hverkuil: could you please review those 3 patches:
https://patchwork.linuxtv.org/patch/48842/
https://patchwork.linuxtv.org/patch/48843/
https://patchwork.linuxtv.org/patch/48846/
they basically address your comments to compat32
I want to apply them before start pulling requests
btw, there's something wrong on your yesterday's cron job:
sparse: OK
  CC       /home/hans/work/build/sparse-git/tools/objtool/str_error_r.o
make[1]: *** [prepare3] Error 1
make: *** [sub-make] Error 2
www.xs4all.nl/~hverkuil/logs/Friday.log
s/yesterday/today/
yeah, I noticed. No idea what went wrong.
I'll look at the compat patches soon.
thank you!
mchehab: reviewed.
thank you!
sailus: https://www.mail-archive.com/linux-media@vger.kernel.org/msg128015.html
Have you looked at that patch series?
"media: imx: Switch to subdev notifiers"?
hverkuil: Thanks for reminding me.
I've reviewed a few patches in the beginning but I have to revisit it.
I'll try to review esp. the core patches as well.
But those core patches definitely need your Ack.
applied all pending pull requests...
and added most compilation patches
now, there are only 2 such patches left:
https://patchwork.linuxtv.org/patch/48859/
https://patchwork.linuxtv.org/patch/48860/
they contain just Kconfig changes to allow building those drivers on x86_64 (and on most archs)
and one big-endian fix that appears only on arm, due to some #ifdefs at the code:
https://patchwork.linuxtv.org/patch/48861/
If nobody comment about them, I'll likely apply them on Monday
with those tree, all media drivers can be built on both i386 and x86_64
There are a number of drivers that only builds on x86:
* WARNING: it won't build some drivers: VIDEO_ATOMISP_OV2722 VIDEO_ATOMISP_GC2235 VIDEO_ATOMISP_MSRLIST_HELPER VIDEO_ATOMISP_MT9M114 VIDEO_ATOMISP_GC0310 VIDEO_ATOMISP_OV2680 VIDEO_ATOMISP_LM3554 VIDEO_ATOMISP_OV5693 VIDEO_ATOMISP INTEL_ATOMISP VIDEO_IPU3_CIO2 STA2X11_VIP VIDEO_ZORAN VIDEO_ZORAN_DC30 VIDEO_ZORAN_ZR36060 VIDEO_ZORAN_BUZ VIDEO_ZORAN_DC10 VIDEO_ZORAN_LML33 VIDEO_ZORAN_LML33R10 VIDEO_ZORAN_AVS6EYES VIDEO_MEYE RADIO_MIRO
PCM20 IR_ENE IR_ITE_CIR IR_FINTEK IR_NUVOTON IR_WINBOND_CIR VIDEO_OMAP2_VOUT_VRFB VIDEO_OMAP2_VOUT VIDEO_VIA_CAMERA
(this is allyesconfig on arm)
well, actually the omap2 builds on arm too, but not with allyesconfig
I may eventually try to make a few other drivers build on arm, but for sure I won't touch on atomisp
(famous last words)
ok, I'm done with COMPILE_TEST
atomisp would require major refactoring to be able to compile outside x86
(it depends on several acpi functions)
as ipu3 also depends on ACPI but doesn't call ACPI functions, I guess atomisp is using a non-standard way to get platform data
RADIO_MIROPCM20 is another challenge - it depends on an ISA DMA infra that exists on some other archs (inclusing on arm - but just for one specific board design)
every other driver are easy to make them to build with COMPILE_TEST
outside x86
hverkuil: FYI, I updated today my sparse/smatch trees
smatch received a few new patches, but nothing new pops up for us
(I ended by reverting a premature support there for spectre issues)
it rised 3 issues - IMHO - false positives