2763915 - Random memory corruption with different stack due to overflowing a page during reformatting - SAP ASE

• An "infected with signal 11" error in the module OptGlobal::OptGlobal() together with a stack trace appears in the ASE errorlog:

Current process (0x1910021) infected with signal 11 (SIGSEGV)
Address 0x0x000000000089b4b2 (OptGlobal::OptGlobal(Node<Lop>*, tree*, OptPool*)+0x4260), siginfo (code, address) = (1, 0x0x0000000000000018) 

• Alternatively, an "infected with signal 11" error in the module LeCacheWkTblIO::updateLOBCacheRow() together with a stack trace appears in the ASE errorlog, ending with a zombie task:

Current process (0x26550098) infected with signal 11 (SIGSEGV)
Address 0x0000000080932c1c (void LeCacheWkTblIO::updateLOBCacheRow(le_row*,int)+0x54), siginfo (code, address) = (1, 0x000000000000000a)

• Various different patterns of stack traces may be seen in the ASE errorlog if a query is using reformat strategy on a table with row size that is potentially bigger than the ASE configured page size.
• Memory corruption may get into different stack traces with different segmentation fault routine names.
• Examples of complete stack traces are available under Attachments.

• SAP Adaptive Server Enterprise (ASE) 16.0
• SAP Adaptive Server Enterprise (ASE) 15.7

kisignal, stacktrace, registers, CR817400,  CR#817400, 817400, SEGV, segmentation, storage, access, violation, fault, segfault , KBA , BC-SYB-ASE , Sybase ASE Database Platform (non Business Suite) , Bug Filed