32bit tasks C/R

Revision as of 15:08, 19 January 2017 by Dsafonov (talk | contribs) (→‎List of failed tests: socket-tcp-last-ack: lucky 64-bit)


Compatible applications

On x86_64 there are two types of compatible applications:

  • ia32 - compiled to run on i686 target, can be executed on x86_64 with IA32_EMULATION config option set.
  • x32 - specially compiled binaries to run on x86_64 machine with CONFIG_X86_X32 config option set.

Both of them uses 4 byte pointers thus can address no more than 4Gb of virtual memory.
But x32 uses full 64-bit register set (and thus can't be launched on i686 host natively).
Both of them requires additional environment on x86_64 as Glibc, libraries, and compiler support.
x32 is rarely distributed (at this moment only Debian x32 port can be easily found).
So, CRIU will support ia32 C/R at this moment, x32 support may be quite easily added on top of ia32 as needed patches have already added in kernel with ia32 C/R support.
The following text uses compatible and 32-bit in the meaning of ia32 applications unless otherwise specified.

Difference between native and compat applications

From the CPU's point of view, 32-bit compatibility mode applications differ to 64-bit application by current CS (code segment selector): if corresponding value of L-bit from flags of entry in descriptors table is set the CPU will be in 64-bit mode when this segment descriptor is being used. There are some other differences between 32 and 64-bit selectors, one can read about them in the article "The 0x33 Segment Selector (Heavens Gate)". Code selectors for both bits are defined in kernel headers as __USER32_CS and __USER_CS and corresponds to descriptors in GDT (Global Descriptors Table). One can change 64-bit mode to compatibility mode by swapping CS value (e.g., with longjump).

From the Linux kernel's point of view, applications differ by values set during exec of application such as mmap_base or thread info flags TIF_ADDR32/TIF_IA32/TIF_X32. Both native and compat applications can do 32 or 64-bit syscalls.

Approaches to C/R compatible applications

C/R of compatible applications can be done differently, this section describes cons/pros of each, to address decision why C/R of 32-bit tasks done that way and not some other.

Restore with exec() of 32-bit dummy binary vs from 64-bit CRIU

Restore of 32-bit application can be done with some daemon that runs in 32-bit mode and communicates with CRIU binary (or 32-bit CRIU subprocess).

Pros:

  • no kernel patches expected (not quite true: vDSO mremap() still needed support)

Cons:

  • CRIU code base does not have special restore daemon to communicate with - code needs to be reworked
  • 64-bit app can have 32-bit child, which could be a parent to 64-bit and so on - need to re-exec native 64-bit CRIU from 32-bit dummy (or 32-bit CRIU)
  • need to send to the daemon properties of restoring processes, open fds to images, share memory with parsed ps_tree and so on... The number of IPC calls will slow down restore
  • restoring becomes more complicated, and if looking forward to restoring user/pid sub-namespaces, it will be too entangled
  • no optimized inheritance for task's properties those erase with exec()
  • will need also another daemon for x32

Restore with a flag to sigreturn() or arch_prctl()

The initial attempt to do 32-bit C/R, was rejected by lkml community by many reasons. It should have swapped thread info flags (such as TIF_ADDR32/TIF_IA32/TIF_X32), unmap native 64-bit vDSO blob from process's address space and map compatible 32-bit vDSO - all according to some bit in sigframe in rt_sigreturn() call or some dedicated for it arch_prctl() call.

Pros:

  • Simple from the point of CRIU: just do sigreturn with a new bit set or call arch_prctl() and do sigreturn

Cons:

  • If 32-bit vDSO image on restored host differ from dumped (in image), need to catch task after sigreturn and make jump trampolines separately - in case of arch_prctl() simpler (that's why arch_prctl was in initial RFC)
  • Too many points of failure for one syscall, too complicated
  • Just adding a way to swap those thread info flags from userspace would result in a new races/bugs (as e.g., TASK_SIZE macro depends on TIF_ADDR32, the mmap code may do unexpected things)

After discussion in lkml, conclusion was: separate changing personality (like thread info flags) from API to map vDSO blobs, remove TIF_IA32 flag that differs 32 from 64-bit tasks and look on syscall's nature: compat syscall, x32 syscall or native syscall.

Seizing with two 32-bit and 64-bit parasites

Pros:

  • no 32-bit calls in 64-bit parasite and vice-versa
  • no need in exit in parasite: ptrace code doesn't allow to set 32-bit regset to 64-bit task and the reverse, running parasite the same nature as task bereaves us from those limits

Cons:

  • need to have two/three (for x32 also) blobs for seizing
  • macros in makefiles to build two parasites
  • serialization of parasite's answers: arguments to parasite differ in size - serialize them, which added not nice-looking and less readable C macros

Current approach

Needs to be done (TODO)

Error dump on x32-bit app dumping

At this moment we'll support only compat ia32 applications, attempt to dump x32 compat binary should result in error.

Bug with mmaping over 4Gb

As 32-bit application is restored from 64-bit CRIU, some task's properties that were filled on exec() are left, which is quite unusual for 32-bit task. One of the things, left from 64-bit binary is precalculated mmap_base which is used to find task's top/bottom address limit during mmap() syscall. That means that compat sys_mmap() may map page over 4Gb address and return 4-byte pointer to low bytes of address. Looks like no one has used compatible mmap in 64-bit binary. Results in broken mmap in restored 32-bit application, which can map vma over 4Gb.

Patches to fix this bug at this moment were posted on lkml, but not yet accepted. See Upstream kernel commits. If they will not go to v4.9-stable, the kerndat test for 32-bit C/R will be reworked to check if the bug present in kernel (which is not nice thing, but ok).

List of failed tests

The table is being kept up-to-date by User:Dsafonov with latest kernel/CRIU patches in his environment, some of which may be yet not in tree or even yet not sent.

Name Fail reason
fpu01 no ia32 version
sse00 no ia32 version
futex-rl sys_get_robust_list() should be compat syscall for 32-bit tasks: kernel keeps two different lists: robust_list and compat_robust_list in task_struct
rtc no 32-bit version of rtc test library
vdso01 no ia32 version
file_locks08 use struct flock64
file_locks07
file_locks06
file_locks00
socket-tcp6-last-ack (incorrect test) uninitialized write for var on stack, looks like worked on x86_64 by luck
socket-tcp-last-ack
sse20 ?
autofs ?
sigpending ?
fpu00 ?
mmx00 ?

Fixes for older kernels

For kernels with backported mainline patches for 32-bit C/R (like vzkernel) there are a couple of things to do like different sizes of vdso/vvar (or vvar may not be even present).

Fault-inject test for vDSO trampolines

Should ensure that they work. Need to be done for both native/compat C/R.

Kernel patch for vsyscall page

That's emulated page, not a vma - affects only in /proc/<pid>/maps for restored process. Depends on !TIF_IA32 && !TIF_X32 - Andy got patches for disabling the emulation on per-pid basics, for now I ran tests with vsyscall=none boot parameter because zdtm.py checks maps before/after C/R.