Closed
Bug 1025034
Opened 10 years ago
Closed 10 years ago
Missing WCNSS_cfg.dat from system.img
Categories
(Firefox OS Graveyard :: GonkIntegration, defect)
Tracking
(b2g-v2.0 fixed, b2g-v2.1 fixed)
RESOLVED
FIXED
2.0 S4 (20june)
People
(Reporter: gerard-majax, Assigned: gerard-majax)
References
Details
(Whiteboard: [systemsfe])
Attachments
(1 file)
51 bytes,
text/x-github-pull-request
|
mwu
:
review+
bajaj
:
approval-gaia-v2.0+
|
Details | Review |
After landing bug 1009396, it seems there is some difference between my local build ennvironment and what pvtbuild are coming from. This results in the system.img missing WCNSS_cfg.dat. From the driver sources, this file gets installed when building proprietary code. But the firmware is also available in our device/qcom/msm8610 tree: > $ sha1sum device/qcom/msm8610/WCNSS_* backup-flame/system/etc/wifi/WCNSS_qcom_cfg.ini backup-flame/system/etc/firmware/wlan/prima/WCNSS_* > 49cfc99caa21159a48c2bd4fcd2cb1b076e94e08 device/qcom/msm8610/WCNSS_cfg.dat > 0fd305b1364c5e8b9cb74ffb2b252010b59669b8 device/qcom/msm8610/WCNSS_qcom_cfg.ini > ffec4e0d2356b94beb8ba7a46199ca728d5d5f43 device/qcom/msm8610/WCNSS_qcom_wlan_nv.bin > 49cfc99caa21159a48c2bd4fcd2cb1b076e94e08 backup-flame/system/etc/firmware/wlan/prima/WCNSS_cfg.dat > 0fd305b1364c5e8b9cb74ffb2b252010b59669b8 backup-flame/system/etc/firmware/wlan/prima/WCNSS_qcom_cfg.ini > 0fd305b1364c5e8b9cb74ffb2b252010b59669b8 backup-flame/system/etc/wifi/WCNSS_qcom_cfg.ini > ffec4e0d2356b94beb8ba7a46199ca728d5d5f43 backup-flame/system/etc/firmware/wlan/prima/WCNSS_qcom_wlan_nv.bin This means we can use those files from the device tree instead of backup folder.
Assignee | ||
Comment 1•10 years ago
|
||
This patch should be okay, currently testing it.
Assignee | ||
Comment 2•10 years ago
|
||
Prior to this patch: > $ ls -hal out/target/product/flame/system/etc/firmware/wlan/prima/WCNSS_* > -rw-r--r-- 1 alex alex 11K juin 10 14:10 out/target/product/flame/system/etc/firmware/wlan/prima/WCNSS_cfg.dat > lrwxrwxrwx 1 alex alex 34 juin 13 17:10 out/target/product/flame/system/etc/firmware/wlan/prima/WCNSS_qcom_cfg.ini -> /data/misc/wifi/WCNSS_qcom_cfg.ini > lrwxrwxrwx 1 alex alex 31 juin 13 17:10 out/target/product/flame/system/etc/firmware/wlan/prima/WCNSS_qcom_wlan_nv.bin -> /persist/WCNSS_qcom_wlan_nv.bin Removing everything, applying patch, > $ ls -hal out/target/product/flame/system/etc/firmware/wlan/prima/WCNSS_* > -rw-r--r-- 1 alex alex 11K juin 13 17:20 out/target/product/flame/system/etc/firmware/wlan/prima/WCNSS_cfg.dat > lrwxrwxrwx 1 alex alex 34 juin 13 17:20 out/target/product/flame/system/etc/firmware/wlan/prima/WCNSS_qcom_cfg.ini -> /data/misc> /wifi/WCNSS_qcom_cfg.ini > lrwxrwxrwx 1 alex alex 31 juin 13 17:20 out/target/product/flame/system/etc/firmware/wlan/prima/WCNSS_qcom_wlan_nv.bin -> /persist/WCNSS_qcom_wlan_nv.bin
Assignee | ||
Comment 3•10 years ago
|
||
Comment on attachment 8439928 [details] [review] Use WCNSS elements from device tree That does work as expected locally.
Attachment #8439928 -
Flags: review?(mwu)
Comment 4•10 years ago
|
||
Comment on attachment 8439928 [details] [review] Use WCNSS elements from device tree Commented on PR.
