Modernize x86log.h and x86log.cpp:

[vittorioromeo]

• Use variadic template functions instead of va_list
  • In x86log_mark_as_data and x86log_printf
• Add noexcept where appropriate/safe
• Use constexpr std::size_t instead of macros for constants
• Nest nullptr check to prevent useless check, add assertion
  • In x86log_printf
• Move definitions of log_comment, data_range_t, x86log_context to x86_log.h from x86_log.cpp
• Adapt usages of x86log_printf in drcbex64.cpp and drcbex86.cpp

[mmicko]

passes gcc 5.2 and vs2015, will check clang now

[Tafoid]

I suspect this submission as being the cause for a crash/exception upon exit for most system which use DRC or DRC backend such as the following driver examples: (aleck64, coolridr, stv, deco_mlc, model3, taitotz, cps3, psikyosh, suprnova and many more). The MIPS core doesn't seem to be influenced by this, examples of such are (kinst, vegas)

#0 0x00000000008759ec in resource_pool::remove ()
at ../../../../../src/emu/emualloc.cpp:127
#1 0x000000000082eb99 in drcuml_state::~drcuml_state ()
at ../../../../../src/devices/cpu/drcuml.cpp:145
#2 0x0000000000cb223c in std::default_delete<drcuml_state>::operator() ()
at C:/buildtools/vendor/mingw64/x86_64-w64-mingw32/include/c++/bits/unique_ptr.h:76
#3 0x0000000000d0c4de in std::unique_ptr<drcuml_state, std::default_delete<drcuml_state> >::~unique_ptr ()
at C:/buildtools/vendor/mingw64/x86_64-w64-mingw32/include/c++/bits/unique_ptr.h:236
#4 0x0000000000ba8c9b in ppc_device::~ppc_device ()
at ../../../../../src/devices/cpu/powerpc/ppc.h:174
#5 0x0000000000bcced3 in mpc8240_device::~mpc8240_device ()
at ../../../../../src/devices/cpu/powerpc/ppc.h:732
#6 0x0000000000bcce4a in mpc8240_device::~mpc8240_device ()
at ../../../../../src/devices/cpu/powerpc/ppc.h:732
#7 0x0000000000bb0379 in simple_list<device_t>::remove ()
at ../../../../../src/lib/util/coretmpl.h:242
#8 0x0000000000bb01d0 in simple_list<device_t>::reset ()
at ../../../../../src/lib/util/coretmpl.h:89
#9 0x0000000000bb03da in simple_list<device_t>::~simple_list ()
at ../../../../../src/lib/util/coretmpl.h:78
#10 0x0000000000867ae3 in device_t::~device_t ()
at ../../../../../src/emu/device.cpp:72

[vittorioromeo]

I'm still looking into any possible mistake - is it possible that simple_list was being used polymorphically somewhere?

It had a virtual destructor that I changed into a non-virtual destructor in aa6d4d0, because I could not find any polymorphically usage of simple_list in the codebase.

Assuming I missed such usages: could you try if reintroducing virtual in ~simple_list

virtual ~simple_list() noexcept { reset(); }

and check if the crash still occurs?

[mmicko]

Thanks, but problem is not there. it is in DRC part, will take care of
that.
It is related to mixing unique_ptr with our auto_alloc system

On Wed, Dec 23, 2015 at 7:03 PM, Vittorio Romeo notifications@github.com
wrote:

I'm still looking into any possible mistake - is it possible that
simple_list was being used polymorphically somewhere?

It had a virtual destructor that I changed into a non-virtual destructor
in aa6d4d0
aa6d4d0,
because I could not find any polymorphically usage of simple_list in the
codebase.

Assuming I missed such usages: could you try if reintroducing virtual in
~simple_list

virtual ~simple_list() noexcept { reset(); }

and check if the crash still occurs?

—
Reply to this email directly or view it on GitHub
#517 (comment).