QEMU and kernel 2.6.15

QEMU and kernel 2.6.15

[Atsushi Nemoto]

Hi.  I'm a QEMU newbie.  Does anybody tried QEMU 0.8.0 with recent
linux-mips kernel ?

I got following output and the kernel hangs.

$ qemu-system-mips -kernel ../build-qemu/arch/mips/boot/vmlinux.bin -m 16 -nographic
(qemu) mips_r4k_init: start
mips_r4k_init: load BIOS '/usr/local/share/qemu/mips_bios.bin' size 131072
Linux version 2.6.15-gb3880251 (anemo@debian) (gcc version 3.4.5) #12 Tue Jan 10 23:50:55 JST 2006
CPU revision is: 00018000
Determined physical RAM map:
 memory: 01000000 @ 00000000 (usable)
On node 0 totalpages: 4096
  DMA zone: 4096 pages, LIFO batch:0
  DMA32 zone: 0 pages, LIFO batch:0
  Normal zone: 0 pages, LIFO batch:0
  HighMem zone: 0 pages, LIFO batch:0
Built 1 zonelists
Kernel command line: console=ttyS0 debug ip=172.20.0.2:172.20.0.1::255.255.0.0
Primary instruction cache 2kB, physically tagged, 2-way, linesize 16 bytes.
Primary data cache 2kB, 2-way, linesize 16 bytes.
Synthesized TLB refill handler (19 instructions).
Synthesized TLB load handler fastpath (31 instructions).
Synthesized TLB store handler fastpath (31 instructions).
Synthesized TLB modify handler fastpath (30 instructions).
PID hash table entries: 128 (order: 7, 2048 bytes)
Using 100.000 MHz high precision timer.
Console: colour VGA+ 80x25
Dentry cache hash table entries: 4096 (order: 2, 16384 bytes)
Inode-cache hash table entries: 2048 (order: 1, 8192 bytes)
Memory: 13916k/16384k available (1702k kernel code, 2468k reserved, 321k data, 112k init, 0k highmem)
Calibrating delay loop... 478.41 BogoMIPS (lpj=2392064)
Mount-cache hash table entries: 512
Checking for 'wait' instruction...  available.


I can get qemu prompt by typing C-a c and 'info register' shows
PC=0x80010d40 (qemu_handle_int+0xe0).

(qemu) info registers
pc=0x80010d40 HI=0x00000000 LO=0x00000000 ds 0000 801b6cd8 0
GPR00: r0 00000000 at 10008401 v0 8027fbe8 v1 00000000
GPR04: a0 80281f0c a1 80281e9c a2 8027fbe8 a3 80281f0c
GPR08: t0 10008400 t1 1000001f t2 00000000 t3 00000000
GPR12: t4 7fffffff t5 ffffffff t6 00100100 t7 7fffffff
GPR16: s0 00000002 s1 80281f08 s2 00000000 s3 00000000
GPR20: s4 00000000 s5 00000000 s6 00000000 s7 00000000
GPR24: t8 00000000 t9 00000001 k0 80281e80 k1 80281e80
GPR28: gp 80280000 sp 80281dd0 s8 80281e80 ra 801b74f8
CP0 Status  0x10008400 Cause   0x80000400 EPC    0x801b74f4
    Config0 0x80008081 Config1 0x1e190c8a LLAddr 0x00000001


Whant's going on ?  An interrupt is not cleared?

Any advise is welcome.  Thanks.
---
Atsushi Nemoto

Re: QEMU and kernel 2.6.15

[Ralf Baechle]

On Wed, Jan 11, 2006 at 12:24:31AM +0900, Atsushi Nemoto wrote:

> Whant's going on ?  An interrupt is not cleared?

Qemu is still work in progress and needs further debugging.

  Ralf

Re: QEMU and kernel 2.6.15

[P. Christeas]

On Tuesday 10 January 2006 5:24 pm, Atsushi Nemoto wrote:
> Hi.  I'm a QEMU newbie.  Does anybody tried QEMU 0.8.0 with recent
> linux-mips kernel ?
>
> I got following output and the kernel hangs.
>
>...
> Memory: 13916k/16384k available (1702k kernel code, 2468k reserved, 321k 
> data, 112k init, 0k highmem) Calibrating delay loop... 478.41 BogoMIPS
> (lpj=2392064)
> Mount-cache hash table entries: 512
> Checking for 'wait' instruction...  available.
>
>
However, this *does* resemble the bug I'm having in real hw.
I generally get instablilities nearly always after the memory mapping and 
'wait' instruction on a 4Kc core with 2.6.15.
I'm not a MIPS expert nor have I the time to learn its internals. So, anything 
that might give me a hint is worth investigating..

