Now that sibling calls aren't supported in ppc64le, change the RHEL 7.5
and 7.6 versions of the new-function tests to disable the sibling call
optimization.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
With the following patch:
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index e008aefc3a9d..7c70e369390d 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -2228,6 +2228,8 @@ static void xs_tcp_shutdown(struct rpc_xprt *xprt)
struct socket *sock = transport->sock;
int skst = transport->inet ? transport->inet->sk_state : TCP_CLOSE;
+ asm("nop");
+
if (sock == NULL)
return;
switch (skst) {
We saw the following panic on a RHEL7.6 kernel:
Unable to handle kernel paging request for data at address 0xd00000000577f390
Faulting instruction address: 0xd000000002e918f4
Oops: Kernel access of bad area, sig: 11 [#1]
SMP NR_CPUS=2048 NUMA pSeries
Modules linked in: kpatch_3_10_0_957_1_3_1_1(OEK) nfsd nfs_acl rpcsec_gss_krb5 auth_rpcgss nfsv4 dns_resolver nfs lockd grace fscache sunrpc virtio_balloon ip_tables xfs libcrc32c virtio_net virtio_console virtio_blk virtio_pci virtio_ring virtio dm_mirror dm_region_hash dm_log dm_mod
CPU: 9 PID: 5961 Comm: kworker/9:1H Kdump: loaded Tainted: G OE K------------ 3.10.0-957.1.3.el7.ppc64le #1
Workqueue: xprtiod xprt_autoclose [sunrpc]
task: c00000000300c3c0 ti: c0000003f1814000 task.ti: c0000003f1814000
NIP: d000000002e918f4 LR: d000000002e57394 CTR: c00000000089d100
REGS: c0000003f1817980 TRAP: 0300 Tainted: G OE K------------ (3.10.0-957.1.3.el7.ppc64le)
MSR: 8000000100009033 <SF,EE,ME,IR,DR,RI,LE> CR: 240f2084 XER: 20000000
CFAR: 000000010bb5270c DAR: d00000000577f390 DSISR: 40000000 SOFTE: 1
GPR00: c00000000000b054 c0000003f1817c00 d00000000579add8 c000000214f0f4d0
GPR04: c0000003fd618200 c0000003fd618200 0000000000000001 0000000000000dc2
GPR08: 0000000000000dc3 0000000000000000 0000000000000000 d00000000577f370
GPR12: c0000003f1814000 c000000007b85100 c00000000012fd88 c0000003f711bb40
GPR16: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
GPR20: 0000000000000001 c0000000013510b0 0000000000000001 fffffffffffffef7
GPR24: 0000000000000000 0000000000000000 0000000000000000 c000000001b60600
GPR28: c000000214f0f000 c000000214f0f4d0 c000000214f0f408 c000000214f0f448
NIP [d000000002e918f4] __rpc_create_common.part.6+0x640/0x533c [sunrpc]
LR [d000000002e57394] xprt_autoclose+0x74/0xe0 [sunrpc]
Call Trace:
[c0000003f1817c00] [c00000000000b054] livepatch_handler+0x30/0x80 (unreliable)
[c0000003f1817c40] [c00000000012333c] process_one_work+0x1dc/0x680
[c0000003f1817ce0] [c000000000123980] worker_thread+0x1a0/0x520
[c0000003f1817d80] [c00000000012fe74] kthread+0xf4/0x100
[c0000003f1817e30] [c00000000000a628] ret_from_kernel_thread+0x5c/0xb4
Instruction dump:
396b4570 f8410018 e98b0020 7d8903a6 4e800420 00000000 73747562 000f49c0
c0000000 3d62fffe 396b4598 f8410018 <e98b0020> 7d8903a6 4e800420 00000000
---[ end trace 98e026b8fa880db7 ]---
The original version of xs_tcp_shutdown() has the following sequence:
0xd000000003cfda44 <xs_tcp_shutdown+148>: addi r1,r1,64
0xd000000003cfda48 <xs_tcp_shutdown+152>: ld r0,16(r1)
0xd000000003cfda4c <xs_tcp_shutdown+156>: ld r29,-24(r1)
0xd000000003cfda50 <xs_tcp_shutdown+160>: ld r30,-16(r1)
0xd000000003cfda54 <xs_tcp_shutdown+164>: ld r31,-8(r1)
0xd000000003cfda58 <xs_tcp_shutdown+168>: mtlr r0
0xd000000003cfda5c <xs_tcp_shutdown+172>: b 0xd000000003cfd768
That is, it restores the stack to the caller's stack frame and then does
a sibling call to the localentry point of xs_reset_transport()). So
when xs_reset_transport() returns, it will return straight to
xs_tcp_shutdown()'s caller (xprt_autoclose).
