Linux port failing on MIPS32 24Kc

Linux port failing on MIPS32 24Kc

[joe seb]

[-- Attachment #1: Type: text/plain, Size: 6519 bytes --]

Hi,

We are trying to port linux 2.6.29.4 version of the kernel from
linux-mips.org site to our MIPS 24K based platform and we see issues when we
use the cache in write-back mode. Cache with write-through configuration
works fine.
We use:
Linux kernel - 2.6.29.4
GNU cross tools - 4.3.2
Busybox - 1.14.1
U-boot - 2009.03

Our platform has 256MB of RAM and its mapped to second 256 MB of the KSEG0
(0x90000000 - 0x9FFFFFFF) and KSEG1 (0xB0000000 - 0xBFFFFFFF), and we
specify that "mem=16M@256M" as boot parameter (we just want to use the first
16MB by the kernel). The cache initialization for the KSEG0 is done in
u-boot.

The error we get when cache is configured as write-back is given below:

--------------------
Linux version 2.6.29.4 (gcc version 4.3.2 (Sourcery G
++ Lite 4.3-51) ) #11 PREEMPT Tue Jul 7 21:16:00 IST 2009
CPU revision is: 0101937c (MIPS 24Kc)
Determined physical RAM map:
User-defined physical RAM map:
 memory: 01000000 @ 10000000 (usable)
Wasting 2097152 bytes for tracking 65536 unused pages
Initrd not found or empty - disabling initrd
Zone PFN ranges:
  Normal   0x00010000 -> 0x00011000
Movable zone start PFN for each node
early_node_map[1] active PFN ranges
    0: 0x00010000 -> 0x00011000
Built 1 zonelists in Zone order, mobility grouping off.  Total pages: 4064
Kernel command line: mem=16M@256M console=ttyS0,9600 cca=3
Primary instruction cache 32kB, VIPT, 4-way, linesize 32 bytes.
Primary data cache 32kB, 4-way, PIPT, no aliases, linesize 32 bytes
Using cache attribute 3
Writing ErrCtl register=00000000
Readback ErrCtl register=00000000
PID hash table entries: 64 (order: 6, 256 bytes)
CPU frequency 50.00 MHz
console [ttyS0] enabled
Dentry cache hash table entries: 2048 (order: 1, 8192 bytes)
Inode-cache hash table entries: 1024 (order: 0, 4096 bytes)
Memory: 13776k/16384k available (1210k kernel code, 2608k reserved, 234k
data, 6
76k init, 0k highmem)
Calibrating delay loop... 33.17 BogoMIPS (lpj=165888)
Mount-cache hash table entries: 512
VFS: Disk quotas dquot_6.5.2
Dquot-cache hash table entries: 1024 (order 0, 4096 bytes)
msgmni has been set to 26
Serial: 8250/16550 driver, 2 ports, IRQ sharing enabled
serial8250: ttyS0 at MMIO 0xa1a30000 (irq = 188) is a 16550A
serial8250: ttyS1 at MMIO 0xa1a40000 (irq = 189) is a 16550A
Freeing unused kernel memory: 676k freed
Algorithmics/MIPS FPU Emulator v1.5
CPU 0 Unable to handle kernel paging request at virtual address cccccccc,
epc ==
 cccccccc, ra == cccccccc
Oops[#1]:
Cpu 0
$ 0   : 00000000 0053934a 00000000 00000001
$ 4   : 00000001 00000000 00000000 0000009a
$ 8   : 00000010 900038c0 00532730 00532730
$12   : 00000018 90284880 00020000 00000034
$16   : cccccccc 9016c320 9016aae0 90213520
$20   : 00000000 9016a960 00000001 90fa9680
$24   : 00000000 9009a414
$28   : 90176000 90177d58 9016aae0 cccccccc
Hi    : 0000000d
Lo    : 0000004a
epc   : cccccccc 0xcccccccc
    Not tainted
ra    : cccccccc 0xcccccccc
Status: 10004403    KERNEL EXL IE
Cause : 10800008
BadVA : cccccccc
PrId  : 0101937c (MIPS 24Kc)
Process rcS (pid: 11, threadinfo=90176000, task=90200500, tls=0053f470)
Stack : cccccccc cccccccc 90200500 9016aae0 90213520 00000000 90213520
00000000
        9016aae0 90213520 00000000 90025c28 90200500 90faf9c0 90200500
00000107
        9016aae0 00000107 9016aae0 900a04c8 00000003 902a4a40 902a4a48
00000000
        902a4a44 900b3ad8 90fb1880 90fa9680 00000000 00000003 90fb1880
90fa9680
        90177f30 90da7b20 cccccccc cccccccc cccccccc cccccccc cccccccc
cccccccc
        ...
Call Trace:
[<90025c28>] mmput+0x9c/0x194
[<900a04c8>] flush_old_exec+0x47c/0x988
[<900b3ad8>] alloc_fd+0x9c/0x1a4
[<90086c88>] handle_mm_fault+0x9a8/0x107c
[<9002f7c4>] do_softirq+0xc8/0xd0
[<900cc60c>] load_elf_binary+0x0/0x1410
[<9009fd9c>] search_binary_handler+0xa0/0x2bc
[<900a138c>] do_execve+0x298/0x300
[<900a4c60>] getname+0x28/0xc8
[<9000c714>] sys_execve+0x4c/0x78
[<90002398>] stack_done+0x20/0x3c

