On Thu, Feb 25, 2010 at 1:33 AM, Rishikesh K Rajak <[email protected]> wrote: > On Thu, Feb 25, 2010 at 02:39:37PM +0530, Silesh C V wrote: >> >> >> >> The end goal is to have these tests fit directly into LTP without >> >> having to fudge around writing a secondary driver, thus maintaining >> >> two pieces of dependent code [which is of course more complicated than >> >> one piece of code], and thus is more of a pain to maintain longterm. >> >> There's a lot of code in the repository like this that was ported and >> >> subsequently not properly adapted to the existing infrastructure, thus >> >> creating additional headache for maintainers and end-users. >> > >> > Ok -- I take that back. I have no idea wtf is going on in this >> > directory nor the ultimate intent of the tests in here, apart from >> > being a lot of ad hoc pain in the arse-ness. >> >> The README in testcases/kernel/device-drivers/ says that these tests >> should not be run with the other tests, and they have to be built >> separately. I agree that this is because they have a kernel space part >> also. But this prevents us from building device_driver tests that >> have only user space part (as in RTC tests), along with the overall >> build. And these tests does not belong in any other directory other >> than kernel/device-drivers. So as long as other tests are there we >> will have to live with building new device driver tests separately.I >> can re-send the patch with Garrett's suggestions incorporated. >> >> Subrata, suggestions ? > > Hi Silesh, > > How about porting this on new makefile infrastructure ? I think this is > time where we can look back and can make one piece of code as garret > said is correct. Please see your fesibility. Let me know if you still > want this to be integrated. Before that can you please little more > descriptive with your patch e,g: what it is doing. > > And also it is always good if you include more documentation for each > function as coding style suggests.
Regardless of whether or not this can fit within the new makefile infrastructure -- please at least consider adding in the bits to use tst_res(3). It will make maintenance considerably easier in the long run. I can port a Makefile over in < 30 mins -- it's the C code that I'd have to fumble around with as I'm not the original developer, I may not understand all the nuances of what's trying to be tested, etc. More documentation would be extremely helpful, in particular pointers to additional reference material so that someone who has _zero_ understanding in how the rtc driver typically works in Linux can at least look up additional materials to develop a better understanding of what's going on... >> > >> > I've learned enough to keep my hands off this because it looks like a mess. Thanks, -Garrett ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Ltp-list mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ltp-list
