py-adb likely broken on android P |
|||||
Issue descriptionhttps://chromium-swarm.appspot.com/bot?id=build315-a9--device1 New P bot, but it's quarantined due to constant protocol_fault, which is probably py-adb telling us that the device isn't following the adb protocol it thinks it should. Same thing happened when we first added an N bot (fixed in bug 626654) Into adb internals I go...
,
Mar 21 2018
Also +Kevin since skia's usually ahead of the curve on these things.
,
Mar 23 2018
got a phone flashed w/ P locally and it's failing with: WARNING:adb.cmd:SYNC FAILURE: 1/27: Unknown CNXN response: CNXN, 16777217, pktsize:262144, 'device::ro.product.name=walleye;ro.product.model=Pixel 2;ro.product.device=walleye;features=cmd,stat_v2,shell_v2' That's py-adb alright
,
Mar 27 2018
The breaking adb change: https://android.googlesource.com/platform/system/core/+/de4719494060c3436e7f8731bdf174102f2e839f They incremented protocol version by one. Time to mirror that in py-adb.
,
Mar 28 2018
I replicated the change in #4 into pyadb, but now it's broken in another way. It successfully auth's with the device and starts sending shell commands to init the cache, but it never gets past more than a couple commands before crashing. (Sometimes it completes 3-4 commands, sometimes it doesn't even complete the first one...) It's timing out when requesting more packets from the device than the device is sending. Something's real broke...
,
Mar 29 2018
FWIW: Skia has not yet tried things flashed to P.
,
Mar 29 2018
,
Mar 29 2018
Why the RVG? There's a P developer release out now: (and that's the image I'm testing) https://developer.android.com/preview/download.html
,
Mar 29 2018
ah, right. was mostly reflexive based on it being a P issue.
,
Apr 2 2018
,
Apr 5 2018
The following revision refers to this bug: https://chromium.googlesource.com/infra/luci/python-adb/+/2d12b5662ec162675f98e33a236536b52ac7c782 commit 2d12b5662ec162675f98e33a236536b52ac7c782 Author: Benjamin Pastene <bpastene@chromium.org> Date: Thu Apr 05 17:26:30 2018 py-adb: Attempt to get py-adb working on P. This: Updates supported protocol versions. Stops writing empty data packets following stand-alone headers. Fix uptime fetching after boot. Bug: 824478 Change-Id: I29517b96a5f0b78f56e775cf74048a950f74141b Reviewed-on: https://chromium-review.googlesource.com/988241 Reviewed-by: Marc-Antoine Ruel <maruel@chromium.org> [modify] https://crrev.com/2d12b5662ec162675f98e33a236536b52ac7c782/adb/contrib/high.py [modify] https://crrev.com/2d12b5662ec162675f98e33a236536b52ac7c782/adb/adb_protocol.py [modify] https://crrev.com/2d12b5662ec162675f98e33a236536b52ac7c782/adb/contrib/adb_commands_safe.py
,
Apr 5 2018
The following revision refers to this bug: https://chrome-internal.googlesource.com/infra/puppet/+/d8077d3e00cb87d019772714d7b644c20817e347 commit d8077d3e00cb87d019772714d7b644c20817e347 Author: Benjamin Pastene <bpastene@chromium.org> Date: Thu Apr 05 17:45:12 2018
,
Apr 6 2018
The following revision refers to this bug: https://chromium.googlesource.com/infra/luci/luci-py.git/+/31845a5c813a52785ce182a883df86d0c1cfb990 commit 31845a5c813a52785ce182a883df86d0c1cfb990 Author: Benjamin Pastene <bpastene@chromium.org> Date: Fri Apr 06 01:39:39 2018 swarming: Roll pyadb. to pick up https://chromium.googlesource.com/infra/luci/python-adb/+/2d12b5662ec162675f98e33a236536b52ac7c782 TBR=maruel@chromium.org Bug: 824478 Change-Id: I6365bab4f3ca5e2290bb1dab4005b0d6bdc5682f Reviewed-on: https://chromium-review.googlesource.com/999299 Reviewed-by: Benjamin Pastene <bpastene@chromium.org> Commit-Queue: Benjamin Pastene <bpastene@chromium.org> [modify] https://crrev.com/31845a5c813a52785ce182a883df86d0c1cfb990/appengine/third_party/python-adb/README.swarming [modify] https://crrev.com/31845a5c813a52785ce182a883df86d0c1cfb990/appengine/third_party/python-adb/adb/adb_protocol.py [modify] https://crrev.com/31845a5c813a52785ce182a883df86d0c1cfb990/appengine/third_party/python-adb/adb/contrib/adb_commands_safe.py [modify] https://crrev.com/31845a5c813a52785ce182a883df86d0c1cfb990/appengine/third_party/python-adb/adb/contrib/high.py
,
Sep 26
We've got P devices on swarming, and everything seems reasonably stable: https://chromium-swarm.appspot.com/botlist?c=id&c=os&c=task&c=status&f=device_os%3AP&l=100&s=id%3Aasc Closing this out. |
|||||
►
Sign in to add a comment |
|||||
Comment 1 by bpastene@chromium.org
, Mar 21 2018