Code: (Bad address in epc)
do_cpu invoked from kernel context![#2]:
Cpu 0
$ 0   : 00000000 90210000 9016a98c 00000001
$ 4   : 00000002 00000003 90168468 00000000
$ 8   : 000007c4 00000004 9016846c 00000001
$12   : ffffff80 00000000 90136f7c 00000010
$16   : 00000000 00000000 90200500 90213520
$20   : 9016a994 90177ca8 00000000 90fa9680
$24   : 00000000 90121648
$28   : 90176000 90177b48 9016aae0 90fa9680
Hi    : 0098963b
Lo    : 38c9b600
epc   : 90fa9680 0x90fa9680
    Tainted: G      D
ra    : 90fa9680 0x90fa9680
Status: 10004403    KERNEL EXL IE
Cause : 1080002c
PrId  : 0101937c (MIPS 24Kc)
Process rcS (pid: 11, threadinfo=90176000, task=90200500, tls=0053f470)
Stack : 9016aae0 900041c4 00000000 00000000 90177ca8 90177ca8 0000000b
90200500
        90200500 cccccccc 00000000 9002cc78 90200658 9016a960 9020065c
90213520
        90152d44 90177ca8 00000001 cccccccc 00000000 90177ca8 90152d44
90177ca8
        90200500 cccccccc 00000000 90177ca8 00000000 90fa9680 9016aae0
9000e0d4
        cccccccc 90220000 ffffffff 00000e89 90177bec cccccccc 9016a960
90010f58
        ...
Call Trace:
[<900041c4>] printk+0x24/0x30
[<9002cc78>] do_exit+0xf4/0x88c
[<9000e0d4>] nmi_exception_handler+0x0/0x34
[<90010f58>] do_page_fault+0x2e0/0x350
[<90070234>] rmqueue_bulk+0x54/0xd8
[<900b0d48>] touch_atime+0xf8/0x174
[<9006c7e8>] generic_file_aio_read+0x4d8/0x8d8
[<90000404>] ret_from_exception+0x0/0x10
[<900038c0>] __bzero+0xc4/0x164
[<9009a414>] do_sync_read+0x0/0x168
[<90025c28>] mmput+0x9c/0x194
[<900a04c8>] flush_old_exec+0x47c/0x988
[<900b3ad8>] alloc_fd+0x9c/0x1a4
[<90086c88>] handle_mm_fault+0x9a8/0x107c
[<9002f7c4>] do_softirq+0xc8/0xd0
[<900cc60c>] load_elf_binary+0x0/0x1410
[<9009fd9c>] search_binary_handler+0xa0/0x2bc
[<900a138c>] do_execve+0x298/0x300
[<900a4c60>] getname+0x28/0xc8
[<9000c714>] sys_execve+0x4c/0x78
[<90002398>] stack_done+0x20/0x3c

