IRC log for #openezx on 20090521

01:33.57*** join/#openezx johanbr (n=j@blk-7-129-240.eastlink.ca)
02:54.31*** join/#openezx flyhorse|linux (n=flyhorse@116.4.54.10)
03:11.01*** join/#openezx arne__ (i=nobody@dslc-082-082-072-143.pools.arcor-ip.net)
04:48.12*** join/#openezx Dominus (i=unices@82-170-225-106.ip.telfort.nl)
05:37.30*** join/#openezx marex (n=marex@vasut.kolej.mff.cuni.cz)
07:53.41*** join/#openezx ao2 (n=u@2001:1418:117:0:0:0:0:1)
07:53.57ao2hi
08:28.44*** join/#openezx Sleep_Walker (n=Sleep@nat/suse/x-8bc593fbdad01dfe)
08:53.41*** join/#openezx pH5 (n=ph5@e178211112.adsl.alicedsl.de)
09:24.39*** join/#openezx pOH9 (n=ph5@e178238207.adsl.alicedsl.de)
09:28.59*** join/#openezx gDD (n=gDD@unaffiliated/gdd)
09:42.48*** join/#openezx l403 (n=l@85.132.159.239)
09:57.46*** join/#openezx marex (n=marex@eduroam81.ms.mff.cuni.cz)
11:31.05WyrMgood morning
11:53.12*** join/#openezx gmzhuo (n=gmzhuo@118.113.165.89)
11:57.14ao2hi WyrM still no comments about pcap-core?
11:57.25WyrMno. :(
11:57.33WyrMwell, I have some comments myself ;)
11:57.35WyrMlol.
11:57.50WyrMbut I will wait until someone else points my mistake ;)
11:57.56ao2:)
11:58.16WyrMthe current code depends on PXA.
11:58.45WyrMDavid Brownell asked me the previous time to make it generic enough to build even on x86.
11:59.12WyrMand the current code will fail to build with allyesconfig on all platforms, except arm. :/
11:59.57WyrMI think we need to pass a "pcap_irq_base" on platform_data.
12:00.11WyrMto remove PCAP_IRQ(0) on ezx-pcap.c
12:01.27WyrMalso, I think we should split it in 2.
12:01.40gmzhuoHi,Wyrm,ao2
12:01.48WyrMone to add the irqs to mach/irqs.h and another that puts the driver on mfd.
12:02.04WyrMhi gmzhuo.
12:02.59ao2hi gmzhuo
12:03.16WyrMao2: maybe if you comment about this failure to build on other platforms...
12:03.39WyrMthis would at least bump the thread on people mail clients ;)
12:05.36ao2WyrM, I was thinking something similar... :)
12:06.20WyrMheh, I prefer comments that requires me to rewrite everything than no comments at all ;)
12:07.48ao2A single patch is easy to be lost, series get more visibility I guess.
12:08.28WyrMbut I think series is a waste of time.
12:08.44WyrMeg, we have acks from Dmitry Torokhov and Alessandro Zumo.
12:08.52WyrMand we will need to get acks again.
12:09.10WyrMI fell like I just made them loose a bunch of time :(
12:09.48WyrMs/fell/feel/
12:11.41ao2I was speaking in general, I agree that most of other stuff in openezx is not ready for submission.
12:11.47WyrMand just to help us, all my msgs to LAKML are held for moderation :(
12:12.44*** join/#openezx pH5_ (n=ph5@e178192176.adsl.alicedsl.de)
12:13.02ao2WyrM, even if you are subscribed?
12:13.12WyrMyes.
12:13.18WyrM"The message headers matched a filter rule"
12:13.50ao2reboot, bbl
13:07.03*** join/#openezx ao2 (n=u@2001:1418:117:0:0:0:0:1)
13:36.01CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * r6540795296 10openezx/arch/arm/mach-pxa/ezx.c: pass irq_base to pcap driver
13:36.01CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * r8bb7d0f7b5 10openezx/ (drivers/mfd/ezx-pcap.c include/linux/mfd/ezx-pcap.h):
13:36.01CIA-33OpenEZX: use irq_base on platform_data, to avoid build failure on other
13:36.01CIA-33OpenEZX: platforms.
13:36.48WyrMsubdrivers needs to be updated too.
13:43.17ao2WyrM, but the irqs definition is still in mach-pxa, right? Will this still build on !ARM ?
13:43.38WyrMao2: the core driver, yes.
13:43.48WyrMit doesn't need the irq definitions anymore.
