03579: rbibb, rbibba, vsskykid: Severe graphic corruption/Access Voilation

ID	Category [?]	Severity [?]	Reproducibility	Date Submitted	Last Update
03579	Crash/Freeze	Major	Have not tried	Dec 14, 2009, 03:45	Jun 6, 2010, 14:09

Tester	Firewave	View Status	Public	Platform	MAME (Self-compiled)
Assigned To	etabeta	Resolution	Fixed	OS	Windows XP (64-bit)
Status [?]	Resolved			Driver	nintendo/vsnes.cpp
Version	0.135u3	Fixed in Version	0.138u2	Build	Debug
		Fixed in Git Commit		Github Pull Request #

Summary	03579: rbibb, rbibba, vsskykid: Severe graphic corruption/Access Voilation
Description	The message from rbibb ----------------------------------------------------- Exception at EIP=00B78FFA (memory_dump+0x5d11): ACCESS VIOLATION While attempting to read memory at 1475E23C ----------------------------------------------------- EAX=1475E23C EBX=00000082 ECX=00000240 EDX=1475DFFC ESI=11E94888 EDI=021A1AE4 EBP=0022F9E8 ESP=0022F9D0
Steps To Reproduce
Additional Information	This is manifest in normal builds buy a severe graphic corruption - but no apparent crash.
Github Commit

Flags
Regression Version	0.133u1
Affected Sets / Systems	rbibb, rbibba, vsskykid

Attached Files

No.05274 Tafoid Administrator Dec 14, 2009, 18:48	Program received signal SIGSEGV, Segmentation fault. 0x00b78fba in read_byte_generic (space=0x188e1348, byteaddress=576) at src/emu/memory.c:558 558 result = (*handler->bankbaseptr)[byteoffset];
No.05275 Haze Senior Tester Dec 14, 2009, 21:20	It's probably related to the suprmrio bug I posted. I have a feeling some banking has been broken with the introduction of the new memory system / install handler / bank changes.
No.05276 Tafoid Administrator Dec 14, 2009, 21:26	The regression goes back to when there was work being done on NES on the MESS side, I believe. Fabio would probably be most knowledgeable about what might be happening.
No.05277 Haze Senior Tester Dec 14, 2009, 23:13	ahh ok, based on where it crashes it looks like it's trying to read banked memory past the end of the bank, or from an invalid pointer, and given that it doesn't seem to copy any valid data for the tiles, that seems rather likely
No.05281 Firewave Senior Tester Dec 15, 2009, 17:41	Backtrace for rbibba including stack crawl: ----------------------------------------------------- Exception at EIP=00B78FFA (memory_dump+0x5d11): ACCESS VIOLATION While attempting to read memory at 14FCE23C ----------------------------------------------------- EAX=14FCE23C EBX=02555122 ECX=00000240 EDX=14FCDFFC ESI=CC2AF332 EDI=02555122 EBP=0022F9E8 ESP=0022F9D0 ----------------------------------------------------- Stack crawl: exception-> 00B78FFA (memory_dump+0x5d11) 0022F9EC: 00B7AA38 (memory_read_byte_8le+0x0018) 0022F9FC: 007BAD11 (ppu2c0x_init_palette+0x0c21) 0022FC9C: 00BC14BB (timer_execute_timers+0x0298) 0022FD0C: 00B6DB83 (cpuexec_timeslice+0x08de) 0022FE1C: 00B561EB (mame_execute+0x029b) 0022FE7C: 00DBF793 (cli_execute+0x01e3) 0022FEEC: 00B077CB (utf8_main+0x00ea) 0022FF1C: 0156213C (main+0x00ec) 0022FF6C: 0040124B (__image_base__+0x124b) 0022FFB4: 00401298 (mainCRTStartup+0x0018)
No.05490 Tafoid Administrator Jan 11, 2010, 18:46	Here is a stack with the new tools, r7969 ----------------------------------------------------- Exception at EIP=00D38DA4 (operator+0x061d): ACCESS VIOLATION While attempting to read memory at 1895E240 ----------------------------------------------------- EAX=1895E240 EBX=00000000 ECX=00000000 EDX=00000240 ESI=00000000 EDI=00000EBC EBP=0022F770 ESP=0022F748 ----------------------------------------------------- Stack crawl: exception-> 00D38DA4 (operator+0x061d) 0022F774: 00D4330C (memory_read_byte_8le(_address_space+0x0018) 0022F794: 0081371B (video_update_playch10(device_config+0x02df) 0022F7B4: 0081464D (ppu2c0x_init_palette(running_machine,+0x0ade) 0022F994: 018825C8 (_Unwind_SjLj_Register+0x0088) 0022F99C: 01895FD0 (0x5b0+0x0000) 0022F9B4: 01882516 (_Unwind_SjLj_Unregister+0x0036) 0022FA54: 00DD2AE9 (timer_execute_timers(running_machine)+0x0292) 0022FA84: 010BE21E 0022FB3C: 01895FD0 (0x5b0+0x0000) 0022FB68: 00DB0F01 (cpuexec_timeslice(running_machine)+0x0a8c) 0022FC04: 00DABD46 (mame_execute(_core_options)+0x04b5) 0022FC9C: 01895FD0 (0x5b0+0x0000) 0022FCF4: 01837962 (core_stricmp(char+0x003a) 0022FDBC: 01895FD0 (0x5b0+0x0000)
No.05594 Tafoid Administrator Jan 29, 2010, 20:27	In 0.136u2, these no longer seem to assert/crash, but the graphic issues still remain in all builds.
No.05597 etabeta Developer Jan 29, 2010, 22:55 edited on: Jan 29, 2010, 22:55	they asserted in debug build, last I tried (when Aaron reported it to the list)...
No.05598 Tafoid Administrator Jan 29, 2010, 23:46	I built debug twice to make sure, they aren't asserting for me as of u2. When a few others didn't assert I had thought something was wrong with my build .. notice all the other 'fixed' I resolved in the last day. When you get a moment, please confirm for yourself.
No.05940 AWJ Developer Apr 5, 2010, 04:02	The severity of this bug should be upgraded from "Minor"--currently, the graphics in all three affected sets are complete garbage, rendering the games unplayable.