Code: 040a001a  e5a8b400  4018e618 <464c457f> 00010101  00000000  00000000
0008
0002  00000001
Fixing recursive fault but reboot is needed!
-------------------------------

We get crashes at different places and the above crash is one of them.
Do you think this failure is due to the wrong cache configuration or related
to the d-cache aliasing problem?

The cache details of our platform:
D-cache: 32KB, 4-way, 32B line size, virtually indexed and physically
tagged, Config7[AR] bit is set (alias is removed by the hardware).
I-cache:  32KB, 4-way, 32B line size,virtually indexed and physically tagged


Is there any similar 24k platform supported in linux kernel which we are
refer for the configurations?


Thanks and Regards
Joe

[-- Attachment #2: Type: text/html, Size: 7442 bytes --]

Re: Linux port failing on MIPS32 24Kc

[Ralf Baechle]

On Wed, Jul 08, 2009 at 01:37:42PM +0530, joe seb wrote:

> We are trying to port linux 2.6.29.4 version of the kernel from
> linux-mips.org site to our MIPS 24K based platform and we see issues when we
> use the cache in write-back mode. Cache with write-through configuration
> works fine.
> We use:
> Linux kernel - 2.6.29.4
> GNU cross tools - 4.3.2
> Busybox - 1.14.1
> U-boot - 2009.03
> 
> Our platform has 256MB of RAM and its mapped to second 256 MB of the KSEG0
> (0x90000000 - 0x9FFFFFFF) and KSEG1 (0xB0000000 - 0xBFFFFFFF), and we
> specify that "mem=16M@256M" as boot parameter (we just want to use the first
> 16MB by the kernel). The cache initialization for the KSEG0 is done in
> u-boot.
> 
> The error we get when cache is configured as write-back is given below:
> 
> --------------------
> Linux version 2.6.29.4 (gcc version 4.3.2 (Sourcery G
> ++ Lite 4.3-51) ) #11 PREEMPT Tue Jul 7 21:16:00 IST 2009
> CPU revision is: 0101937c (MIPS 24Kc)
> Determined physical RAM map:
> User-defined physical RAM map:
>  memory: 01000000 @ 10000000 (usable)
> Wasting 2097152 bytes for tracking 65536 unused pages
> Initrd not found or empty - disabling initrd
> Zone PFN ranges:
>   Normal   0x00010000 -> 0x00011000
> Movable zone start PFN for each node
> early_node_map[1] active PFN ranges
>     0: 0x00010000 -> 0x00011000
> Built 1 zonelists in Zone order, mobility grouping off.  Total pages: 4064
> Kernel command line: mem=16M@256M console=ttyS0,9600 cca=3
> Primary instruction cache 32kB, VIPT, 4-way, linesize 32 bytes.
> Primary data cache 32kB, 4-way, PIPT, no aliases, linesize 32 bytes
> Using cache attribute 3
> Writing ErrCtl register=00000000
> Readback ErrCtl register=00000000
> PID hash table entries: 64 (order: 6, 256 bytes)
> CPU frequency 50.00 MHz
> console [ttyS0] enabled
> Dentry cache hash table entries: 2048 (order: 1, 8192 bytes)
> Inode-cache hash table entries: 1024 (order: 0, 4096 bytes)
> Memory: 13776k/16384k available (1210k kernel code, 2608k reserved, 234k
> data, 6
> 76k init, 0k highmem)
> Calibrating delay loop... 33.17 BogoMIPS (lpj=165888)
> Mount-cache hash table entries: 512
> VFS: Disk quotas dquot_6.5.2
> Dquot-cache hash table entries: 1024 (order 0, 4096 bytes)
> msgmni has been set to 26
> Serial: 8250/16550 driver, 2 ports, IRQ sharing enabled
> serial8250: ttyS0 at MMIO 0xa1a30000 (irq = 188) is a 16550A
> serial8250: ttyS1 at MMIO 0xa1a40000 (irq = 189) is a 16550A
> Freeing unused kernel memory: 676k freed
> Algorithmics/MIPS FPU Emulator v1.5
> CPU 0 Unable to handle kernel paging request at virtual address cccccccc,
> epc ==
>  cccccccc, ra == cccccccc
> Oops[#1]:
> Cpu 0
> $ 0   : 00000000 0053934a 00000000 00000001
> $ 4   : 00000001 00000000 00000000 0000009a
> $ 8   : 00000010 900038c0 00532730 00532730
> $12   : 00000018 90284880 00020000 00000034
> $16   : cccccccc 9016c320 9016aae0 90213520
> $20   : 00000000 9016a960 00000001 90fa9680
