Hi Emil, Chih-Wei, what about the series of latest three patches we have in oreo-x86 branch? [oreo-x86 branch] http://git.osdn.net/view?p=android-x86/external-libdrm.git;a=history;f=Android.mk;h=f832b24e99007c75ce3a8d9a3ece34d4e475e957;hb=refs/heads/oreo-x86 My doubt is about the one "android: make libdrm*.so available to the vendor partition", should we keep it (and send to mesa/drm) so that we will simply have our x86 branches aligned to mesa/drm, or should we skip it in both cases? Once this is assessed I will either send the series of 3 or just the android logging one. Mauro Mauro On Wed, Apr 17, 2019 at 7:50 PM Emil Velikov <emil.l.velikov@xxxxxxxxx> wrote: > > On Mon, 18 Mar 2019 at 13:32, Tapani Pälli <tapani.palli@xxxxxxxxx> wrote: > > On 3/18/19 3:25 PM, Robert Foss wrote: > > > Hey, > > > > > > On 3/18/19 2:11 PM, Mauro Rossi wrote: > > >> Hi, > > >> > > >> On Mon, Mar 18, 2019 at 10:58 AM Robert Foss > > >> <robert.foss@xxxxxxxxxxxxx> wrote: > > >>> > > >>> Hey Mauro, > > >>> > > >>> On 3/18/19 9:38 AM, Mauro Rossi wrote: > > >>>> Hi Robert, > > >>>> > > >>>> On Mon, Mar 18, 2019 at 9:21 AM Robert Foss > > >>>> <robert.foss@xxxxxxxxxxxxx> wrote: > > >>>>> > > >>>>> On a second note, this does not apply on libdrm/master due > > >>>>> to: > > >>>>> > > >>>>> LOCAL_SHARED_LIBRARIES := \ > > >>>>> libcutils > > >>>> > > >>>> Sorry, we have an additional Google patch, not present in libdrm/master > > >>>> that adds libdrm_platform module, but it is for a specific Google > > >>>> issue. [1] > > >>>> > > >>>> However with libdrm module we have both liblog and libcutils shared > > >>>> dependencies > > >>>> > > >>>> [1] > > >>>> http://git.osdn.net/view?p=android-x86/external-libdrm.git;a=commit;h=8ccbfeab9fb2bddf4585339a0bcbea2f1e3ffa1e > > >>>> > > >>> > > >>> Do you know if [1] causes incompatibility issues with earlier android > > >>> verions? > > >>> If not I would suggest upstreaming it too. > > >> > > >> I used those patches to build with nougat-x86 and there was no issue. > > >> > > >> To be precise I did a double rookie mistake, because __android_log_vprint > > >> not used in upstream libdrm and libdrm_platform not used either. > > >> > > >> Now starting from my mistakes, let's see if there is anything useful > > >> to libdrm project > > >> > > >> In our builds Chih-Wei Huang said that libdrm_platform is not used, > > >> meaning not added to packages list, > > >> however with oreo-x86 the build error appeared and the liblog dependency. > > >> > > >> __android_log_vprint is used with __ANDROID__ braces > > >> in a special patch [2] by Chih-Wei Huang which adds capability to > > >> print logs > > >> in logcat > > >> > > >> If it's not too invasive in libdrm, it could be useful. > > >> Cheers > > >> Mauro > > >> > > >> [2] > > >> http://git.osdn.net/view?p=android-x86/external-libdrm.git;a=commitdiff;h=bcee43063ffd52a8677029c9ae6f4203563460f4;hp=81d7264033db4946a00003bf1ee82eb6c21260f9 > > >> > > > > > > [2] Seems like a good idea to me. > > > Logcat really is the only intended path for logging on Android, and > > > redirecting our logs there does make sense to me. > > > > > > But, I'm not sure about I like the way [2] disregards log-levels in > > > drmMsg(). > > > > > > > Yeah, I think it's a good idea. I see that in mesa we include > > "android/log.h", not "log/log.h", will need to make sure we get that > > correctly, maybe older versions did "log/log.h"? > > > I would also be in favour of adding logcat support to libdrm. > > Mauro, please respin when you have the time. > > Thanks > Emil _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
