30 seconds longer than needed, I’ve been keeping an eye on it with temp+cpu app, processor shows minimal cpu usage but freq at 998?

  16. Ping from Slice:

    why is it that even though the kernel is overclocked to 1228 my phone continually crashes at everything over 1152, is it my phone?

  17. Ping from rEVOlution89:

    Hey Ziggy just a quick question for ya, and just email me if you want. I have been using your kernels since the 12/5 Evo beta’s and the past few updates still give me great battery life, but what have you changed since the 12/5. I run the Evo CFS with setcpu just to keep it set to Smartass on boot etc. I had crazy awesome battery life with the 12/5 and with the past few updates it has still been good but nothing out of this world like 12/5, any suggestion here?

  18. Ping from Quick:

    Ziggy, have you had any other reports of screen tearing? It happens on the first 1/4″ of the top of the screen.

    I don’t have any issues with other kernels, just your 2 latest betas. I have an 003 EVO with a Nova screen.

    I see it when I’m flipping between homescreens and when scrolling text (in a browser or on Facebook).

  19. Ping from Philyteach:

    Ziggy, the kernel seems to be working fine for me…and I’m getting great battery life. Only thing I want for Christmas is a tun.ko 🙂 I actually use my EVO at work and sometimes need to use OpenVPN.

  20. Ping from PhilE:

    Just an FYI, the new giggle maps 5.0 seems to be draining the battery like crazy. Just in case anyone thought recent bad battery life had anything to-do with trying out these new kernels.

    I have an incredible and my friend with an Evo had the same battery drain with the new maps.

  21. Ping from Michael:

    Hey Ziggy. I love your kernel, except for one problem. I have a GSM HTC Desire and when your kernel is installed the camera only goes up to 3MP. With the default kernel, it goes to 5MP. If you fix this it will make me switch to this full time.

  22. Ping from Mike:

    Hey Ziggy, just flashed this on my Dinc running Skyraider 3.5. So far so good, nice and fast. I have a noob question though, is this OC’d? How can I change that back to stock? I’m not quite sure if I feel comfortable OC’ing the phone, can this make the life of the phone last less?

  23. Ping from Terepin:

    I’m having issue with BFS version for Desire with Z ROM. Entire display is in blue shade. No problem with CFS, but BFS feels snappier, but I can’t use it because of this bug. Can you have a look at it?


  24. Ping from Vince:

    Been running BETA Kernel Inc BFS 19 Dec 10 (157) on the Incredible for more than a week straight now. Stock ROM. Nothing to report except a quicker running phone that is very responsive. Using SetCPU set to max 998, min 128. Smartass gov, unchecked set on boot. No profiles. Way better kernel than stock. Battery lasts longer.

  25. Ping from Shawn K.:

    Ziggy, please check this out, an emulator developer discovered a performance issue in HTC kernels:

    I’m also anxiously awaiting your next release.

  26. Ping from wtf911:

    I’m sure you’ve seen this? …I’m hoping to see it in your next release. 🙂 …trickle charging for the Evo

  27. Ping from Therev:

    +1 for this

    Ziggy, please check this out, an emulator developer discovered a performance issue in HTC kernels:

  28. Ping from Therev:

    Check here for multi touch fix. Apparently update for clockwork blocked multi touch patch

    • Ping from

      Flashing the gave me 5 points where before I had 2. However, I get the black screen with white bars for the recovery screen. I suppose I could just use ROM Manager… but I normally don’t and just use normal recovery.

  29. Ping from Redcat:

    Looks like many are experiencing reboot lockup with SBC enabled kernels. Read some of the last pages in the threads below. I’ll pass on the SBC until that gets ironed out.

  30. Ping from Redcat:

    One other question; For the EVO kernels, is the havs module installed too? I searched, but couldn’t find a recent reference to it.


  31. Ping from billfnbragg:

    Whens the next release for the evo ziggy?

  32. Ping from wtf911:

    lol SBC makes our batteries charge to their full potential which HTC foolishly hasn’t had them do…it is not harmful to the longevity of our batteries nor is it to blame for people having problems with their phone. kindly choose to add it or at least make it available.

  33. Ping from VisedMonk:

    I’ve flashed Ziggy’s latest CFS evo kernel as well as several of the SBC HAVS more aggressive kernels, and while the full 100% charge of the SBC kernels is nice, I seem to get better battery life from Ziggy’s kernels. The SBC kernels seem to drop quickly once they leave the 100% mark.

    I dunno. That’s just my experience, and ymmv…

  34. Ping from PrfectOne:

    to really rate the effectiveness of the SBC kernels you need to get something that tells you the current being used, like Battery Monitor Widget. Even sbc kernels will sit at 100% but not truly be 100%. The only way to tell is when the current drops to 2mA and the battery is listed at at least 4200mV. This may take 30 minutes or more *after* the meter reads 100%.

    So if your sbc experience is such that it drops quickly after it is charged, I’d guess it is because it isn’t 100%, no different than the stock charging situation.

    Me personally, I don’t really care for sbc because the brainiacs at HTC must have done a bit more engineering than our favorite kernel hackers. And the little bit of extra life I am getting isn’t worth it to me to make the point. I want the best life between 90% and 5%. So far the different to me running an sbc kernel is maybe an hour or so over a two day span. Big deal.

    my 2cents (yes tl;dr)

  35. Ping from VisedMonk:

    I have battery monitor widget, and I know the SBC kernels were charging to the true 100% for me. I would have a voltage of 4202 or 4203 mV off the charger in the morning.

    All I meant was that, with my normal usage, my battery would take longer than with Ziggy’s kernels to fall from 100%, but that once it did, my battery discharges more quickly with an SBC kernel than with ZIggy’s. Like you, I care more about the 90% to 5% discharge than between 100% and 90%. That’s why I returned to Ziggy’s kernels, which imao are FAR SUPERIOR from 90% to 5% than the SBC kernels.

  36. Ping from

    Most of the time I don’t care about that upper 4-6% the various battery hack kernels improve on. I’m never far from a charger anyway. But… I discovered by accident that if I listen to Pandora while charging, it pretty much does what the hacked kernels do. I’ll pull it off the charger and shutdown Pandora, and it will stay at 99% for awhile and stop dropping at a normal rate from there. I would bet any low level activity like what Pandora is doing will trigger the same behavior. However, it does not do the same thing if I am just using the phone while it’s charging. Weird.

    That’s not an entirely normal charge pattern for me, but I’ve used it a few times when I knew I’d need that extra little amount of time.

Leave a Reply