Attachment #8439928 -
Flags: review?(mwu)
Comment 6•10 years ago
|
||
I was just going to comment on bug 1026278, but I'll make the comment here instead. I'm going to guess that the pvt builds are using a stale base image (i.e. backup-flame directory). If I build master locally with a backup-flame created from v121-2 base image and do a full flash, then wifi works properly.
Assignee | ||
Comment 8•10 years ago
|
||
(In reply to Dave Hylands [:dhylands] from comment #6) > I was just going to comment on bug 1026278, but I'll make the comment here > instead. > > I'm going to guess that the pvt builds are using a stale base image (i.e. > backup-flame directory). Wrong guess :). There is nothing wrong with the backup-flame, for two reasons: - the bug is about the firmware bits not copied - we build kernel and wifi driver from source
Comment 9•10 years ago
|
||
So I guess I don't understand. In both cases, a full flash was done, so the entire system.img file was written. If the firmware wasn't copied, then that's a build issue. Why would my local build be different from the pvt build? Or is the firmware on a different partition?
Assignee | ||
Comment 10•10 years ago
|
||
(In reply to Dave Hylands [:dhylands] from comment #9) > So I guess I don't understand. > > In both cases, a full flash was done, so the entire system.img file was > written. If the firmware wasn't copied, then that's a build issue. Why would > my local build be different from the pvt build? > > Or is the firmware on a different partition? No, the question is: did you previously build from scratch successully? If yes, then the WCNSS_cfg.dat file was already properly copied in the out/target/product/flame/system/ directory. I guess pvt builds are always done from scratch, and that explains why I got lurred thinking my patch was okay.
Comment 11•10 years ago
|
||
OK - I just tried rm -rf out, and rebuild, and now my local flash doesn't work (wifi doesn't work). So there must have been leftover files from a previous base image.
Comment 12•10 years ago
|
||
And I confirmed that applying this patch makes wifi work again (after an rm -rf out)
Assignee | ||
Comment 13•10 years ago
|
||
Michael, I commented on the PR: while we include msm8610.mk, the flame.mk file redefined PRODUCT_COPY_FILES. This is why I'm re-copying the missing files, as it has already been done for other components.
Flags: needinfo?(mwu)
Updated•10 years ago
|
Whiteboard: [systemsfe]
Comment 14•10 years ago
|
||
Comment on attachment 8439928 [details] [review] Use WCNSS elements from device tree We shouldn't override the list from msm8610.mk, afaict. Going to r+ this to fix things for now, but it shouldn't be too bad to switch to the msm8610.mk list. I think we have to add the goldfish device repo and it'll work.
Attachment #8439928 -
Flags: review+
Flags: needinfo?(mwu)
Assignee | ||
Comment 15•10 years ago
|
||
(In reply to Michael Wu [:mwu] from comment #14) > Comment on attachment 8439928 [details] [review] > Use WCNSS elements from device tree > > We shouldn't override the list from msm8610.mk, afaict. Going to r+ this to > fix things for now, but it shouldn't be too bad to switch to the msm8610.mk > list. I think we have to add the goldfish device repo and it'll work. Filed the followup as bug 1027530 to fix this.
Assignee | ||
Comment 16•10 years ago
|
||
https://github.com/mozilla-b2g/device-flame/commit/3f58e19523d0435d214db49475f95c73a2036525
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Assignee | ||
Updated•10 years ago
|
status-b2g-v2.0:
--- → affected
Assignee | ||
Comment 17•10 years ago
|
||
Comment on attachment 8439928 [details] [review] Use WCNSS elements from device tree NOTE: Please see https://wiki.mozilla.org/Release_Management/B2G_Landing to better understand the B2G approval process and landings. [Approval Request Comment] [Bug caused by] (feature/regressing bug #): 1009396 [User impact] if declined: no wifi when building and flashing from scratch v2.0 [Testing completed]: landed on master for more than one month, no issue [Risk to taking this patch] (and alternatives if risky): few, works very well on master [String changes made]: none
Attachment #8439928 -
Flags: approval-gaia-v2.0?(bbajaj)
Assignee | ||
Comment 18•10 years ago
|
||
This should have been uplifted with bug 1009396.
Updated•10 years ago
|
Attachment #8439928 -
Flags: approval-gaia-v2.0?(bbajaj) → approval-gaia-v2.0+
Comment 19•10 years ago
|
||
v2.0: https://github.com/mozilla-b2g/device-flame/commit/e67827b94d0e5d2fadcacc6049262679c02eec30
status-b2g-v2.1:
--- → fixed
You need to log in
before you can comment on or make changes to this bug.
Description
•