> $24   : 00000000 9009a414
> $28   : 90176000 90177d58 9016aae0 cccccccc
> Hi    : 0000000d
> Lo    : 0000004a
> epc   : cccccccc 0xcccccccc
>     Not tainted
> ra    : cccccccc 0xcccccccc
> Status: 10004403    KERNEL EXL IE
> Cause : 10800008
> BadVA : cccccccc
> PrId  : 0101937c (MIPS 24Kc)
> Process rcS (pid: 11, threadinfo=90176000, task=90200500, tls=0053f470)
> Stack : cccccccc cccccccc 90200500 9016aae0 90213520 00000000 90213520
> 00000000
>         9016aae0 90213520 00000000 90025c28 90200500 90faf9c0 90200500
> 00000107
>         9016aae0 00000107 9016aae0 900a04c8 00000003 902a4a40 902a4a48
> 00000000
>         902a4a44 900b3ad8 90fb1880 90fa9680 00000000 00000003 90fb1880
> 90fa9680
>         90177f30 90da7b20 cccccccc cccccccc cccccccc cccccccc cccccccc
> cccccccc
>         ...
> Call Trace:
> [<90025c28>] mmput+0x9c/0x194
> [<900a04c8>] flush_old_exec+0x47c/0x988
> [<900b3ad8>] alloc_fd+0x9c/0x1a4
> [<90086c88>] handle_mm_fault+0x9a8/0x107c
> [<9002f7c4>] do_softirq+0xc8/0xd0
> [<900cc60c>] load_elf_binary+0x0/0x1410
> [<9009fd9c>] search_binary_handler+0xa0/0x2bc
> [<900a138c>] do_execve+0x298/0x300
> [<900a4c60>] getname+0x28/0xc8
> [<9000c714>] sys_execve+0x4c/0x78
> [<90002398>] stack_done+0x20/0x3c
> 
> Code: (Bad address in epc)
> do_cpu invoked from kernel context![#2]:
> Cpu 0
> $ 0   : 00000000 90210000 9016a98c 00000001
> $ 4   : 00000002 00000003 90168468 00000000
> $ 8   : 000007c4 00000004 9016846c 00000001
> $12   : ffffff80 00000000 90136f7c 00000010
> $16   : 00000000 00000000 90200500 90213520
> $20   : 9016a994 90177ca8 00000000 90fa9680
> $24   : 00000000 90121648
> $28   : 90176000 90177b48 9016aae0 90fa9680
> Hi    : 0098963b
> Lo    : 38c9b600
> epc   : 90fa9680 0x90fa9680
>     Tainted: G      D
> ra    : 90fa9680 0x90fa9680
> Status: 10004403    KERNEL EXL IE
> Cause : 1080002c
> PrId  : 0101937c (MIPS 24Kc)
> Process rcS (pid: 11, threadinfo=90176000, task=90200500, tls=0053f470)
> Stack : 9016aae0 900041c4 00000000 00000000 90177ca8 90177ca8 0000000b
> 90200500
>         90200500 cccccccc 00000000 9002cc78 90200658 9016a960 9020065c
> 90213520
>         90152d44 90177ca8 00000001 cccccccc 00000000 90177ca8 90152d44
> 90177ca8
>         90200500 cccccccc 00000000 90177ca8 00000000 90fa9680 9016aae0
> 9000e0d4
>         cccccccc 90220000 ffffffff 00000e89 90177bec cccccccc 9016a960
> 90010f58
>         ...
> Call Trace:
> [<900041c4>] printk+0x24/0x30
> [<9002cc78>] do_exit+0xf4/0x88c
> [<9000e0d4>] nmi_exception_handler+0x0/0x34
> [<90010f58>] do_page_fault+0x2e0/0x350
> [<90070234>] rmqueue_bulk+0x54/0xd8
> [<900b0d48>] touch_atime+0xf8/0x174
> [<9006c7e8>] generic_file_aio_read+0x4d8/0x8d8
> [<90000404>] ret_from_exception+0x0/0x10
> [<900038c0>] __bzero+0xc4/0x164
> [<9009a414>] do_sync_read+0x0/0x168
> [<90025c28>] mmput+0x9c/0x194
> [<900a04c8>] flush_old_exec+0x47c/0x988
> [<900b3ad8>] alloc_fd+0x9c/0x1a4
> [<90086c88>] handle_mm_fault+0x9a8/0x107c
> [<9002f7c4>] do_softirq+0xc8/0xd0
> [<900cc60c>] load_elf_binary+0x0/0x1410
> [<9009fd9c>] search_binary_handler+0xa0/0x2bc
> [<900a138c>] do_execve+0x298/0x300
> [<900a4c60>] getname+0x28/0xc8
> [<9000c714>] sys_execve+0x4c/0x78
> [<90002398>] stack_done+0x20/0x3c
> 
> Code: 040a001a  e5a8b400  4018e618 <464c457f> 00010101  00000000  00000000
> 0008
> 0002  00000001
> Fixing recursive fault but reboot is needed!
> -------------------------------
> 
> We get crashes at different places and the above crash is one of them.
> Do you think this failure is due to the wrong cache configuration or related
> to the d-cache aliasing problem?
> 
> The cache details of our platform:
> D-cache: 32KB, 4-way, 32B line size, virtually indexed and physically
> tagged, Config7[AR] bit is set (alias is removed by the hardware).

