Monday, 2022-12-19

T42<edp_17> mal : I think bluez5 is backported to my 3.4 kernel. I've checked all of these and found 0001 and 00012:20
T42<edp_17> 0003 in the kernel, however 0002 is not there :
T42<edp_17> So, I think I should use bleuz5 rather than bluez4. What do you think?12:21
malof course you can try if it works12:43
T42<edp_17> Does CRYPTO_CMAC required? I don't have it in my kernel.12:43
malit shouldn't be difficult to patch it in12:46
T42<edp_17> I know but is it required?12:47
malnot sure but there must be a reason why it's in the patches12:53
T42<edp_17> Okay, I'll then switch back to bluez5.12:56
T42<edp_17> Hm. I've switched back to bluez5. Bluetooth service fails to start. When I try to turn BT on, shortly it disconnects the wifi. Log about it :
T42<edp_17> Any suggestion? (Previously at least, BT wasn't interfering the wifi :( )15:45
T42<edp_17> As the wifi disconnected, the previous log doesn't show the last bit. Once I re-connected the wifi, I was able to capture the missing bit. So, this shows all :
T42<edp_17> Oh, and once the wifi disconnected, it cannot connect until I do it manually.15:52
mal"Failed to access management interface" means that kernel doesn't have new enough bluetooth support, are you sure you have backported drivers in the kernel17:18
malthe patches you linked are just part of the things needed17:19
spiiroinpiggz: that "uiexception_type = call -> linger" is related display blanking policy changes around end of call (and expected)17:36
spiiroinbasically it deals with e.g.: the other end terminates call while phone is on ear -> we still want to unblank screen when phone is lifted from ear -> "linger" handles that17:37
T42<adampigg> spiiroin: long delay meany i fixed the problem in ofono :D17:50
T42<adampigg> s/meant17:51
T42<adampigg> ofono wasnt emitting disconnected signal for reason=unknown, which, is always the case on the PP, so, i tweaked that17:51
spiiroin@adampigg ofono sounds like the correct place to handle things like that17:59
T42<adampigg> yeah .... that uiexception log was the only log i was getting, with the behavior of when a call was ended, the ui would stick around for 10 seconds afterwards18:00
T42<edp_17> mal : How can I check wheter all necessary parts backported?18:00
malI mean usually backported drivers are built as modules like is mentioned in faq18:05
T42<edp_17> mal : Yeah, I've seen that and compiled hci_smd driver as module. This is what I have on device:
T42<edp_17> So, I think bluez5 has properly backported. What do you think?21:02
T42<elros34> hci smd? don't you have broadcom with some different interface?21:03
T42<edp_17> Also, seemingly the driver is loaded :
T42<edp_17> @elros34 : This is what I have in the device's kernel. I don't know (not really sure) whether I should use broadcom.21:04
T42<edp_17> Might that's why it doesn't work. Might be I should use hci_smd for bt and broadcom for wifi. I don't know.21:05
T42<edp_17> If this is possible.21:05
T42<elros34> what modules where built when you make these backports?21:06
T42<elros34> you know probably nobody else tried this backport with broadcom21:07
T42<edp_17> I didn't do the backport. I only added this : crypto-add-CMAC-support-to-CryptoAPI patch21:07
T42<edp_17> That hci_smd was already in the kernel except this CMAC thingy.21:07
T42<edp_17> That's why I asked mal today whether this cmac is really needed.21:08
T42<edp_17> Then, I added it into the kernel as it was straight forward.21:08
T42<edp_17> And also, I switched back to bluez5.21:08
T42<elros34> so you only added some CMAC (whatever  that is), how this could bring you backported bluetooth from newer kern? You need to build something don't you think21:09
T42<edp_17> However, since this cmac in the kernel, BT interferes with wifi. (altghough this can be coincidence)21:09
T42<elros34> ther is whole repo to clone and build from mal github21:10
T42<edp_17> Yes, I only added the cmac. to build that hci_smd I set this in defconfig: CONFIG_BT_HCISMD=m21:10
T42<elros34> was it even enabled previously?21:10
T42<edp_17> No, it was unset.21:11
T42<edp_17> I've checked these three patches and they were there (except the cmac) CONFIG_BT_HCISMD=m21:11
T42<edp_17> sorry, my keyboard :( . So I checked these three:
T42<elros34> forget about patches, build backported subsystem, if it fail then try to figure out what patches do you need21:11
T42<edp_17> 10.4 from hadk-faq ?21:12
T42<edp_17> I didn't do that.21:12
T42<edp_17> Okay, I'll try that again.21:12

Generated by 2.17.1 by Marius Gedminas - find it at!