Skip to contentSkip to conversation

regression for m68k/coldfire

DISCUSSION Waldemar Brodkorb 5 participants 10 messages Feb 2, 2017 – Feb 6, 2017

lore
Share & tools
lore
#0 lore raw

Hi,

your commit 80cca775cdc4f8555612d2943a2872076b33e0ff breaks Linux booting in qemu-system-m68k:

qemu-system-m68k -nographic -M mcf5208evb -cpu m5208 -kernel qemu-m68k-mcf5208-initramfspiggyback-kernel

[    0.000000] Linux version 4.9.6-1 (wbx@vopenadk) (gcc version
5.4.0 (GCC) ) #2 Thu Feb 2 21:08:59 CET 2017
[    0.000000] uClinux with CPU COLDFIRE(m520x)
[    0.000000] COLDFIRE port done by Greg Ungerer, [email protected]
[    0.000000] Flat model support (C) 1998,1999 Kenneth Albanowski,
D. Jeff Dionne
[    0.000000] Built 1 zonelists in Zone order, mobility grouping
on.  Total pages: 16320
[    0.000000] Kernel command line: console=ttyS0,115200 ro
rootfstype=tmpfs
[    0.000000] PID hash table entries: 512 (order: -2, 2048 bytes)
[    0.000000] Dentry cache hash table entries: 16384 (order: 3,
65536 bytes)
[    0.000000] Inode-cache hash table entries: 8192 (order: 2, 32768
bytes)
[    0.000000] Memory: 127920K/131072K available (1292K kernel code,
96K rwdata, 216K rodata, 560K init, 179K bss, 3152K reserved, 0K
cma-reserved)
[    0.000000] Virtual kernel memory layout:
[    0.000000]     vector  : 0x40000000 - 0x40000400   (   1 KiB)
[    0.000000]     kmap    : 0x00000000 - 0xffffffff   (4095 MiB)
[    0.000000]     vmalloc : 0x00000000 - 0xffffffff   (4095 MiB)
[    0.000000]     lowmem  : 0x40000000 - 0x48000000   ( 128 MiB)
[    0.000000]       .init : 0x401b2000 - 0x4023e000   ( 560 KiB)
[    0.000000]       .text : 0x40020000 - 0x40163350   (1293 KiB)
[    0.000000]       .data : 0x40163350 - 0x401b1640   ( 313 KiB)
[    0.000000]       .bss  : 0x4023e000 - 0x4026af04   ( 180 KiB)
[    0.000000] SLUB: HWalign=16, Order=0-3, MinObjects=0, CPUs=1,
Nodes=8
[    0.000000] NR_IRQS:256
[    0.000000] clocksource: pit: mask: 0xffffffff max_cycles:
0xffffffff, max_idle_ns: 3669622406313 ns
[    0.020000] Calibrating delay loop... 238.38 BogoMIPS
(lpj=1191936)
[    0.080000] pid_max: default: 4096 minimum: 301
[    0.080000] Mount-cache hash table entries: 2048 (order: 0, 8192
bytes)
[    0.080000] Mountpoint-cache hash table entries: 2048 (order: 0,
8192 bytes)
[    0.160000] clocksource: jiffies: mask: 0xffffffff max_cycles:
0xffffffff, max_idle_ns: 19112604462750000 ns
[    0.170000] NET: Registered protocol family 16
[    0.200000] pps_core: LinuxPPS API ver. 1 registered
[    0.200000] pps_core: Software ver. 5.3.6 - Copyright 2005-2007
Rodolfo Giometti <[email protected]>
[    0.200000] PTP clock support registered
[    0.220000] clocksource: Switched to clocksource pit
[    0.230000] NET: Registered protocol family 2
[    0.240000] TCP established hash table entries: 2048 (order: 0,
8192 bytes)
[    0.240000] TCP bind hash table entries: 2048 (order: 0, 8192
bytes)
[    0.240000] TCP: Hash tables configured (established 2048 bind
2048)
[    0.240000] UDP hash table entries: 512 (order: 0, 8192 bytes)
[    0.240000] UDP-Lite hash table entries: 512 (order: 0, 8192
bytes)
[    0.250000] NET: Registered protocol family 1
[    2.560000] random: fast init done
[    3.400000] futex hash table entries: 16 (order: -6, 192 bytes)
[    3.400000] workingset: timestamp_bits=27 max_order=14
bucket_order=0
[    3.460000] ColdFire internal UART serial driver
[    3.470000] mcfuart.0: ttyS0 at MMIO 0xfc060000 (irq = 90,
base_baud = 2083333) is a ColdFire UART
[    3.500000] console [ttyS0] enabled
[    3.500000] mcfuart.0: ttyS1 at MMIO 0xfc064000 (irq = 91,
base_baud = 2083333) is a ColdFire UART
[    3.510000] mcfuart.0: ttyS2 at MMIO 0xfc068000 (irq = 92,