Re: QEMU and kernel 2.6.15

[Daniel Jacobowitz]

On Wed, Jan 11, 2006 at 12:24:31AM +0900, Atsushi Nemoto wrote:
> Hi.  I'm a QEMU newbie.  Does anybody tried QEMU 0.8.0 with recent
> linux-mips kernel ?

You've configured the kernel for QEMU, right?  And are usin QEMU from
CVS?

> Checking for 'wait' instruction...  available.
> 
> 
> I can get qemu prompt by typing C-a c and 'info register' shows
> PC=0x80010d40 (qemu_handle_int+0xe0).
> 
> (qemu) info registers
> pc=0x80010d40 HI=0x00000000 LO=0x00000000 ds 0000 801b6cd8 0
> GPR00: r0 00000000 at 10008401 v0 8027fbe8 v1 00000000
> GPR04: a0 80281f0c a1 80281e9c a2 8027fbe8 a3 80281f0c
> GPR08: t0 10008400 t1 1000001f t2 00000000 t3 00000000
> GPR12: t4 7fffffff t5 ffffffff t6 00100100 t7 7fffffff
> GPR16: s0 00000002 s1 80281f08 s2 00000000 s3 00000000
> GPR20: s4 00000000 s5 00000000 s6 00000000 s7 00000000
> GPR24: t8 00000000 t9 00000001 k0 80281e80 k1 80281e80
> GPR28: gp 80280000 sp 80281dd0 s8 80281e80 ra 801b74f8
> CP0 Status  0x10008400 Cause   0x80000400 EPC    0x801b74f4
>     Config0 0x80008081 Config1 0x1e190c8a LLAddr 0x00000001

It worked for me the last time I tried, but that was a couple
of weeks ago.  The port may have gotten broken...


-- 
Daniel Jacobowitz
CodeSourcery

Re: QEMU and kernel 2.6.15

[Atsushi Nemoto]

>>>>> On Wed, 11 Jan 2006 09:43:56 -0500, Daniel Jacobowitz <dan@debian.org> said:

dan> You've configured the kernel for QEMU, right?  And are usin QEMU
dan> from CVS?

Yes, I configured the kernel with qemu_defconfig.  I tried both QEMU
0.8.0 and tried current CVS today, but got same results.

Here is my instructions:

kernel:
make O=../build-qemu qemu_defconfig
make O=../build-qemu

QEMU:
./configure --target-list=mips-softmmu --disable-gfx-check
make

Then:
mips-softmmu/qemu-system-mips -kernel /home/git/build-qemu/arch/mips/boot/vmlinux.bin -m 16 -nographic

dan> It worked for me the last time I tried, but that was a couple of
dan> weeks ago.  The port may have gotten broken...

While QEMU 0.8.0 was released on Dec 19 and it seems there was not so
much changes on kernel's arch/mips in last few weeks, we should be
very close ...

---
Atsushi Nemoto

Re: QEMU and kernel 2.6.15

[Daniel Jacobowitz]

On Thu, Jan 12, 2006 at 12:09:04AM +0900, Atsushi Nemoto wrote:
> mips-softmmu/qemu-system-mips -kernel /home/git/build-qemu/arch/mips/boot/vmlinux.bin -m 16 -nographic

Just to check, could you try -m 32 or -m 128?  It shouldn't rely on
more than 16MB, but the boundary condition may be wrong.

Beyond that, I have no idea what might be wrong.

-- 
Daniel Jacobowitz
CodeSourcery

Re: QEMU and kernel 2.6.15

[Atsushi Nemoto]

>>>>> On Wed, 11 Jan 2006 15:13:54 -0500, Daniel Jacobowitz <dan@debian.org> said:

dan> Just to check, could you try -m 32 or -m 128?  It shouldn't rely
dan> on more than 16MB, but the boundary condition may be wrong.

dan> Beyond that, I have no idea what might be wrong.

I tried both but it did not help.  Thank you anyway.
---
Atsushi Nemoto

Re: QEMU and kernel 2.6.15

[Atsushi Nemoto]

>>>>> On Tue, 10 Jan 2006 15:36:52 +0000, Ralf Baechle <ralf@linux-mips.org> said:

ralf> Qemu is still work in progress and needs further debugging.

Do you mean QEMU itself or Qemu port of the kernel?

It looks timer setting in arch/mips/qemu is somewhat broken.

qemu_timer_setup() installs the timer interrupt handler on irq 0, but
do_qemu_int() also calls ll_timer_interrupt() directly.

Which timer interrupt source does the kernel expect?  r4k counter
interrupt or i8259 timer?

---
Atsushi Nemoto