The patched version of the function has this instead (dumped from a
vmcore):
0xd000000003df0834 <xs_tcp_shutdown+148>: addi r1,r1,64
0xd000000003df0838 <xs_tcp_shutdown+152>: ld r0,16(r1)
0xd000000003df083c <xs_tcp_shutdown+156>: ld r29,-24(r1)
0xd000000003df0840 <xs_tcp_shutdown+160>: ld r30,-16(r1)
0xd000000003df0844 <xs_tcp_shutdown+164>: ld r31,-8(r1)
0xd000000003df0848 <xs_tcp_shutdown+168>: mtlr r0
0xd000000003df084c <xs_tcp_shutdown+172>: b 0xd000000003df0ad0
After restoring the stack, instead of branching directly to
xs_reset_transport(), it (rightfully) branches to a toc stub. A stub is
needed because the function it's branching to is in another module
(branching from the patch module to the sunrpc module).
The stub is:
0xd000000003df0ad0 <xs_tcp_shutdown+816>: addis r11,r2,-1
0xd000000003df0ad4 <xs_tcp_shutdown+820>: addi r11,r11,26328
0xd000000003df0ad8 <xs_tcp_shutdown+824>: std r2,24(r1)
0xd000000003df0adc <xs_tcp_shutdown+828>: ld r12,32(r11)
0xd000000003df0ae0 <xs_tcp_shutdown+832>: mtctr r12
0xd000000003df0ae4 <xs_tcp_shutdown+836>: bctr
And the "std r2,24(r1)" corrupts the caller's stack.
This stub makes sense for a normal call, because the stack would be
owned by the caller of the stub, so it's ok to write r2 to it. But
because this is a sibling call, the stack has been restored and r2 gets
incorrectly saved to the original caller's stack (i.e., xprt_autoclose's
stack).
So xprt_autoclose() -- which is in the sunrpc module -- gets the
livepatch module's toc pointer written to its stack. It panics on when
it tries to use that vlue on its very next call.
Fix it by disallowing sibling calls from patched functions on ppc64le.
In theory we could instead a) generate a custom stub, or b) modify the
kernel livepatch_handler code to save/restore the stack r2 value, but
this is easier for now.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Use make's .DELETE_ON_ERROR special target to make sure we don't have
any .OUTPUT.o leftovers for failed tests.
Fixes: #999
Signed-off-by: Artem Savkov <asavkov@redhat.com>
Apply a sed filter to remove "[<localentry>: 8] " info from
readelf --wide --symbols output. This ensures consistent column
data for the awk script creating the new_symbols file.
Fixes#994
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
The list of prerequsite RPMs which are needed to build the kernel RPM is
constantly growing. But at least some of those RPMs aren't strictly
necessary for building the kernel, at least for kpatch-build's purposes.
Requiring them all to be installed is a bit overkill, and sometimes
causes kpatch-build to fail when it doesn't need to.
If the build does fail, we can always check the kpatch.log file and
update the dependencies listed in the README as needed.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
The ppc64le unit tests are failing because the lookup code is getting
confused by the "[<localentry>: 8]" string in the readelf symbol table
output.