base_baud = 2083333) is a ColdFire UART
qemu: hardware error: mcf_fec_read: Bad address 0x200

CPU #0:

D0 = 00000200   A0 = 4780fdec   F0 = 0000000000000000 (           0)
D1 = 4780ff94   A1 = 401703de   F1 = 0000000000000000 (           0)
D2 = 46dbb000   A2 = 4780f800   F2 = 0000000000000000 (           0)
D3 = 4019d612   A3 = fc030200   F3 = 0000000000000000 (           0)
D4 = 00000000   A4 = 4019d608   F4 = 0000000000000000 (           0)
D5 = 40043f7a   A5 = 400222a0   F5 = 0000000000000000 (           0)
D6 = 400df204   A6 = 4783de78   F6 = 0000000000000000 (           0)
D7 = 401701c0   A7 = 4783de28   F7 = 0000000000000000 (           0)
PC = 400e5ff2   SR = 2009 -N--C FPRESULT =            0
Aborted
wbx@vopenadk:~/m68k/openadk$     
wbx@vopenadk:~/m68k/openadk$ qemu-system-m68k --version
QEMU emulator version 2.8.0
Copyright (c) 2003-2016 Fabrice Bellard and the QEMU Project
developers

Any ideas how to fix it?

best regards
Waldemar

_______________________________________________
uClinux-dev mailing list [email protected]
http://mailman.uclinux.org/mailman/listinfo/uclinux-dev
This message was resent by [email protected] To unsubscribe see:
http://mailman.uclinux.org/mailman/options/uclinux-dev

#1 lore raw
Quoted 4 lines

Hi,

your commit 80cca775cdc4f8555612d2943a2872076b33e0ff breaks Linux booting in qemu-system-m68k:

Hi.

Do you compile with CONFIG_M5272?

Is f85de6666347c974cdf97b1026180995d912d7d0 also in your tree?

Nikita

#2 lore raw
In reply to Waldemar Brodkorb

Hi Waldemar,

Quoted 4 lines · Waldemar Brodkorb

your commit 80cca775cdc4f8555612d2943a2872076b33e0ff breaks Linux booting in qemu-system-m68k:

qemu-system-m68k -nographic -M mcf5208evb -cpu m5208 -kernel qemu-m68k-mcf5208-initramfspiggyback-kernel

[snip]

Quoted 27 lines
[    3.460000] ColdFire internal UART serial driver
[    3.470000] mcfuart.0: ttyS0 at MMIO 0xfc060000 (irq = 90,
base_baud = 2083333) is a ColdFire UART
[    3.500000] console [ttyS0] enabled
[    3.500000] mcfuart.0: ttyS1 at MMIO 0xfc064000 (irq = 91,
base_baud = 2083333) is a ColdFire UART
[    3.510000] mcfuart.0: ttyS2 at MMIO 0xfc068000 (irq = 92,

base_baud = 2083333) is a ColdFire UART
qemu: hardware error: mcf_fec_read: Bad address 0x200

CPU #0:

