On Tue, 06 Jan 2009 17:49:04 +0200 Boaz Harrosh <bharrosh@xxxxxxxxxxx> wrote: > FUJITA Tomonori wrote: > > On Tue, 06 Jan 2009 16:56:23 +0200 > > Boaz Harrosh <bharrosh@xxxxxxxxxxx> wrote: > > > >> FUJITA Tomonori wrote: > >>> On Tue, 6 Jan 2009 16:17:16 +0200 > >>> Why do we need in-kernel test suite? > >> It must be in kernel, because it tests the in-kernel library. > >> However it could be in it's own Module, which makes it more > >> complicated, but doable. > > > > Why do we need the code to tests the in-kernel library? > > > > I expect that you guys have tested the osd in-kernel library heavily > > and will. I don't think people want to do the same. Then why do we > > need to have the test suite in mainline? > > > > one - There are things that can regress. Lots of stuff in current > kernel are only used by OSD, like bidi and varlen. all the way down > to LLDs. So this should be a regression test for future kernels. SMP has been using BIDI support. FC will use it too soon, I guess. If you want to test bidi and varlen, you can just send SCSI commands via bsg. That's what I did. I don't see how it can be a reason why we need the in-kernel test suite. > two - To test compatibility, some thing is broken, the test can help > pinpoint the problems. So it's a debugging tool. > > Sure, the lib itself is heavily tested but this is to test the > round trip. I think that it's fine if people want a debugging tool. I've not seen any developer who say that. IMHO, it's better to have only useful debugging stuff in mainline for many people like scsi_debug. -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