13:44.06ao2ok, now I get it
13:44.06WyrMthe subdrivers will also need to get irqs from pdata (or presource).
13:44.18ao2it will derive the numbers from irq_base
13:44.24WyrMits not fixed anymore.
13:44.26WyrMyes.
13:45.18WyrMim wondering if I should export something on ezx-pcap.c, so drivers know irq_base.
13:45.32WyrMor if I should set presource for each subdriver.
13:45.43ao2but what about the IRQ names? Will you move them to ezx-pcap.h?
13:45.47WyrMthe offset regarding irq_base is fixed.
13:46.10WyrMao2: yes, will be moved to ezx-pcap.h
13:46.47ao2ok, and don't the subdriver just need to know names?
13:46.57WyrMao2: no.
13:47.10WyrMbecause irq_base is different on each architecture.
13:47.20WyrMand ezx-pcap.h doesn't know irq_base.
13:48.00WyrMso, PCAP_IRQ_* definitions will be an offset of irq_base.
13:48.13WyrMs/of/from/
13:48.28ao2ok, so something like PCAP_IRQ(TS) macto is needed to pass irq, that adjust the numeber according to irq_base
13:48.43WyrMbut this can't be a macro.
13:48.43ao2s/macto/macro/
13:49.02WyrMezx-pcap.h doesn't know the irq_base.
13:49.12ao2right, again, sorry
13:49.14WyrMonly ezx-pcap.c
13:49.27gmzhuoany architecture have PCAP should define PCAP_IRQ_BASE,so the ezx-pcap.h should get PCAP_IRQ_BASE
13:50.10WyrMgmzhuo: yes, this makes sense. The problem is that any architecture that doesn't define it will cause a build error.
13:50.42WyrMreally there are only 2 solutions.
13:50.56WyrMset irq with presource for each subdriver.
13:51.18WyrMor export a function on ezx-pcap.c to tell the subdrivers what irq_base it is using.
13:51.54gmzhuoif the architecture doesn't define it,that's mean's it's no pcap,and no pcap subdevice?
13:52.06WyrMgmzhuo: yes, I agree.
13:52.20WyrMbut people want pcap driver to be generic and build on all platforms.
13:52.21WyrM:/
13:53.52WyrMI think I will:
13:53.52WyrMunsigned int pcap_get_irq(int offset) {
13:53.52WyrM<PROTECTED>
13:53.52WyrM}
13:53.52WyrMEXPORT_SYMBOL_GPL(pcap_get_irq);
13:54.20WyrMand drivers call request_irq(pcap_get_irq(PCAP_IRQ_TS), .......
13:54.25*** join/#openezx sabrod (n=sab@82.234.143.184)
13:55.06WyrMbut I will name it pcap2irq. ;)
13:55.08WyrMas it is the analogous to irq2pcap()
13:55.47WyrMI think its better than using a platform_resource for each subdriver.
13:57.27WyrMno, not a good name.
13:57.35WyrMpcap_irq() is better I think.
14:10.42CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * r2cb7b19b29 10openezx/arch/arm/mach-pxa/ezx.c: PCAP_IRQ() no longer exists, use IRQ_BOARD_START instead
14:10.48WyrMok, this is enough so subdrivers can know their own irqs.
14:10.51CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * r9351db3ae6 10openezx/ (3 files in 3 dirs):
14:10.51CIA-33OpenEZX: add a function on ezx-pcap.c so subdrivers know their irqs.
14:10.51CIA-33OpenEZX: remove PCAP_IRQ() macro, move PCAP_IRQ_* macros to ezx-pcap.h as offsets
14:10.51CIA-33OpenEZX: from irq_base.
14:11.36ao2ok, thanks
14:13.36gmzhuono pcap_irq function is so good!
14:13.51WyrMgmzhuo: what you mean?
14:14.04WyrMrestricting platform build is not good.
14:15.12gmzhuoI think IRQ_BOARD_START or IRQ_BOARD_END is better
14:15.28gmzhuothan pcap_irq function
14:15.54*** join/#openezx marex (n=marex@vasut.kolej.mff.cuni.cz)
14:16.41WyrMbut IRQ_BOARD_START is pxa stuff.
14:16.55WyrMthis would make our driver fail build on other platforms.
14:17.43gmzhuoany other irq chip example?
14:17.49WyrMyes.
14:17.55WyrMtwl4030.
14:18.04WyrMbut it uses presource for all subdrivers.
14:18.16WyrMand keeps all subdrivers inside its core.