D0 = 00000200   A0 = 4780fdec   F0 = 0000000000000000 (           0)
D1 = 4780ff94   A1 = 401703de   F1 = 0000000000000000 (           0)
D2 = 46dbb000   A2 = 4780f800   F2 = 0000000000000000 (           0)
D3 = 4019d612   A3 = fc030200   F3 = 0000000000000000 (           0)
D4 = 00000000   A4 = 4019d608   F4 = 0000000000000000 (           0)
D5 = 40043f7a   A5 = 400222a0   F5 = 0000000000000000 (           0)
D6 = 400df204   A6 = 4783de78   F6 = 0000000000000000 (           0)
D7 = 401701c0   A7 = 4783de28   F7 = 0000000000000000 (           0)
PC = 400e5ff2   SR = 2009 -N--C FPRESULT =            0
Aborted
wbx@vopenadk:~/m68k/openadk$     
wbx@vopenadk:~/m68k/openadk$ qemu-system-m68k --version
QEMU emulator version 2.8.0
Copyright (c) 2003-2016 Fabrice Bellard and the QEMU Project
developers
Quoted 1 line

Any ideas how to fix it?

This is a limitation in the FEC support in QEMU. This works on real ColdFire hardware (which do support the FEC MIB stats registers from offset 0x200 - so not 5272). I sent this patch to the qemu dev list a couple of weeks back which fixes qemu:

http://lists.gnu.org/archive/html/qemu-devel/2017-01/msg01781.html

I do not believe it has been picked up by QEMU mainline yet (I will probably have to resend and push a little to get that done).

Regards
Greg

_______________________________________________
uClinux-dev mailing list [email protected]
http://mailman.uclinux.org/mailman/listinfo/uclinux-dev
This message was resent by [email protected] To unsubscribe see:
http://mailman.uclinux.org/mailman/options/uclinux-dev

#3 lore raw
Quoted 10 lines · Greg Ungerer

This is a limitation in the FEC support in QEMU. This works on real ColdFire hardware (which do support the FEC MIB stats registers from offset 0x200 - so not 5272). I sent this patch to the qemu dev list a couple of weeks back which fixes qemu:

http://lists.gnu.org/archive/html/qemu-devel/2017-01/msg01781.html

I do not believe it has been picked up by QEMU mainline yet (I will probably have to resend and push a little to get that done).

QEMU upstream can sometimes take a while before they are merging patches, but usually it helps containing the maintainer of this part of the source tree directly.

Unfortunately, mcf5208 is currently unmaintained [1]:

mcf5208
S: Orphan
F: hw/m68k/mcf5208.c
F: hw/m68k/mcf_intc.c
F: hw/char/mcf_uart.c
F: hw/net/mcf_fec.c

But you can try getting into touch with Laurent Vivier who is the new maintainer of the m68k target. I have CC'ed him.

Adrian

Quoted 1 line

[1] http://git.qemu.org/?p=qemu.git;a=blob;f=MAINTAINERS

Signature · 4 lines
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - [email protected]
`. `'   Freie Universitaet Berlin - [email protected]
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913
#4 lore raw
Quoted 25 lines · John Paul Adrian Glaubitz
Quoted 10 lines · Greg Ungerer

This is a limitation in the FEC support in QEMU. This works on real ColdFire hardware (which do support the FEC MIB stats registers from offset 0x200 - so not 5272). I sent this patch to the qemu dev list a couple of weeks back which fixes qemu:

http://lists.gnu.org/archive/html/qemu-devel/2017-01/msg01781.html

I do not believe it has been picked up by QEMU mainline yet (I will probably have to resend and push a little to get that done).

QEMU upstream can sometimes take a while before they are merging patches, but usually it helps containing the maintainer of this part of the source tree directly.

Unfortunately, mcf5208 is currently unmaintained [1]:

mcf5208
S: Orphan
F: hw/m68k/mcf5208.c
F: hw/m68k/mcf_intc.c
F: hw/char/mcf_uart.c
F: hw/net/mcf_fec.c

Quoted 7 lines

But you can try getting into touch with Laurent Vivier who is the new maintainer of the m68k target. I have CC'ed him.

Adrian

Quoted 1 line

[1] http://git.qemu.org/?p=qemu.git;a=blob;f=MAINTAINERS

Use scripts/get_maintainer.pl on your patch to find the people.

