This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] microblaze: Use 'SI' instead of 'VOID' for operand 1 of 'call_value_intern'

On 10/21/14 22:55, Chen Gang wrote:
> On 09/25/2014 08:12 AM, Chen Gang wrote:
>> OK, thanks, next month, I shall try Qemu for microblaze (I also focus on Qemu, and try to make patches for it).
> Excuse me, after tried upstream qemu, it cann't run microblaze correctly,
> even for Xilinx qemu branch, I cann't run correctly either. I tried to
> consult related members in qemu mailing list, but got no result.
> After compared upstream branch and Xilinx branch, I am sure upstream
> microblaze qemu is lack of several related main features about
> microblaze.
> For qemu, I have to only focus on upstream, and try bug fix patches, new
> features and version merging are out of my current border, so sorry, I
> have to stop trying qemu for microblaze gcc test, at present.

Oh, sorry, after communicate with the qemu members, there are multiple
upstream branches for qemu, so I can only say: for upstream develop
master qemu, it does not support microblaze main features.

There are several upstream qemu distributions may support microblaze.  I
do not mainly focus on only using qemu, so I will stop qemu and try sim.

>>> So I guess the root cause is: I only use cross-compiling environments
>>> under fedora x86_64, no any real or virtual target for test.
>> Yes, if you want to test on a target, you will need a target.  You can either have a simulator (see binutils and sim/* for an example of how to write one) or target hardware in some form.

I will continue sim, and should try to finish within 2014-10-31. Sorry,
my other things which related open source maybe be delayed (maybe can
not finish within this month, if happens, need finish within next month).

> After trying sim, for me, it is really useful way for test, although I
> also met issues:
> For a hello world C program, microblaze gcc succeeded building, gdb can
> load and display the source code, and disassembe code successfully, but
> sim reported failure, the related issue is below:
>   [root@localhost test]# /upstream/release/bin/microblaze-gchen-linux-run ./test
>   Loading section .interp, size 0xd vma 0x100000f4
>   Loading section .note.ABI-tag, size 0x20 vma 0x10000104
>   Loading section .hash, size 0x24 vma 0x10000124
>   Loading section .dynsym, size 0x40 vma 0x10000148
>   Loading section .dynstr, size 0x3c vma 0x10000188
>   Loading section .gnu.version, size 0x8 vma 0x100001c4
>   Loading section .gnu.version_r, size 0x20 vma 0x100001cc
>   Loading section .rela.dyn, size 0x24 vma 0x100001ec
>   Loading section .rela.plt, size 0x24 vma 0x10000210
>   Loading section .init, size 0x58 vma 0x10000234
>   Loading section .plt, size 0x44 vma 0x1000028c
>   Loading section .text, size 0x3d0 vma 0x100002d0
>   Loading section .fini, size 0x34 vma 0x100006a0
>   Loading section .rodata, size 0x12 vma 0x100006d4
>   Loading section .eh_frame, size 0x4 vma 0x100006e8
>   Loading section .ctors, size 0x8 vma 0x100016ec
>   Loading section .dtors, size 0x8 vma 0x100016f4
>   Loading section .jcr, size 0x4 vma 0x100016fc
>   Loading section .dynamic, size 0xd0 vma 0x10001700
>   Loading section .got, size 0xc vma 0x100017d0
>   Loading section .got.plt, size 0x18 vma 0x100017dc
>   Loading section .data, size 0x10 vma 0x100017f4
>   Start address 0x100002d0
>   Transfer rate: 14424 bits in <1 sec.
>   ERROR: Unknown opcode
>   program stopped with signal 4.
> For me, I guess it is sim's issue, and I shall try to fix it in the next
> month, so sorry, I can not finish emulator for microblaze within this
> month. :-(
> Welcome any ideas, suggestions or completions.
> Thanks.

Chen Gang

Open, share, and attitude like air, water, and life which God blessed

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]