[PATCH 0/2] Add C-SKY support
Mao Han
han_mao@c-sky.com
Thu Apr 4 06:19:00 GMT 2019
Hi,
On Wed, Apr 03, 2019 at 10:54:11PM +0200, Mark Wielaard wrote:
> Hi,
>
> On Mon, Apr 01, 2019 at 10:31:55AM +0800, Mao Han wrote:
> > This patch add backend support for C-SKY architecture.
>
> Thanks. I don't know anything about the C-SKY architecture. Do you
> have an URL to the ELF/DWARF supplemental spec that describes the ELF
> structures and DWARF register numbers C-SKY uses? It would be good to
> have that in the commit message and/or in a comment in the backend.
ELF structures can be found inside the ABI manual.
CSKY V2 ABI manual:
https://github.com/c-sky/csky-doc/blob/master/C-SKY_V2_CPU_Applications_Binary_I
+nterface_Standards_Manual.pdf
CSKY architecture user guide:
https://github.com/c-sky/csky-doc/blob/master/CSKY%20Architecture%20user_guide.p
+df
The DWARF part of the document is out of date. GCC source code can
be used as reference:
const int csky_dbx_regno[FIRST_PSEUDO_REGISTER] =
{
0, 1, 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15,
16, 17, 18, 19, 20, 21, 22, 23,
24, 25, 26, 27, 28, 29, 30, 31,
-1, -1, 36, 37, 35, 38, 39, 40,
41, 42, 43, 44, 45, 46, 47, 48,
49, 50, 51, 52, 53, 54, 55, 56,
57, 58, 59, 60, 61, 62, 63, 64,
65, 66, 67, 68, -1, -1, 72
};
The number in the array is the DWARF register numbers, the corresponding
register can be found in the corresponding below.
#define FIXED_REGISTERS \
/* r0 r1 r2 r3 r4 r5 r6 r7 */ \
{ 0, 0, 0, 0, 0, 0, 0, 0, \
/* r8 r9 r10 r11 r12 r13 r14 r15 */ \
0, 0, 0, 0, 0, 0, 1, 0, \
/* r16 r17 r18 r19 r20 r21 r22 r23 */ \
0, 0, 0, 0, 0, 0, 0, 0, \
/* r24 r25 r26 r27 r28 r29 r30 tls */ \
0, 0, 1, 1, 1, 1, 1, 1, \
/* reserve c hi lo */ \
1, 1, 0, 0, \
/* reserved */ \
1, 1, 1, 1, 1, 1, 1, 1, \
1, 1, 1, 1, 1, 1, 1, 1, \
/* vr0 vr1 vr2 vr3 vr4 vr5 vr6 vr7 */ \
0, 0, 0, 0, 0, 0, 0, 0, \
/* vr8 vr9 vr10 vr11 vr12 vr13 vr14 vr15 */ \
0, 0, 0, 0, 0, 0, 0, 0 , \
/* reserved */ \
1, 1, \
/* epc */ \
1 \
}
These code come from gcc/config/csky/csky.c and gcc/config/csky/csky.h:
https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/config/csky/csky.c;h=e4ac93323c0
+287d8d3d8191d51260983490fcac1;hb=73c7feb34d49e52ca5b6c2342cda882a9b06a036
> In general please read the CONTRIBUTING file and add a Signed-off-by
> line if you agree with and can contribute the patch under the
> appropriate licenses. Please also add a copyright line to the new
> files.
>
Ok, I'll add the Signed-off-by line and the copyright.
> Does qemu support full system or user-mode only C-SKY?
>
Yes, we have support for both system-mode and user-mode.
Our linux system is available at:
https://gitlab.com/c-sky/buildroot
Our linux ci enviroment:
https://gitlab.com/c-sky/buildroot/-/jobs/187862406
> > FAIL: run-exprlocs-self.sh
> > ==========================
> >
> > /home/vmh/elfutilswk/tests/varlocs: dwarf_getlocation_attr for addr: +.debug_addr section missing
> > *** failure in /home/vmh/elfutilswk/tests/varlocs --exprlocs -e +/home/vmh/elfutilswk/src/readelf
> > FAIL run-exprlocs-self.sh (exit status: 1)
>
> This is somewhat surprising. Do you build with -gdwarf-5? If you
> could make the /home/vmh/elfutilswk/src/readelf binary available
> somewhere we could investigate what goes on.
>
The default compile option is dwarf 4. I'v upload the binary to
https://github.com/c-sky/test-result/blob/master/elfutils/readelf
As I used the non-upstream version gcc to do the test, the machine
code is 39(EM_RCE) not 252(EM_CSKY).
> >
> > biarch testing disabled
> > SKIP run-backtrace-native-core-biarch.sh (exit status: 77)
>
> I haven't looked at the code yet, so maybe none of the backtrace tests
> would work even if you could run in "full system mode". If you do have
> a way to generate a C-SKY core file it would be interesting to see if
> you can get a backtrace from it using eu-stack.
>
Ok, I'll try if I can get a backtrace with C-SKY core file.
Thanks,
Mao Han
More information about the Elfutils-devel
mailing list