This patch is on the network part, so you should cc:
Jason Wang <[email protected]> (odd fixer:Network devices)

If you cc me, you add more chances to have a review ;)

Thanks,
Laurent

#5 lore raw

Hi,
Laurent Vivier wrote,

Quoted 34 lines · John Paul Adrian Glaubitz
Quoted 25 lines · Greg Ungerer
Quoted 10 lines

This is a limitation in the FEC support in QEMU. This works on real ColdFire hardware (which do support the FEC MIB stats registers from offset 0x200 - so not 5272). I sent this patch to the qemu dev list a couple of weeks back which fixes qemu:

http://lists.gnu.org/archive/html/qemu-devel/2017-01/msg01781.html

I do not believe it has been picked up by QEMU mainline yet (I will probably have to resend and push a little to get that done).

QEMU upstream can sometimes take a while before they are merging patches, but usually it helps containing the maintainer of this part of the source tree directly.

Unfortunately, mcf5208 is currently unmaintained [1]:

mcf5208
S: Orphan
F: hw/m68k/mcf5208.c
F: hw/m68k/mcf_intc.c
F: hw/char/mcf_uart.c
F: hw/net/mcf_fec.c

Quoted 7 lines

But you can try getting into touch with Laurent Vivier who is the new maintainer of the m68k target. I have CC'ed him.

Adrian

Quoted 1 line

[1] http://git.qemu.org/?p=qemu.git;a=blob;f=MAINTAINERS

Quoted 6 lines

Use scripts/get_maintainer.pl on your patch to find the people.

This patch is on the network part, so you should cc:
Jason Wang <[email protected]> (odd fixer:Network devices)

If you cc me, you add more chances to have a review ;)

I tried the patch on top of 2.8.0 qemu release and it works for me with Linux Kernel 4.9.
Thanks for the hint and patch!
I hope Greg's patch get included in the next qemu release.

Btw: Laurent, are you m68k with mmu support are going to be included upstream? I always carry an old binary for any m68k with mmu testing.

best regards
Waldemar

#6 lore raw
Quoted 39 lines · Waldemar Brodkorb

Hi,
Laurent Vivier wrote,

Quoted 36 lines · John Paul Adrian Glaubitz
Quoted 25 lines · Greg Ungerer
Quoted 10 lines

This is a limitation in the FEC support in QEMU. This works on real ColdFire hardware (which do support the FEC MIB stats registers from offset 0x200 - so not 5272). I sent this patch to the qemu dev list a couple of weeks back which fixes qemu:

http://lists.gnu.org/archive/html/qemu-devel/2017-01/msg01781.html

I do not believe it has been picked up by QEMU mainline yet (I will probably have to resend and push a little to get that done).

QEMU upstream can sometimes take a while before they are merging patches, but usually it helps containing the maintainer of this part of the source tree directly.

Unfortunately, mcf5208 is currently unmaintained [1]:

mcf5208
S: Orphan
F: hw/m68k/mcf5208.c
F: hw/m68k/mcf_intc.c
F: hw/char/mcf_uart.c
F: hw/net/mcf_fec.c

Quoted 7 lines

But you can try getting into touch with Laurent Vivier who is the new maintainer of the m68k target. I have CC'ed him.

Adrian

Quoted 1 line

[1] http://git.qemu.org/?p=qemu.git;a=blob;f=MAINTAINERS

Use scripts/get_maintainer.pl on your patch to find the people.

Quoted 7 lines
Quoted 4 lines

This patch is on the network part, so you should cc:
Jason Wang <[email protected]> (odd fixer:Network devices)

If you cc me, you add more chances to have a review ;)

I tried the patch on top of 2.8.0 qemu release and it works for me with Linux Kernel 4.9.

If you can send a "Tested-by:" to the mailing list, it would be great.

Quoted 5 lines

Thanks for the hint and patch!
I hope Greg's patch get included in the next qemu release.

Btw: Laurent, are you m68k with mmu support are going to be included upstream? I always carry an old binary for any m68k with mmu testing.

I'm working to have the FPU included for now, that will allow to have the linux-user qemu enabled for 680x0 upstream.

