Alexander Graf <ag...@suse.de> writes: > On 30.10.2011, at 15:30, Stefan Weil wrote: > >> Am 30.10.2011 14:41, schrieb Alexander Graf: >>> On 30.10.2011, at 13:07, Stefan Weil wrote: >>>> Valgrind is a tool which can automatically detect many kinds of bugs. >>>> >>>> Running QEMU on Valgrind with x86_64 hosts was not possible because >>>> Valgrind aborts when memalign is called with an alignment larger than >>>> 1 MiB. QEMU normally uses 2 MiB on Linux x86_64. >>>> >>>> Now the alignment is reduced to the page size when QEMU is running on >>>> Valgrind. >>>> >>>> valgrind.h is a copy from Valgrind svn trunk r12226 with trailing >>>> whitespace stripped but otherwise unmodified, so it still raises lots >>>> of errors when checked with scripts/checkpatch.pl. >>> >>> Can't we just require valgrind header files to be around when kvm is >>> enabled? I would rather not copy code from other projects. Alternatively >>> you could take the header and shrink it down to maybe 5 lines of inline asm >>> code that would be a lot more readable :). You're #ifdef'ing on x86_64 >>> already anyways. >> >> The patch is currently required for x86_64 hosts running Linux. >> I estimate that this is one of the most frequently used QEMU host platforms, >> and in most cases, KVM will be configured because this is the default >> and also because it is reasonable for this platform. >> >> How many of these hosts will have the Valgrind header around? >> I estimate less than 20 %, so configure would have to test whether >> valgrind.h is available or not. I think providing valgrind.h is >> a much better (and simpler) solution. > > Hrm. I see your point. > >> Stripping valgrind.h is not a good idea: the file is specially designed >> to be included in other projects like QEMU. As soon as the 2 MiB alignment >> is used for other hosts (ppc64, ...), you would have to take more and more >> from the original code. The file was not designed to be readable. >> Although it contains lots of comments which improve readability, >> there remains code which is less easy to read. I cite one of those >> comments: >> >> /* The following defines the magic code sequences which the JITter >> spots and handles magically. Don't look too closely at them as >> they will rot your brain. >> >> Instead of rotting my brain, I prefer using a copy of the original code.
Yes, copies are evil. > Could we maybe use a git submodule to point to the valgrind repo and fetch it > from there? Anyone sophisticated enough to make use of valgrind should be able to install valgrind.h just fine. It's not rocket science: # yum provides \*/valgrind.h [...] # yum install valgrind-devel [...]
