valgrind: Syscall param ppoll(ufds.events) points to uninitialised byte(s)

[fanquake]

Using master, GCC 12.2.1, Valgrind 3.21.0 (system), aarch64, Alpine Linux (musl libc):

./test/functional/test_runner.py --combinedlogslen=19999 --timeout-factor=12 --jobs=9 --valgrind
...
24/268 - p2p_dns_seeds.py failed, Duration: 9 s
....
 node0 2023-07-12T12:59:04.448921Z [init] [noui.cpp:56] [noui_InitMessage] init message: Done loading 
 node0 2023-07-12T12:59:04.453823Z [addcon] [util/thread.cpp:20] [TraceThread] addcon thread start 
 node0 2023-07-12T12:59:04.456041Z [msghand] [util/thread.cpp:20] [TraceThread] msghand thread start 
 test  2023-07-12T12:59:04.763000Z TestFramework.node0 (DEBUG): Stopping node 
 test  2023-07-12T12:59:04.763000Z TestFramework (ERROR): Assertion failed 
                                   Traceback (most recent call last):
                                     File "/bitcoin/test/functional/test_framework/test_framework.py", line 552, in start_nodes
                                       node.wait_for_rpc_connection()
                                     File "/bitcoin/test/functional/test_framework/test_node.py", line 235, in wait_for_rpc_connection
                                       raise FailedToStartError(self._node_msg(
                                   test_framework.test_node.FailedToStartError: [node 0] bitcoind exited with status 1 during initialization
                                   During handling of the above exception, another exception occurred:
                                   Traceback (most recent call last):
                                     File "/bitcoin/test/functional/test_framework/test_framework.py", line 130, in main
                                       self.setup()
                                     File "/bitcoin/test/functional/test_framework/test_framework.py", line 295, in setup
                                       self.setup_network()
                                     File "/bitcoin/test/functional/test_framework/test_framework.py", line 389, in setup_network
                                       self.setup_nodes()
                                     File "/bitcoin/test/functional/test_framework/test_framework.py", line 411, in setup_nodes
                                       self.start_nodes()
                                     File "/bitcoin/test/functional/test_framework/test_framework.py", line 555, in start_nodes
                                       self.stop_nodes()
                                     File "/bitcoin/test/functional/test_framework/test_framework.py", line 570, in stop_nodes
                                       node.stop_node(wait=wait, wait_until_stopped=False)
                                     File "/bitcoin/test/functional/test_framework/test_node.py", line 346, in stop_node
                                       self.stop(wait=wait)
                                       ^^^^^^^^^
                                     File "/bitcoin/test/functional/test_framework/test_node.py", line 192, in __getattr__
                                       assert self.rpc_connected and self.rpc is not None, self._node_msg("Error: no RPC connection")
                                                                     ^^^^^^^^^^^^^^^^^^^^
                                   AssertionError: [node 0] Error: no RPC connection
 test  2023-07-12T12:59:04.764000Z TestFramework (DEBUG): Closing down network thread 
 test  2023-07-12T12:59:04.814000Z TestFramework (INFO): Stopping nodes 
 test  2023-07-12T12:59:04.814000Z TestFramework.node0 (DEBUG): Stopping node 

 node0 stderr ==30770== Thread 16 b-dnsseed:
==30770== Syscall param ppoll(ufds.events) points to uninitialised byte(s)
==30770==    at 0x405DCCC: ??? (in /lib/ld-musl-aarch64.so.1)
==30770==  Address 0x7d219a4 is on thread 16's stack
==30770== 
{
   <insert_a_suppression_name_here>
   Memcheck:Param
   ppoll(ufds.events)
   obj:/lib/ld-musl-aarch64.so.1
}
==30770== 
==30770== Exit program on first error (--exit-on-first-error=yes) 