I have a 68040 MMU implementation that I will send after this step. I didn't have a look to the ColdFire MMU (is there one?), but as 68040 is not the same as the 68030 one, I don't expect it works with coldfire.

Laurent

#7 lore raw

Hi Laurent,

Quoted 1 line · Laurent Vivier

Le 03/02/2017 à 16:17, Waldemar Brodkorb a écrit :

[snip]

Quoted 9 lines
Quoted 2 lines

Btw: Laurent, are you m68k with mmu support are going to be included upstream? I always carry an old binary for any m68k with mmu testing.

I'm working to have the FPU included for now, that will allow to have the linux-user qemu enabled for 680x0 upstream.

I have a 68040 MMU implementation that I will send after this step. I didn't have a look to the ColdFire MMU (is there one?), but as 68040 is not the same as the 68030 one, I don't expect it works with coldfire.

There is a ColdFire MMU in some v4 ColdFire parts, for example the 547x, 548x, 5441x and 5445x families. It is not compatible with the older 68020/86030/68040 MMU. It is supported in mainline Linux - at least for the 547x and 548x parts.

It would be really nice to have support for ColdFire MMU in QEMU :-)

Regards
Greg
_______________________________________________
uClinux-dev mailing list [email protected]
http://mailman.uclinux.org/mailman/listinfo/uclinux-dev
This message was resent by [email protected] To unsubscribe see:
http://mailman.uclinux.org/mailman/options/uclinux-dev

#8 lore raw
Quoted 15 lines · Greg Ungerer

Hi Laurent,

Quoted 1 line · Laurent Vivier

Le 03/02/2017 à 16:17, Waldemar Brodkorb a écrit :

[snip]

Quoted 9 lines
Quoted 2 lines

Btw: Laurent, are you m68k with mmu support are going to be included upstream? I always carry an old binary for any m68k with mmu testing.

I'm working to have the FPU included for now, that will allow to have the linux-user qemu enabled for 680x0 upstream.

I have a 68040 MMU implementation that I will send after this step. I didn't have a look to the ColdFire MMU (is there one?), but as 68040 is not the same as the 68030 one, I don't expect it works with coldfire.

Quoted 6 lines

There is a ColdFire MMU in some v4 ColdFire parts, for example the 547x, 548x, 5441x and 5445x families. It is not compatible with the older 68020/86030/68040 MMU. It is supported in mainline Linux - at least for the 547x and 548x parts.

It would be really nice to have support for ColdFire MMU in QEMU :-)

I can work on this. Do you have a Coldfire distro I can install?

Thanks,
Laurent

#9 lore raw

Hi Laurent,

Quoted 16 lines · Laurent Vivier
Quoted 15 lines · Greg Ungerer

Hi Laurent,

Quoted 1 line · Laurent Vivier

Le 03/02/2017 à 16:17, Waldemar Brodkorb a écrit :

[snip]

Quoted 9 lines
Quoted 2 lines

Btw: Laurent, are you m68k with mmu support are going to be included upstream? I always carry an old binary for any m68k with mmu testing.

I'm working to have the FPU included for now, that will allow to have the linux-user qemu enabled for 680x0 upstream.

I have a 68040 MMU implementation that I will send after this step. I didn't have a look to the ColdFire MMU (is there one?), but as 68040 is not the same as the 68030 one, I don't expect it works with coldfire.

Quoted 8 lines
Quoted 6 lines

There is a ColdFire MMU in some v4 ColdFire parts, for example the 547x, 548x, 5441x and 5445x families. It is not compatible with the older 68020/86030/68040 MMU. It is supported in mainline Linux - at least for the 547x and 548x parts.

It would be really nice to have support for ColdFire MMU in QEMU :-)

I can work on this. Do you have a Coldfire distro I can install?

This is the one I use:

http://www.uclinux.org/pub/uClinux/dist/

(toolchains for it at http://www.uclinux.org/pub/uClinux/m68k-elf-tools/)

Or if you don't want to compile it there is a ready to run binary for the m5475evb at:

http://www.uclinux.org/ports/coldfire/binary.html

Regards
Greg