14:18.32WyrMs/subdrivers/subdrivers resources/
14:19.01WyrMsee twl4030-core.c, twl4030-irq.c, and rtc-twl4030.c
14:19.32CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * r0582305010 10openezx/drivers/input/keyboard/pcap_keys.c: use pcap_irq(), fix pstat checking, PCAP_IRQ_* is not bitmask
14:19.33CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * rddfb30b322 10openezx/drivers/rtc/rtc-pcap.c: use pcap_irq()
14:30.06WyrMbuild error on ezx.c
14:30.17WyrMI guess we can't use IRQ_BOARD_START.
14:30.23WyrMor it is redefined somewhere?
14:31.42WyrMerr
14:31.48WyrMsorry, stupid typo.
14:35.07CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * rf970f0fedb 10openezx/arch/arm/mach-pxa/ezx.c: fix typo
14:38.28gmzhuoezx/mach/isp1583 is broken :(
14:39.16gmzhuoHow can I use the MOBPORT irq for ISP1583?
14:40.27WyrMwell, in fact, it shouldn't use pcap irq at all.
14:40.28WyrM:)
14:40.48WyrMbut for now, you can use pcap_irq(PCAP_IRQ_MOBPORTB)
14:42.03gmzhuolet it useable not so correct now,I will resolve it when we have usb transive driver.
14:43.03CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * radf8700e72 10openezx/include/linux/mfd/ezx-pcap.h:
14:43.03CIA-33OpenEZX: fix:
14:43.03CIA-33OpenEZX: drivers/input/keyboard/pcap_keys.c:65: error: implicit declaration of
14:43.03CIA-33OpenEZX: function ‘pcap_irq’
14:44.57WyrMok, now it builds, no warnings :)
14:51.35CIA-33OpenEZX: 03guiming zhuo <gmzhuo@gbsoft.localdomain> 07ezx/current * r5296b4d2fd 10openezx/arch/arm/mach-pxa/ezx.c: fix isp1583 bus check
15:03.20*** join/#openezx flo_lap (n=fuchs@f049038042.adsl.alicedsl.de)
15:17.56*** join/#openezx johanbr (n=j@JBrannlund2.MathStat.Dal.Ca)
16:15.35*** join/#openezx gmzhuo-- (n=gmzhuo@125.70.86.27)
16:41.08CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * r90bba2c1d3 10openezx/ (.topdeps .topmsg): create
16:41.11CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * ref70ca7d54 10openezx/.topdeps: add dependency on ezx/increase-nrirqs
16:41.11CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * r698036f1d8 10openezx/arch/arm/mach-pxa/include/mach/irqs.h: move irqs.h change to a separate patch
16:41.12CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * r17ca13982f 10openezx/arch/arm/mach-pxa/include/mach/irqs.h: fix typo
16:41.13CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * r07edf1d4d5 10openezx/arch/arm/mach-pxa/Kconfig: not needed.
17:08.12*** join/#openezx sabrod (n=sab@mic92-8-82-234-143-184.fbx.proxad.net)
17:10.34WyrMhi sabrod
17:10.38WyrMlong time no see :)
17:33.05*** join/#openezx sabro1 (n=sab@mic92-8-82-234-143-184.fbx.proxad.net)
18:12.53*** join/#openezx l403 (n=l@85.132.159.239)
19:02.54*** join/#openezx Dominus (i=unices@82-170-225-106.ip.telfort.nl)
19:05.48*** join/#openezx SuN|2 (i=unices@82-170-225-106.ip.telfort.nl)
19:19.46CIA-33OpenEZX: 03Daniel Ribeiro <drwyrm@gmail.com> 07ezx/current * rd29c8a1c74 10openezx/drivers/mfd/ezx-pcap.c: fix irq release on remove()
20:09.48*** join/#openezx flo_lap (n=fuchs@f049038042.adsl.alicedsl.de)
21:49.25ao2WyrM, did you leave out LAK in v2 PCAP mail on purpose?
21:49.36WyrMyeah.
21:49.43WyrMno more arm/pxa bits in it.
22:18.42*** join/#openezx l403 (n=l@85.132.159.239)
22:23.54*** join/#openezx sabrod (n=sab@mic92-8-82-234-143-184.fbx.proxad.net)
22:27.12*** join/#openezx WyrM1 (n=wyrm@189.13.234.34)
22:29.27WyrMao2: the worst is that I found some more mistakes on the code.
22:29.28WyrMheh
22:30.00WyrMwill not send again, I will wait at least 5 days before sending another version now.

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.