Closed arfvidsonUIS closed 7 years ago
Branch has been pushed to githubissue-83 and is ready for review
Just new branch that only modifies the controlvmchannel.h pound defines. Branch name is: githubissue-83-v2. Please review my work. I'll continue working on removing the rest of pound defines so I will be adding additional patches to this branch and will update the github ticket. 💃 :D
Can you squash the last one with the reserved error messages with the first one that removes them? I would like to keep that the same patch so we don't have any time when people don't know they are reserved Also in the fix comments can you get rid of references to SWITCH_ messages ?
Erik, are you still in question mode with this one or is there a v3? Also, I'm still seeing SWITCH_ messages, are you planning on doing that as a separate patch or should we create a new issue for it?
These commits to githubissue-83 branch:
were committed by Greg to staging-next on 9/12 as:
Note: I am leaving this issue open, as I believe the above commits are only PART of the fix for this issue.
From what I see in the branches githubissue-83-v2 branch and githubissue-83-v3 branch, it appears as if these patches are also part of this issue. Greg accepted these into staging-next on 9/12 as:
I'm leaving this issue open because I'm not sure if other patches are needed.
Yeah, I'm just cleaning thing ups will be pushing a new branch soon. I've been working on a Presentation and trac and having gotten around to completing this yet. Please don't close this issue yet. Thanks for the follow up
just pushed the branch to https://github.com/davidker/unisys/commits/githubissue-83-v3. Ready to be inspected
Got the updated branch and the patches are no long complaining about checkpatch warnings. The unused #define count went from 57 to 32. I'll submit these patches, but do you consider the story done yet?
updated branch githubissue-83-v3 and ready for review
check by running ./test_usage.sh 2>/dev/null | cut -d -f 1 | xargs -I '{}' grep -r -n '{}' . I left maxnum on purpose doesn't make sense to have maxmin without a the maxnum.
Erik, what about MAXNUM in iochannel.h? ffee0261160901b1c2a93177c02f9ebecb624fc3
I think that one gets touched as well, I'm also seeing about 11 messages in controlvmchannel.h, am I missing something or are they intended to stay?
kershnda@USTR-ERL-4458[2933.master]:~/kernel/master/unisys/drivers/staging/unisys$ ~/bin/test_usage.sh 2>/dev/null | wc -l 12 kershnda@USTR-ERL-4458[2933.master]:~/kernel/master/unisys/drivers/staging/unisys$ ~/bin/test_usage.sh 2>/dev/null CONTROLVM_RESP_ERROR_CHANNEL_INVALID not used, found in ./visorbus/controlvmchannel.h? CONTROLVM_RESP_ERROR_VIRTPCI_DRIVER_FAILURE not used, found in ./visorbus/controlvmchannel.h? CONTROLVM_RESP_ERROR_VIRTPCI_DRIVER_CALLBACK_ERROR not used, found in ./visorbus/controlvmchannel.h? CONTROLVM_RESP_ERROR_GENERIC_DRIVER_CALLBACK_ERROR not used, found in ./visorbus/controlvmchannel.h? CONTROLVM_RESP_ERROR_BUS_DEVICE_ATTACHED not used, found in ./visorbus/controlvmchannel.h? CONTROLVM_RESP_ERROR_CHANNEL_TYPE_UNKNOWN not used, found in ./visorbus/controlvmchannel.h? CONTROLVM_RESP_ERROR_CHANNEL_SIZE_TOO_SMALL not used, found in ./visorbus/controlvmchannel.h? CONTROLVM_RESP_ERROR_CHIPSET_SHUTDOWN_FAILED not used, found in ./visorbus/controlvmchannel.h? CONTROLVM_RESP_ERROR_CHIPSET_SHUTDOWN_ALREADY_ACTIVE not used, found in ./visorbus/controlvmchannel.h? CONTROLVM_RESP_ERROR_CHIPSET_STOP_FAILED_BUS not used, found in ./visorbus/controlvmchannel.h? CONTROLVM_RESP_ERROR_CHIPSET_STOP_FAILED_SWITCH not used, found in ./visorbus/controlvmchannel.h? MAXNUM not used, found in ./include/iochannel.h?
