Hi all, I have created a new cache indexing policy, in which a few ways are used by multiple sets. I have a new parameter in each cache block (called Replaceably Entry class in gem5) that determines to which set it currently belongs to, which is used to regenerate the address of the data it currently holds.
This is the error I get when I run it: (SE Mode) 0: system.remote_gdb: listening for remote gdb on port 7000 info: Entering event queue @ 0. Starting simulation... warn: ignoring syscall access(...) warn: ignoring syscall access(...) warn: ignoring syscall access(...) warn: ignoring syscall mprotect(...) Inconsistency detected by ld.so: ../sysdeps/x86_64/dl-machine.h: 540: elf_machine_rela_relative: Assertion `ELFW(R_TYPE) (reloc->r_info) == R_X86_64_RELATIVE' failed! Exiting @ tick 16697500 because exiting with last active thread context Simulated exit code not 0! Exit code is 127 Does anyone have any idea why I am seeing this? Please feel free to reply to me for more information on the implementation. Some more info: 1.) I have added debugging statements and can confirm that the regenerated address is correct for every cache block. 2.) I have also observed that there is no error as long as one replaceable entry is owned by only one set for e.g. if each set can't use its own last two ways but can use last two ways of next set i.e. the set 0 uses last two ways of set 1, set 1 uses last two ways of set 2 and so on, there is no error as there is always one owner only. If I modify the earlier configuration where a set uses all its ways plus the last two ways of the next set, I start to see this error. This also confirms that the address regeneration is working fine. Thanks, Yogesh
_______________________________________________ gem5-users mailing list -- [email protected] To unsubscribe send an email to [email protected] %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
