tcg_out_ld(s, TCG_TYPE_I64, r1, r1, offsetof(CPUTLBEntry, addend) -
offsetof(CPUTLBEntry, addr_read));
should be modified to
tcg_out_ld(s, TCG_TYPE_I64, r1, r1, offsetof(CPUTLBEntry, addend));
Actually, it should be the entire offset from CPUState. The ADDR_WRITE
offset before was entirely st
BUG?
In tcg_out_qemu_ld
tcg_out_ld(s, TCG_TYPE_I64, r1, r1,
offsetof(CPUTLBEntry, addend)
- offsetof(CPUTLBEntry, addr_read));
should be modified to
tcg_out_ld(s, TCG_TYPE_I64, r1, r1,
offsetof(CPUTLBEntry, addend));
since the when calling tcg_out
On Sat, Jan 30, 2010 at 12:04 AM, Stefan Weil wrote:
[...]
>
> that was a good suggestion. bntest raises a segmentation fault
> (NULL pointer, p == 0, see below) with qemu-x86_64 on a x86_64 host.
Compile bntest statically and it should work.
x86_64 user mode is completely broken for dynamically
On Sat, Jan 30, 2010 at 12:04:29AM +0100, Stefan Weil wrote:
> Edgar E. Iglesias schrieb:
> > On Fri, Jan 29, 2010 at 09:04:58AM -0800, Richard Henderson wrote:
> >> On 01/28/2010 05:55 PM, identifier scorpio wrote:
> >>> Your code is more skilled and I learn a lot from it, but I haven't yet
> >>>
Edgar E. Iglesias schrieb:
> On Fri, Jan 29, 2010 at 09:04:58AM -0800, Richard Henderson wrote:
>> On 01/28/2010 05:55 PM, identifier scorpio wrote:
>>> Your code is more skilled and I learn a lot from it, but I haven't yet
>>> recognized the key difference between your version and mine. In last
>>
On Fri, Jan 29, 2010 at 09:04:58AM -0800, Richard Henderson wrote:
> On 01/28/2010 05:55 PM, identifier scorpio wrote:
> > Your code is more skilled and I learn a lot from it, but I haven't yet
> > recognized the key difference between your version and mine. In last
> > mail, you've said that "Main
+} else if (cond == TCG_COND_EQ || cond == TCG_COND_NE) {
+tcg_out_mem_long(s, INSN_LDA, TMP_REG1, arg1, -arg2);
+opc = (cond == TCG_COND_EQ ? INSN_BEQ : INSN_BNE);
Bug here. What was intended was to add "arg1 = TMP_REG1".
But since the constraints use "I" for ui
On 01/25/2010 05:19 PM, Richard Henderson wrote:
+} else if (~val == (uint8_t)~val) {
+tcg_out_fmt_opi(s, INSN_BIC, ra, ~val, rc);
Bug here. I've applied the following to my local tree.
r~
commit 2ecce92da6eee4b3496c7655da45259308abb536
Author: Richard Henderson
Date: Fri Jan
On 01/28/2010 05:55 PM, identifier scorpio wrote:
1. adding constraints for operation mov_i32/movi_i32/mov_i64/movi_i64
2. replacing "goto do_arith" with "break" for INDEX_op_not_i32/i64 case
handling in tcg_out_op
Oops.
3. removing the redundant declaration for "alpha_tb_set_jmp_target()" at
I'v reviewed your code and applied the patch to 0.12.0.rc1. After fixing some
small bugs I run it on my debian/alpha platform, and it hung after running
abort 3 TBs. I'll continue to trace your code.
The bugs that are fixed are listed as follow:
1. adding constraints for operation mov_i32/movi_i
On 01/22/2010 07:47 AM, identifier scorpio wrote:
Is there any good method to find the bug except "guess and try"?
-singlestep -d in_asm,cpu,exec
Use these options on both Alpha host and x86_64 host and diff the
resulting output files. That will show you which target instruction is
emulated d
Hi, Richard.
I almost re-write the whole code under your help and now it seems somehow
elegant. Unfortunately, it still can't run MS windows, I wonder whether there
are some other special places in QEMU to be modifed for alpha case, beside the
3 files (cpu-all.h/tcg-target.c/tcg-target.h) I wri
identifier scorpio schrieb:
> Thank Mr. Weil for your reply.
>
> >
> > Maybe you can also try the TCG interpreter (TCI) from
> > http://repo.or.cz/w/qemu/ar7.git.
> > In theory, it supports any host architecture with or
> > without native TCG
> > support.
> >
> > It was tested successful with som
Thank Mr. Weil for your reply.
>
> Maybe you can also try the TCG interpreter (TCI) from
> http://repo.or.cz/w/qemu/ar7.git.
> In theory, it supports any host architecture with or
> without native TCG
> support.
>
> It was tested successful with some basic tests on x86,
> mips, ppc and arm,
> s
On 01/20/2010 09:19 AM, identifier scorpio wrote:
Below i'll append my newly generated patch against stable-0.10,
in case it is mangled, i also put it in the attachment.
For the record, the inline portion was again mangled; the attachment is
fine.
in qemu_ld/st, we must use $16,$17,$18 as t
identifier scorpio schrieb:
> Hello.
>
> I ported TCG to alpha platform, the patch is currently based on stable-0.10
> branch, and now it can run linux-0.2.img testing image on my alpha XP1000
> workstation. but it still can't run MS-windows, and I hope someone,
> especially those guys that are
On Tue, 19 Jan 2010, Richard Henderson wrote:
> On 01/19/2010 12:47 AM, identifier scorpio wrote:
> > I ported TCG to alpha platform, the patch is currently based on stable-0.10
> > branch,
> > and now it can run linux-0.2.img testing image on my alpha XP1000
> > workstation.
> > but it still can'
On 01/19/2010 12:47 AM, identifier scorpio wrote:
I ported TCG to alpha platform, the patch is currently based on stable-0.10
branch,
and now it can run linux-0.2.img testing image on my alpha XP1000 workstation.
but it still can't run MS-windows, and I hope someone, especially those guys
that
Hello.
I ported TCG to alpha platform, the patch is currently based on stable-0.10
branch, and now it can run linux-0.2.img testing image on my alpha XP1000
workstation. but it still can't run MS-windows, and I hope someone, especially
those guys that are working on target-alpha, may help me to
19 matches
Mail list logo