Update githubissue-83-v3 branch with the additional pound defines that needed to be removed
I also added min_t macro kernel.h so that MINNUM macro could be removed
On another note iochannel.h comments are inconsistent. ie /*
when most of them are /*paragraph
copyright is at 2013.
Some paragraph start with Uppercase when other do not.
I know I'm nitpicking and is a quick fix. Let me know if you think any of this are necessary to fix and I'll create githubissue. Thing is that if we start fixing this little things we might uncover additional nitpicks I'm not sure if it is worth our time to pursue this.
BTW added feature to your script so we get the linenumber of the unused pound defines ./test_usage.sh 2>/dev/null | cut -d -f 1 | xargs -I '{}' grep -r -n '{}' .
Thanks for the addition to the script, I'll update it as well.
Go ahead and write the KanBoard issues for the nitpicks, but make them lower priority than Greg's emails. I would like to have a full list of all the nits that we know of and can address. I'm not sure what happens with the copyrights on the files since we are in the upstream, I wonder if we should email Jes about that nit.
I've updated the script to include the line number. It can found in the svn repo maintainer_scripts/bin.
Fails to compile, pulling the patches and just sending series up as is.
n file included from ../drivers/staging/unisys/visornic/visornic_main.c:29:0: ../drivers/staging/unisys/include/iochannel.h: In function 'add_physinfo_entries': ../drivers/staging/unisys/include/iochannel.h:536:41: error: macro "min_t" requires 3 arguments, but only 2 given (u16)min_t(len, (u32)PI_PAGE_SIZE); ^ ../drivers/staging/unisys/include/iochannel.h:536:13: error: 'min_t' undeclared (first use in this function) (u16)min_t(len, (u32)PI_PAGE_SIZE); ^ ../drivers/staging/unisys/include/iochannel.h:536:13: note: each undeclared identifier is reported only once for each function it appears in make[5]: * [drivers/staging/unisys/visornic/visornic_main.o] Error 1 make[4]: * [drivers/staging/unisys/visornic] Error 2 make[4]: * Waiting for unfinished jobs.... LD kernel/built-in.o In file included from ../drivers/staging/unisys/visorhba/visorhba_main.c:27:0: ../drivers/staging/unisys/include/iochannel.h: In function 'add_physinfo_entries': ../drivers/staging/unisys/include/iochannel.h:536:41: error: macro "min_t" requires 3 arguments, but only 2 given (u16)min_t(len, (u32)PI_PAGE_SIZE); ^ ../drivers/staging/unisys/include/iochannel.h:536:13: error: 'min_t' undeclared (first use in this function) (u16)min_t(len, (u32)PI_PAGE_SIZE); ^ ../drivers/staging/unisys/include/iochannel.h:536:13: note: each undeclared identifier is reported only once for each function it appears in LD drivers/base/built-in.o make[5]: * [drivers/staging/unisys/visorhba/visorhba_main.o] Error 1 make[4]: * [drivers/staging/unisys/visorhba] Error 2 make[3]: * [drivers/staging/unisys] Error 2 make[2]: * [drivers/staging] Error 2 make[2]: * Waiting for unfinished jobs.... make[1]: * [drivers] Error 2 make[1]: Leaving directory `/home/kershnda/kernel/master/unisys/.build' make: * [sub-make] Error 2 FAILED make
UPDATED: WITH FIX
just remove offset in remove unused pound defines fix comments to 72 look at channel.h and iochannel.h spar firmware pound defines leave everything get rid off
I'm closing this issue. It has been mostly done by Erik's patches and the 2 Code-A-Thons we have done.
Greg's comments