Clear random header value by AIO read error

[masaori335]

Fix #7865. We observed the crash loop on start-up with a bad disk.

Backtrace

[ 00 ] libpthread-2.17.so waitpid 
[ 01 ] traffic_server crash_logger_invoke(int, siginfo_t*, void*) ( Crash.cc:165 ) 
[ 02 ] libpthread-2.17.so 
[ 03 ] traffic_server CacheDisk::~CacheDisk() ( List.h:293 ) 
[ 04 ] traffic_server CacheDisk::~CacheDisk() ( CacheDisk.cc:112 ) 
[ 05 ] traffic_server CacheProcessor::diskInitialized() ( Cache.cc:767 ) 
[ 06 ] traffic_server CacheDisk::openStart(int, void*) ( CacheDisk.cc:213 ) 
[ 07 ] traffic_server AIOCallbackInternal::io_complete(int, void*) ( eventsystem/I_Continuation.h:160 ) 
[ 08 ] traffic_server EThread::process_event(Event*, int) ( I_Continuation.h:160 ) 
[ 09 ] traffic_server EThread::process_queue(Queue<Event, Event::Link_link>*, int*, int*) ( UnixEThread.cc:170 ) 
[ 10 ] traffic_server EThread::execute_regular() ( UnixEThread.cc:230 ) 
[ 11 ] traffic_server spawn_thread_internal(void*) ( Thread.cc:85 ) 
[ 12 ] libpthread-2.17.so start_thread 

Log

[ET_AIO 0:99] WARNING: cache disk operation failed READ -1 61
[ET_NET 9] WARNING: could not read disk header for disk /dev/disk/****: declaring disk bad
[ET_NET 9] WARNING: failed operation: READ (opcode=1), span: /dev/disk/**** (fd=99)

Conditions

1. Some HDD on a box got broken
2. (re)start ATS for some reason

Scenario

1. On ET_NET thread, CacheDisk::open schedule AIO read to load disk header info into DiskHeader *header member.
2. On ET_AIO thread, cache_op try to read the disk but got an error (errno=61). When this error happens, the given buffer (header) has a random value.
3. On ET_NET thread, CacheDisk::openStart handles the callback and declares it's a bad disk because of the result. It deletes the CacheDisk object.
4. The CacheDisk destructor refers to the header which has a random value on a loop and got heap-buffer-overflow.

Affected Version

8.1.x, 9.0.x, and master

[zwoop]

Cherry-picked to v9.1.x branch.
Cherry-picked to v9.2.x

[ezelkow1]

If we want this for 8.x can we get a backport?

[masaori335]

@ezelkow1 Here you're > #8653