This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
RE: Broken SH2a patches
- From: Andrew STUBBS <andrew dot stubbs at st dot com>
- To: "'Nick Clifton'" <nickc at redhat dot com>
- Cc: "'Alexandre Oliva'" <aoliva at redhat dot com>, <binutils at sources dot redhat dot com>, "'Joern RENNECKE'" <joern dot rennecke at st dot com>
- Date: Tue, 11 Jan 2005 18:02:14 -0000
- Subject: RE: Broken SH2a patches
- Organization: STMicroelectronics
Hi Nick,
I have found another problem with your patch.
Your patch has:
#define arch_sh2a_or_sh3e (arch_sh2a_sh4_base)
#define arch_sh2a_or_sh4 (arch_sh2a_sh4_base
|arch_sh_dp_fpu)
Which should be:
#define arch_sh2a_or_sh3e (arch_sh2a_sh4_base|arch_sh_no_mmu
|arch_sh_sp_fpu)
#define arch_sh2a_or_sh4 (arch_sh2a_sh4_base|arch_sh_no_mmu
|arch_sh_dp_fpu)
Otherwise the -isa=sh2a-or-sh3e and -isa=sh2a-or-sh4 options cause the
assembler to reject all instructions, which isn't helpful.
I am developing a new script for the testsuite which will hopefully point
all these issues out in future, should anybody make any broken changes. It
hasn't found any issues with your patch other than the ones I have already
pointed out (although I have had to modifiy sh-opc.h a little to make it
work, so maybe I am not testing the same thing).
Interestingly, one thing my script has exposed is that mul.l is defined
twice! Once for sh2 and once for sh1, which is obviously wrong.
--
Andrew Stubbs
andrew.stubbs@st.com
andrew.stubbs@superh.com