Remove the string so the lookup code can parse it correctly.
kpatch-build already uses the same sed, so they should match.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Increment version to v0.7.0 due to changed behavior of kpatch.service and
kpatch load subcommand. Full list of changes:
- Multiple memory leak fixes in kpatch-build
- livepatch-patch-hook compatability fixes for kernels 5.1+
- Making kpatch-build compatible with custom gcc names
- Added rhel-rebased integration tests
- kpatch.service will no longer unload modules on stop
- kpatch load will no longer fail if a module is already loaded and enabled
- kpatch-build will now check for *_fixup section changes on ppc64le and will
fail on such changes
- Add support for R_X86_64_PLT32
- don't allow jump labels
- ppc64le-specific kpatch-build fixes
Signed-off-by: Artem Savkov <asavkov@redhat.com>
Create an error if a patched function uses a jump label. We need this
until upstream livepatch supports them with a .klp.arch section.
Fixes#946.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Make 3.8 doesn't support the "file" function, which is used by the unit
test Makefile to output the contents of a custom env file, if one
exists. Use shell instead.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
"make unit" automatically rebases the submodule, which causes a lot of
surprises. Instead, just print a warning if it's out of date.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
After Linux commit 47cdd64be483 ("dynamic_debug: refactor
dynamic_pr_debug and friends"), the name of the static local variable
used for dynamic printks is no longer "descriptor".
Make the is_special_static() check broader such that it doesn't care
about the variable name, and just whitelists any variable in the
__verbose section.
Fixes#990.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
This test is no longer needed. I had previously converted it to a unit
test and forgot to remove its integration test counterpart.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Starting with binutils 2.31, the Linux kernel may have R_X86_64_PLT32
relocations. Make sure we support them. This should be as simple as
treating R_X86_64_PLT32 exactly like R_X86_64_PC32 everywhere. For more
details see upstream commit torvalds/linux@b21ebf2.
This also fixes the following issue seen on Fedora 29:
```
$ kpatch-build/kpatch-build -t vmlinux ./test/integration/fedora-27/convert-global-local.patch
Using cache at /home/jpoimboe/.kpatch/src
Testing patch file(s)
Reading special section data
Building original source
Building patched source
Extracting new and modified ELF sections
ERROR: slub.o: 1 function(s) can not be patched
slub.o: function __kmalloc has no fentry/mcount call, unable to patch
/home/jpoimboe/git/kpatch/kpatch-build/create-diff-object: unreconcilable difference
ERROR: 1 error(s) encountered. Check /home/jpoimboe/.kpatch/build.log for more details.
```
Fixes#975.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
kpatch_verify_activeness_safety() calls kpatch_backtrace_address_verify()
for each address in the call traces of the processes.
Among other things, kpatch_backtrace_address_verify() searches the whole
set of functions for the ones being replaced (func->op == KPATCH_OP_UNPATCH).
This is a waste of time when the patch is loaded or unloaded rather than
replaced. Let us do the searching only if patch replacement is in
progress.
Signed-off-by: Evgenii Shatokhin <eshatokhin@virtuozzo.com>
If atomic replacement is used for the old-style patches (the patches
that depend on kpatch.ko), the kernel might crash if the new patch
changes a smaller set of functions than the patch being replaced.
kpatch_apply_patch() does check if the functions from the patch to be
replaced are currently running. However, the functions are removed from
'kpatch_func_hash' in kpatch_register() only after stop_machine() and
kpatch_apply_patch() have finished:
ret = stop_machine(kpatch_apply_patch, kpmod, NULL);
/*
* For the replace case, remove any obsolete funcs from the hash and
* the ftrace filter, and disable the owning patch module so that it
* can be removed.
*/
if (!ret && replace) {
struct kpatch_module *kpmod2, *safe;
hash_for_each_rcu(kpatch_func_hash, i, func, node) {
if (func->op != KPATCH_OP_UNPATCH)
continue;
if (func->force)
force = 1;
hash_del_rcu(&func->node);
WARN_ON(kpatch_ftrace_remove_func(func->old_addr));
}
<...>
As a result, the kernel may end up with an inconsistent set of patched
functions. Some of the functions from the replaced patch could
still be running, while some would be already reverted to the original
ones.
