There is a fundemental flaw in how unaligned accesses are supported, but this
is still an improvement.
--HG--
extra : convert_revision : 1c20b524ac24cd4a812c876b067495ee6a7ae29f
The only cases where this was the correct behavior are now handled with the
"B" operand type, and doing things this way was breaking some instructions,
notably a shift.
--HG--
extra : convert_revision : 072346d4f541edaceba7aecc26ba8d2cd756e481
Make instructions observe segment prefixes, default segment rules, segment
base addresses.
Also fix some microcode and add sib and riprel "keywords" to the x86
specialization of the microassembler.
--HG--
extra : convert_revision : be5a3b33d33f243ed6e1ad63faea8495e46d0ac9
This lets you index into a group of registers without having to know
explicitly which one is the lowest in that group.
--HG--
extra : convert_revision : e3cad25a1c5910955204c37177b049ca9834cfd9
The arch_prctl system call is used to set and get the FS and GS segment
bases. The FS segment is use for TLS, so glibc needs to be able to set it
up.
--HG--
extra : convert_revision : 79501491a15967a7a862add846ff88a934fb1b37
After very carefully reading through the Linux source, I'm pretty confident I now know -exactly- how the initial stack frame is constructed, filled, and aligned.
--HG--
extra : convert_revision : 3c654ade7e458bdd5445026860f11175f383a65f
R11 is just junk after the start of exectuion because we're "returning" from
an execve call and linux destroys the contents of rcx and r11 on system calls.
--HG--
extra : convert_revision : 6bf69a50ce56e0355dfdd41524163874340beec0
These are the only floating point instructions that get used in my simple hello world test. These instructions are for setting up the floating point control register. Their not being implemented doesn't affect anything because floating point isn't used.
--HG--
extra : convert_revision : 4dfb9ef2a5665f034946c504978029e8799e64cd
The instructions now ask for the appropriate flags to be set, and the microops do the "right thing" with the CF and OF flags, namely zero them.
--HG--
extra : convert_revision : 85138a832f44c879bf8a11bd3a35b58be6272ef3
The initial stack frame for x86 is now substantially more correct. The fixes made here can be back ported to SPARC and possible the other ISAs as well. The auxiliary vector types were moved to the LiveProcess base class because they are independent of ISA. Some of the types may only apply to Linux, though, so they may have to be moved.
--HG--
extra : convert_revision : 89ace35fcc8eb9586d2fee8eeccbc3686499ef24
POPA used st instead of ld, and it didn't skip rsp. push rsp needs to store the -original- value of the stack pointer.
--HG--
extra : convert_revision : 376370c99b6ab60fb2bc4cd4f0a6dce71153ad06
Merge was returning the value to merge in, not the actual result of the merge.
--HG--
extra : convert_revision : 230b4b5064037d099ae7859edabdf5be84603849
The stack base on my development machine starts one page below where it needs to. I don't know why it does, but I've duplicated it in m5.
--HG--
extra : convert_revision : c4783ba885b90f17e843f61e07af0bc3330a74bc
The type constants should go into an architecture independent spot since they are universal to all Linux elf binaries. The right value for some of the vectors needs to be determined. Also, x86 does not store argc or argv_array_base in registers like some other architectures.
--HG--
extra : convert_revision : 8d3f6a3e028d881d3c41e8ddf4f29d25738b529c