Aliases disabled in hardware, yet you suspect aliases?  Doesn't make sense.

> I-cache:  32KB, 4-way, 32B line size,virtually indexed and physically tagged
> 
> 
> Is there any similar 24k platform supported in linux kernel which we are
> refer for the configurations?

There has been a kernel bug for a while which on platforms with a non-zero
memory start address would effectively disable part of of the cache code.
Your description above, including the changed behaviour between write-though
and write-back caches is consistent with that bug.  Commit
67227819d6dd07f6ec225ea59c67aff3ba936e25 fixes this issue.  For your
convenience I append it below.

I'd appreciate feedback on your test results with this patch.

(Why do people use non-zero starting addresses for memory?  Handling of
cache error exceptions is hard enough as it is but with no memory in the
low 32k the design idea of the cache architecture that stores relative to
$zero can be used goes down the drain and (not considering platform-specific
solutions here) only be handled by burning the scarce resource of a TLB
entry for an extremly rare event ...)

  Ralf

Re: Linux port failing on MIPS32 24Kc

[joe seb]

[-- Attachment #1.1: Type: text/plain, Size: 694 bytes --]

Hi,

We made the following changes and tried,

-> applied the patch given by you.
-> changed the PHYS_OFFSET to 0x10000000 to match our memory offset.
-> cache write back mode enabled

Still we face the same problem. It crashes at different points when it
enters the user space. I have attached one of the logs.

