Path resolution bug in klocalizer can lead to AssertionError

[lolrepeatlol]

Description

• Depending on the commits provided, klocalizer can attempt to look for a non-existent file path, which causes an AssertionError in Python.
• It is likely that this AssertionError is due to a path resolution error.

Steps to reproduce

Steps followed

To get a repaired configuration file for a commit range, I followed the steps here:

1. Clone the Linux kernel with git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git and enter the directory.
2. Create a diff from a range of patches using git diff {commit1}..{commit2} > patchset.diff. I initially encountered this issue with the range 541b80216cd1d511841c3c8d9559a7b13f4f79f2 to 0df2ac59bebfac221463ef57ed3554899b41d75f.
3. Check out the source code at the latest of the two commits and create a kernel configuration file with a command like make allnoconfig.
4. Run klocalizer with klocalizer --repair .config -a x86_64 --include-mutex patchset.diff.

   What I expected to happen

5. I expected klocalizer to repair the kernel configuration file.

   What actually happened

6. klocalizer runs into an AssertionError:

   WARNING: No cached formulas for e98bee58dd93 available for download :(
   Traceback (most recent call last):
   File "/home/alexei/.local/bin/klocalizer", line 1726, in <module>
   klocalizerCLI()
   File "/home/alexei/.local/bin/klocalizer", line 810, in klocalizerCLI
   assert os.path.isfile(srcfile_fullpath)
   AssertionError
   alexei@turing:~/LinuxKernels/hybridtest/linuxhybrid$

Additional details

• I applied the following debug code to the klocalizerCLI file around line 810:

  global_scope_lines = {} # { unit : set(int) }, lines outside any ifdefs
  for target_list in [include_list, include_mutex_list, exclude_list]:
  for i, (unit, line_list) in enumerate(target_list):
    if line_list:
      srcfile = unit2srcfile(unit)
      srcfile_fullpath = os.path.join(linux_ksrc, srcfile)
      print(f"DEBUG: srcfile: {srcfile}, srcfile_fullpath: {srcfile_fullpath}. About to assert os.path.isfile...")  # debugging line
      assert os.path.isfile(srcfile_fullpath)  # affected line

  which produced this debugging output:

  DEBUG: srcfile: net/mptcp/pm_netlink.c, srofile_fullpath: ./net/mptcp/pm_netlink.c. About to assert os.path.isfile...
  DEBUG: srcfile: net/sctp/input.c, srcfile_fullpath: ./net/sctp/input.c. About to assert os.path.isfile...
  DEBUG: srcfile: sound/core/seq/seq_ump_convert.c, srcfile_fullpath: ./sound/core/seq/seq_ump_convert.c. About to assert os.path.isfile...
  DEBUG: srcfile: sound/core/ump_convert.c, srofile_fullpath: ./sound/core/ump_convert.c. About to assert os.path.isfile...
  DEBUG: srcfile: sound/firewire/amdtp-stream.c, srofile_fullpath: ./sound/firewire/amdtp-stream.c. About to assert os.path.isfile...
  DEBUG: srcfile: sound/pci/hda/hda_generic.c, srcfile_fullpath: ./sound/pci/hda/hda_generic.c. About to assert os.path.isfile...
  DEBUG: srcfile: sound/pci/hda/hda_intel.c, srcfile_fullpath: ./sound/pci/hda/hda_intel.c. About to assert os.path.isfile...
  DEBUG: srcfile: sound/pci/hda/patch_conexant.c, srcfile_fullpath: ./sound/pci/hda/patch_conexant.c. About to assert os.path.isfile...
  DEBUG: srcfile: sound/pci/hda/patch_realtek.c, srofile_fullpath: ./sound/pci/hda/patch_realtek.c. About to assert os.path.isfile...
  DEBUG: srcfile: sound/usb/stream.c, srcfile_fullpath: ./sound/usb/stream.c. About to assert os path.isfile...
  DEBUG: srcfile: arch/x86/kvm/../../../virt/kvmguest_memfd.c, srcfile_fullpath: ./arch/x86/kvm/../../../virt/kvmguest_memfd.c. About to assert os.path.isfile...
  Traceback (most recent call last):
  File "/home/alexei/.local/bin/klocalizer", line 1728, in <module>
  klocalizerCLI()
  File "/home/alexei/.local/bin/klocalizer", line 811, in klocalizerCLI
  assert os.path.isfile(srcfile_fullpath) # affected line
  AssertionError
  alexei@turing:~/LinuxKernels/hybridtest/linuxhybrid$

• I noticed that /arch/x86/kvm/../../../virt/kvmguest_memfd.c does not exist and this file can't be found either as a compiled unit file, or a C file, in the Linux kernel, even when many KVM and virtualization options are turned on in the kernel build.

• The closest file in the provided diff was /virt/kvm/guest_memfd.c.

[lolrepeatlol]

Addressed by the professor in #279.