dibakar gope wrote: > Hi All, > > I have few queries regarding the regression tests for X86. > > (1) I could build the x86 in FS mode (I am using a bunch of > x86-specific patches from > http://www.csl.cornell.edu/~vince/projects/m5/m5_x86_64_se_status.html). > I guess that the pre-compiled linux kernel (that can be downloaded > from M5 site) was complied for alpha arch only. So I actually > downloaded the linux-dist tarball from M5 site for x86 build. This > tarball has a .config.M5 that can be used for compiling the kernel, > but that .config.m5 is ALPHA-specific. > > So in order to compile the linux for x86, I used the config of my > native linux machine kernel as a basis for our x86 config kernel and > got the vmlinux for x86. Following commands are used for that:- > > cp /boot/config ./.config > make menuconfig > make-kpkg clean > fakeroot make-kpkg --initrd --append-to-version=-v2.6.27 kernel_image > kernel_headers > > I used that vmliux for X86_FS build and did not get any error during > the build process. > > So my query is, are there any x86-specific patches (configurations) > that I should have considered for compiling the linux kernel for x86. >
No, or at least not that I'm aware of. It wouldn't surprise me if you found a minor bug in M5 with certain configurations or possibly got the kernel to use some feature of x86 that's not implemented, but with a "normal" configuration you should be ok. As far as the patches from that webpage, they're valid fixes of real issues to the best of my knowledge, but most of them only matter in SE mode. The three that would matter in FS are the fsincos, fnstsw, and haddps related ones, and those instructions won't be important during boot. They could quite easily used by the workload you run once you've booted, though. M5 started as an Alpha only simulator, and the other ISAs were added later. They've historically always been added to SE mode first and eventually work their way into FS, so the FS specific portions of M5, especially peripheral parts like kernel configurations or precompiled kernels, are going to tend to still be Alpha centric. > > (2)Then I tried to test that X86_FS m5.opt using regression tests. All > the several test programs present for the regression tests have > config.ini files only for alpha in m5-dev tarball, but they don't have > the same for X86. But using the following command, I can generate > those x86-specific config-ini for the test-programs used in FS mode > regression. > > build/X86_FS/m5.opt -re configs/example/fs.py > --cmd=tests/test-progs/<test program name>/bin/x86/<test program > binary> > > But the problem is that the m5-dev tarball (m5/tests/test-progs/*) > does not have the test program binaries (m5/tests/quick/*) (for > example, 10.linux-boot,80.netperf-stream,50.memtest etc) except hello > (which is not used for FS mode regression). So I could not generate > the config.ini for x86 in order to run the regression tests. > > So my query is, have anyone worked on the X86 regression tests / faced > the same problem? Before I use the x86_FS.opt for SPEC2000/2006 > benchmarks, I want that to pass the regression tests first. > X86_FS doesn't have any regressions at all right now, something I'm sure Nate at least is anxious for me to fix. If you want to test it out, try booting a kernel. If it boots, it works :-). I can send you some configs that I used for a few versions of the kernel, which I should probably put on the wiki somewhere (I notice there are a lot of things I've been meaning to get to but haven't) The --cmd option doesn't work in FS mode. If you want to run a particular benchmark, the prescribed way is to write an init script and feed that into the kernel once it's booted using special backdoors built into M5 and the kernel. It would get everything set up and run your benchmark for you from within the simulation itself. Unfortunately, while some of the pieces that would let you do that in X86_FS are there, they aren't all there so that probably won't work. You could log into the simulated system and run the commands yourself, although that introduces a lot of variability that you'd want to avoid. You could also possibly put an actual init script on the disk image directly and run that, but then you'd have to modify the image for each run. There are a lot of finishing (and maybe not so finishing) touches like this that need to be added to X86_FS for it to be as useful and usable as it can be, and hopefully a lot of those will be coming soon. X86 development in M5 should pick up noticeably in a week or so. > > Thanks and Regards, > > Dibakar Gope > Texas A&M University > _______________________________________________ > m5-users mailing list > [email protected] > http://m5sim.org/cgi-bin/mailman/listinfo/m5-users > _______________________________________________ m5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