But in cache write through mode it works.

One comment about the patch, it tries to change the start of the valid pfn
range from ARCH_PFN_OFFSET to min_low_pfn.
But in bootmem_init() present in arch/mips/kernel/setup.c, line 324,
min_low_pfn = ARCH_PFN_OFFSET;

So appears like the previous pfn_valid check was also doing something
similar. Correct me if I am wrong.

Regards,
Joe

[-- Attachment #1.2: Type: text/html, Size: 790 bytes --]

[-- Attachment #2: log_1 --]
[-- Type: application/octet-stream, Size: 3635 bytes --]

Linux version 2.6.29.4 (gcc version 4.3.2 (Sourcery G
++ Lite 4.3-51) ) #8 PREEMPT Wed Jul 8 17:24:31 IST 2009
CPU revision is: 0101937c (MIPS 24Kc)
Determined physical RAM map:
User-defined physical RAM map:
 memory: 01000000 @ 10000000 (usable)
Initrd not found or empty - disabling initrd
Zone PFN ranges:
  Normal   0x00010000 -> 0x00011000
Movable zone start PFN for each node
early_node_map[1] active PFN ranges
    0: 0x00010000 -> 0x00011000
Built 1 zonelists in Zone order, mobility grouping off.  Total pages: 4064
Kernel command line: mem=16M@256M console=ttyS0,9600
Primary instruction cache 32kB, VIPT, 4-way, linesize 32 bytes.
Primary data cache 32kB, 4-way, PIPT, no aliases, linesize 32 bytes
Using cache attribute 3
Writing ErrCtl register=00000000
Readback ErrCtl register=00000000
PID hash table entries: 64 (order: 6, 256 bytes)
CPU frequency 50.00 MHz
console [ttyS0] enabled
Dentry cache hash table entries: 2048 (order: 1, 8192 bytes)
Inode-cache hash table entries: 1024 (order: 0, 4096 bytes)
Memory: 13768k/16384k available (1212k kernel code, 2616k reserved, 240k data, 6
76k init, 0k highmem)
Calibrating delay loop... 33.17 BogoMIPS (lpj=165888)
Mount-cache hash table entries: 512
VFS: Disk quotas dquot_6.5.2
Dquot-cache hash table entries: 1024 (order 0, 4096 bytes)
msgmni has been set to 26
Serial: 8250/16550 driver, 2 ports, IRQ sharing enabled
serial8250: ttyS0 at MMIO 0xa1a30000 (irq = 188) is a 16550A
serial8250: ttyS1 at MMIO 0xa1a40000 (irq = 189) is a 16550A
Freeing unused kernel memory: 676k freed
Algorithmics/MIPS FPU Emulator v1.5
vmalloc fault 
CPU 0 Unable to handle kernel paging request at virtual address ccccccf4, epc ==
 90081b5c, ra == 9008bd58 fo - 0
Oops[#1]:
Cpu 0
$ 0   : 00000000 90fab68a cccccccc ccccccf0
$ 4   : 9021241c 9021256c cccccccc 9021241c
$ 8   : 00000001 90240000 00400000 00000121
$12   : 00040000 00000000 92e9a1a3 3723702a
$16   : 90ddb58c 9021241c 90210c60 00000000
$20   : 90212490 90210c60 00000000 90de54a0
$24   : 00000000 00000000                  
$28   : 90204000 90205c48 08001875 9008bd58
Hi    : 0098963a
Lo    : ceb11100
epc   : 90081b5c vma_prio_tree_add+0x20/0x74
    Not tainted
ra    : 9008bd58 vma_link+0x8c/0x11c
Status: 10004403    KERNEL EXL IE 
Cause : 00800008
BadVA : ccccccf4
PrId  : 0101937c (MIPS 24Kc)
Process rcS (pid: 10, threadinfo=90204000, task=90202978, tls=0053f470)
Stack : 00000002 30e26b23 00000000 9021241c 90212488 900698f4 90212488 00400000
        90212490 9021241c 00000001 9008d818 90210c94 90106f58 90215520 90210c60
        90212488 00000000 90de54a0 00000000 00000000 00400000 00000121 90ddb4f4
        00000000 90210c94 00000000 fffff000 00400000 00121000 90205f30 90de54a0
        00000005 00000000 00000000 900cce58 90212470 00000000 00121000 00006012
        ...
Call Trace:
[<90081b5c>] vma_prio_tree_add+0x20/0x74
[<9008bd58>] vma_link+0x8c/0x11c
[<9008d818>] mmap_region+0x3a8/0x5c4
[<900cce58>] elf_map+0x148/0x1a0
[<900cd448>] load_elf_binary+0x598/0x1410
[<900a04ec>] search_binary_handler+0xa0/0x2bc
[<900cb710>] load_script+0x2d0/0x2e0
[<900a04ec>] search_binary_handler+0xa0/0x2bc
[<900a1adc>] do_execve+0x298/0x300
[<9000c69c>] sys_execve+0x4c/0x78
[<90002398>] stack_done+0x20/0x3c


