7 comments:
Patch Set #1, Line 9: filed
filed, or filled?
compilation would fail
with x86_64 + debugging enabled.
If some combination of settings results in a build failure when it shouldn't, it means it's not being build-tested. Please add a config file under configs/ to make sure it doesn't break again.
CB:43977 added an example config to build-test several options. Watch out for the filename, CB:43611 had to rename a file because Jenkins ignored it.
Code that interacts with GDB was accessing the eip and eflags filed in
the eregs struct.
Because those registers are rip and rflags in x86_64, compilation would fail
with x86_64 + debugging enabled.
Please wrap lines at 72 characters.
Patch Set #1, Line 14: compilation
Does it work as intended, though?
Patch Set #1, Line 14: accessing the fields with a macro
IMHO, the macros are a bit of an eyesore. How about using inline functions that take in a `struct eregs` and return the register value?
Patch Set #1, Line 15: expends
expands
File src/arch/x86/exception.c:
Patch Set #1, Line 382: uint32_t
There's lots of assumptions about register sizes. Since `gdb_stub_registers` is an array of `uint32_t`, the following writes will truncate for x86_64.
To view, visit change 45032. To unsubscribe, or for help writing mail filters, visit settings.