- Timestamp:
- Aug 2, 2013 12:26:46 AM (11 years ago)
- Location:
- trunk/src
- Files:
-
- 2 edited
Legend:
- Unmodified
- Added
- Removed
-
trunk/src/VBox/HostDrivers/Support/SUPDrvTracer.cpp
r44529 r47518 346 346 /* The loader and/or ld64-97.17 seems not to generate fixups for our 347 347 __VTGObj section. Detect this by comparing them with the 348 u64VtgObjSectionStart member and assume max image size of 4MB. */ 348 u64VtgObjSectionStart member and assume max image size of 4MB. 349 Seems to be worked around by the __VTGPrLc.End and __VTGPrLc.Begin 350 padding fudge, meaning that the linker misplaced the relocations. */ 349 351 if ( (int64_t)u64Tmp != (int32_t)u64Tmp 350 352 && pVtgHdr->u64VtgObjSectionStart != uVtgHdrAddr -
trunk/src/bldprogs/VBoxTpG.cpp
r47517 r47518 484 484 "VTG_GLOBAL g_aVTGObj_LinkerPleaseNoticeMe, data\n" 485 485 " [section __VTG __VTGPrLc.Begin align=16]\n" 486 " dq 0, 0 ; Paranoia, related to the fudge below.\n" 486 487 "VTG_GLOBAL g_aVTGPrLc, data\n" 487 488 " [section __VTG __VTGPrLc align=16]\n" … … 489 490 " [section __VTG __VTGPrLc.End align=16]\n" 490 491 "VTG_GLOBAL g_aVTGPrLc_End, data\n" 491 " dq 0, 0 ; Fudge to work around unidentified linker /loader issue, causing a partial address\n"492 " ; to be written to the start of the first entry, generally making the line number\n"493 " ; invalid. (1 byte is actually enough)\n"492 " dq 0, 0 ; Fudge to work around unidentified linker where it would otherwise generate\n" 493 " ; a fix up of the first dword in __VTGPrLc.Begin despite the fact that it were\n" 494 " ; an empty section with nothing whatsoever to fix up.\n" 494 495 " %%endif\n" 495 496 " [section __VTG __VTGObj]\n"
Note:
See TracChangeset
for help on using the changeset viewer.