# Dfu-util on linux fails to flash \[Solved\] **URL:** \ **Category:** Firmware\ **Created:** [April 27, 2014, 5:57pm UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080 "2014-04-27T17:57:33Z")\ **Posts on this page:** 11\ **Page:** 2
**Author:** ![zachary](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/zachary/32/20717_2.png) [@zachary](https://community.particle.io/u/zachary)\ **Post date:** [May 1, 2014, 10:09pm UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/21 "2014-05-01T22:09:07Z")
Here was the email answer from Tormod Volden of the dfu-util project: > Thanks for informing me about this issue! > > First I couldn't understand why DEFAULT\_POLLTIMEOUT is relevant for > the Core, since it is only used for a few (broken) devices that > doesn't report bwPollTimeout correctly. Then I saw in the posted log > that the Core is using an OpenMOKO USB ID! dfu-util thinks your core > is an old openmoko phone 🙂 > > We have to narrow down the device range for the OpenMOKO quirk. I > guess you are only using one device ID. Probably best if we could > figure out exactly which ID's were used for the broken openMOKOs. > > Can you try just disabling the quirk in quirks.c and test again? I am > curious to what speed and reliability you'll get now by using the > (hopefully) proper timing reported by the device. I'd appreciate as many people as possible trying out, as he says, "disabling the quirk in quirks.c" and testing with DEFAULT\_POLLTIMEOUT at its default of 5. ---
**Author:** ![Hypnopompia](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/hypnopompia/32/9039_2.png) [@Hypnopompia](https://community.particle.io/u/Hypnopompia)\ **Post date:** [May 2, 2014, 1:54am UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/22 "2014-05-02T01:54:14Z")
OK, here are the results: WITH quirks enabled and DEFAULT\_POLLTIMEOUT = 50: runtime: 0m21.690s WITHOUT quirks enabled and DEFAULT\_POLLTIMEOUT = 5, it works! runtime: 0m10.911s So yeah, we just want the quirk disabled for the spark core. It’s twice as fast to flash with. ---
**Author:** ![Dave](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/dave/32/181_2.png) [@Dave](https://community.particle.io/u/Dave)\ **Post date:** [May 2, 2014, 2:58am UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/23 "2014-05-02T02:58:44Z")
Woot! Faster dfu-util makes me very happy. 🙂 ---
**Author:** ![zachary](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/zachary/32/20717_2.png) [@zachary](https://community.particle.io/u/zachary)\ **Post date:** [May 2, 2014, 4:25am UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/24 "2014-05-02T04:25:25Z")
Nice! Thank you @Hypnopompia! That’s so AWEXOME!!! ✈ 🚀 :flipper: 🐨 🌞 The manufacturing line will go soooooo much faster! Think about that 10 seconds times 10,000 Cores… zomg. I’ll let Tormod know. ---
**Author:** ![drkemp](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/drkemp/32/2020_2.png) [@drkemp](https://community.particle.io/u/drkemp)\ **Post date:** [May 2, 2014, 6:08pm UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/25 "2014-05-02T18:08:51Z")
I confirm that disabling quirks and using the 5ms default works and is fast. I get 18 seconds and 9 seconds. ---
**Author:** ![Hypnopompia](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/hypnopompia/32/9039_2.png) [@Hypnopompia](https://community.particle.io/u/Hypnopompia)\ **Post date:** [May 2, 2014, 7:01pm UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/26 "2014-05-02T19:01:15Z")
That’s 27 minutes of flashing time saved. 🙂 ---
**Author:** ![Hypnopompia](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/hypnopompia/32/9039_2.png) [@Hypnopompia](https://community.particle.io/u/Hypnopompia)\ **Post date:** [May 2, 2014, 7:16pm UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/27 "2014-05-02T19:16:30Z")
I’ve got a patched fork over at [https://github.com/Hypnopompia/dfu-util/tree/spark](https://github.com/Hypnopompia/dfu-util/tree/spark) Not sure how to submit a pull request since the remote isn’t on github though. 🙂 ---
**Author:** ![bko](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/bko/32/20704_2.png) [@bko](https://community.particle.io/u/bko)\ **Post date:** [May 2, 2014, 7:29pm UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/28 "2014-05-02T19:29:19Z")
I think you meant to say 27.77 **hours** of time saved! Which is a huge win. ---
**Author:** ![Dave](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/dave/32/181_2.png) [@Dave](https://community.particle.io/u/Dave)\ **Post date:** [May 2, 2014, 7:47pm UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/29 "2014-05-02T19:47:03Z")
It’s actually more than that, since it gets called a lot during manufacturing, it’s probably more in the range of saving 20-40 seconds per core… ---
**Author:** ![Hypnopompia](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/hypnopompia/32/9039_2.png) [@Hypnopompia](https://community.particle.io/u/Hypnopompia)\ **Post date:** [May 2, 2014, 10:01pm UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/30 "2014-05-02T22:01:55Z")
Oh. You’re right. I can’t math today.Oops! 🙂 ---
**Author:** ![whyameye](https://sea2.discourse-cdn.com/flex026/user_avatar/community.particle.io/whyameye/32/5485_2.png) [@whyameye](https://community.particle.io/u/whyameye)\ **Post date:** [June 25, 2015, 5:50am UTC](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080/31 "2015-06-25T05:50:57Z")
None of these tricks work for me. I still get dfu-util: Error during special command “ERASE\_PAGE” get\_status [Previous page](https://community.particle.io/t/dfu-util-on-linux-fails-to-flash-solved/4080.md?page=1)