254/268 - feature_config_args.py failed, Duration: 20 s
....
 node0 2023-07-12T13:16:16.303713Z [opencon] [util/thread.cpp:20] [TraceThread] opencon thread start 
 node0 2023-07-12T13:16:16.309602Z [opencon] [net.cpp:1633] [ThreadOpenConnections] Fixed seeds are disabled 
 test  2023-07-12T13:16:16.366000Z TestFramework (ERROR): Unexpected exception caught during testing 
                                   Traceback (most recent call last):
                                     File "/bitcoin/test/functional/test_framework/test_framework.py", line 131, in main
                                       self.run_test()
                                     File "/bitcoin/test/functional/feature_config_args.py", line 376, in run_test
                                       self.test_args_log()
                                     File "/bitcoin/test/functional/feature_config_args.py", line 182, in test_args_log
                                       self.start_node(0, extra_args=[
                                     File "/bitcoin/test/functional/test_framework/test_framework.py", line 537, in start_node
                                       node.wait_for_rpc_connection()
                                     File "/bitcoin/test/functional/test_framework/test_node.py", line 235, in wait_for_rpc_connection
                                       raise FailedToStartError(self._node_msg(
                                   test_framework.test_node.FailedToStartError: [node 0] bitcoind exited with status 1 during initialization
 test  2023-07-12T13:16:16.367000Z TestFramework (DEBUG): Closing down network thread 
 test  2023-07-12T13:16:16.417000Z TestFramework (INFO): Stopping nodes 
 test  2023-07-12T13:16:16.417000Z TestFramework.node0 (DEBUG): Stopping node 

 node0 stderr ==41539== Thread 15 b-addcon:
==41539== Syscall param ppoll(ufds.events) points to uninitialised byte(s)
==41539==    at 0x405DCCC: ??? (in /lib/ld-musl-aarch64.so.1)
==41539==  Address 0x8121724 is on thread 15's stack
==41539== 
{
   <insert_a_suppression_name_here>
   Memcheck:Param
   ppoll(ufds.events)
   obj:/lib/ld-musl-aarch64.so.1
}
==41539== 
==41539== Exit program on first error (--exit-on-first-error=yes) 

Full logs: https://gist.github.com/fanquake/8e8ce3efcdbf77a85103a36c29c3e68c.
Looking at other discussion, seems like this is likely a Valgrind false-positive, so we may just want to suppress.

[maflcko]

To reproduce without tests:

# valgrind --gen-suppressions=all --quiet --error-exitcode=1 --suppressions=$PWD/contrib/valgrind.supp --exit-on-first-error=yes ./src/bitcoind -regtest -noconnect -addnode=some.node  -printtoconsole=0
==47801== Thread 25 b-addcon:
==47801== Syscall param ppoll(ufds.events) points to uninitialised byte(s)
==47801==    at 0x405DCCC: ??? (in /lib/ld-musl-aarch64.so.1)
==47801==  Address 0x823b724 is on thread 25's stack
==47801== 
{
   <insert_a_suppression_name_here>
   Memcheck:Param
   ppoll(ufds.events)
   obj:/lib/ld-musl-aarch64.so.1
}
==47801== 
==47801== Exit program on first error (--exit-on-first-error=yes)

[maflcko]

Maybe someone can transform this into a standalone minimal cpp file to submit to valgrind or musl?

[maflcko]

Done:

/bitcoin-core # cat /tmp/a.cpp 
#include <sys/types.h>
#include <sys/socket.h>
#include <netdb.h>

int main(){
    addrinfo ai_hint{};
    ai_hint.ai_socktype = SOCK_STREAM;
    ai_hint.ai_protocol = IPPROTO_TCP;
    ai_hint.ai_family = AF_UNSPEC;
    ai_hint.ai_flags = true? AI_ADDRCONFIG : AI_NUMERICHOST;
    addrinfo* ai_res{nullptr};
    return getaddrinfo("foo.bar", nullptr, &ai_hint, &ai_res);
}
/bitcoin-core # clang++  /tmp/a.cpp -o /tmp/exe && valgrind --gen-suppressions=all --quiet /tmp/exe 
==77232== Syscall param ppoll(ufds.events) points to uninitialised byte(s)
==77232==    at 0x405DCCC: ??? (in /lib/ld-musl-aarch64.so.1)
==77232==  Address 0x1ffefff804 is on thread 1's stack
==77232== 
{
   <insert_a_suppression_name_here>
   Memcheck:Param
   ppoll(ufds.events)
   obj:/lib/ld-musl-aarch64.so.1
}

[fanquake]

Opened a bug report with Valgrind: https://bugs.kde.org/show_bug.cgi?id=472219

[maflcko]

Ok, let's continue discussion there.