I observed kernel crashes in such situations when I was trying to
replace a patch with a new one without a faulty fix.
Let us remove the replaced patched functions from 'kpatch_func_hash'
in kpatch_apply_patch() to avoid such issues.
Signed-off-by: Evgenii Shatokhin <eshatokhin@virtuozzo.com>
Make kpatch_apply_patch() aware of whether the patch should replace other
patches.
This will be used by subsequent fixes.
Signed-off-by: Evgenii Shatokhin <eshatokhin@virtuozzo.com>
Currently we do not support changes to functions referring to any of the
*_fixup sections on ppc64le. This patch introduces check for such
changes during the patchability check, where we abort building the
patch module.
This patch implements the phase 1 fix of 3 phases discussed at
https://github.com/dynup/kpatch/issues/974:
"
Phase 1 fix:
For kernel versions which don't have livepatch-specific powerpc code
(currently all kernel versions), kpatch-build needs to assert an error
if it detects that one of the following sections refers to a patched
function: __ftr_fixup, __mmu_ftr_fixup, __fw_ftr_fixup.
"
Suggested-by: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
For "kpatch load" invocations, don't set failing return status if the
kpatch module is already loaded and enabled. Make note of the existing
livepatch module and then verify that is has completed its transition
before continuing. This allows the user to more gracefully re-run
"kpatch load" commands to pick up new kpatch modules.
Fixes: #979
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
Add support for __spec_barrier_fixup (barrier nospec fixup) special
section on ppc64le.
Suggested-by: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Fix the size of special group __lwsync_fixup on ppc64le.
Suggested-by: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
This commit contains centos-7 patches rebased and adjusted to work with
recent rhel minors so that integration tests actually pass on those.
Signed-off-by: Artem Savkov <asavkov@redhat.com>
commit f8213c87f6 ("lookup: Fix format string for symtab_read() on
PPC64LE") fixed the symbol table lookup when readelf adds ppc64le
"[<localentry>: 8]" info for functions like so:
23: 0000000000000008 96 FUNC LOCAL DEFAULT [<localentry>: 8] 4 cmdline_proc_show
however, it seems that readelf 2.30-57.el8 displays this in a slightly
different format:
24493: c000000000587970 96 FUNC LOCAL DEFAULT 2 cmdline_proc_show [<localentry>: 8]
Instead of adding more cases to kpatch-build's lookup.c scanf format,
let's just delete this information from the symtab file with a quick and
dirty sed regex. This allows us to handle both observed cases (and
perhaps others) while removing the arch-specific scanf formatting in
lookup.c
Fixes: f8213c87f6 ("lookup: Fix format string for symtab_read() on PPC64LE")
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
The kpatch.service file shouldn't unload patch modules on service stop
(this is also executed by systemd on reboot). Patch modules may not be
designed to be safely unloaded and/or may patch kernel routines that
need to continue to run throughout system bring down.
Suggested-by: disaster123
Signed-off-by: Joe Lawrence <joe.lawrence@redhat.com>
commit 767d9669bd ("kpatch-build: use readelf instead of eu-readelf")
replaced eu-readelf with readelf for constructing symbol table. The
format of symbol table entries differs a little on Power when the symbol
is a function with binding type LOCAL. For example, consider:
23: 0000000000000008 96 FUNC LOCAL DEFAULT [<localentry>: 8] 4 cmdline_proc_show
An extra column preceding index of the symbol denoting symbol value to
be local entry point offset of the function is printed, with the
current sscanf format string in lookup::symtab_read the values will
mismatch ending with in accurate lookup table getting constructed. This
patch fixes it by introducing an Power specific format string for
function symbols with bind type LOCAL.
Fixes: 767d9669 ("kpatch-build: use readelf instead of eu-readelf")
Signed-off-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