Code: 8ca60030  10c00008  24c30024 <8cc20028> 24840024  ace30024  acc40028  ac44
0000  03e00008 
note: rcS[10] exited with preempt_count 1
BUG: scheduling while atomic: rcS/10/0x00000002
Call Trace:
[<90003fd8>] dump_stack+0x8/0x38
[<900048a8>] schedule+0x560/0x72c
[<9000782c>] __down_read+0xe8/0x188
[<9002a9b8>] exit_mm+0x3c/0x268
[<9002cec0>] do_exit+0xf4/0x88c
[<9000e05c>] nmi_exception_handler+0x0/0x34

Re: Linux port failing on MIPS32 24Kc

[David VomLehn]

On Wed, Jul 08, 2009 at 11:37:56AM +0100, Ralf Baechle wrote:
> On Wed, Jul 08, 2009 at 01:37:42PM +0530, joe seb wrote:
> 
> > We are trying to port linux 2.6.29.4 version of the kernel from
> > linux-mips.org site to our MIPS 24K based platform and we see issues when we
> > use the cache in write-back mode. Cache with write-through configuration
> > works fine.
...
> > CPU 0 Unable to handle kernel paging request at virtual address cccccccc,
> > epc ==
> >  cccccccc, ra == cccccccc
...
> > We get crashes at different places and the above crash is one of them.
> > Do you think this failure is due to the wrong cache configuration or related
> > to the d-cache aliasing problem?

In a narrow sense, when you get a page fault at an address that equals the
value of the EPC register and the ra register is the same value, you have
probably smashed your stack. Specifically, you smashed the location where
the ra value was stored on function entry. When you restored it to ra and
did a jr ra, you put the value in the pc register.

One possible cause is a stack overflow. This is unlikely, but I've seen
drivers put big enough buffers on the stack that it intruded into the
thread_info area. This causes all sorts of badness to happen.

> (Why do people use non-zero starting addresses for memory?  Handling of
> cache error exceptions is hard enough as it is but with no memory in the
> low 32k the design idea of the cache architecture that stores relative to
> $zero can be used goes down the drain and (not considering platform-specific
> solutions here) only be handled by burning the scarce resource of a TLB
> entry for an extremly rare event ...)

(Because hardware people are, uh, insufficiently acquainted with the MIPS
architecture and don't know what the conventions are. Or why they are that
way. Then, you're stuck with that architecture forever. Sigh.)

>   Ralf

David VomLehn

Re: Linux port failing on MIPS32 24Kc

[David VomLehn]

On Wed, Jul 08, 2009 at 06:17:42PM +0530, joe seb wrote:
> Hi,
> 
> We made the following changes and tried,
> 
> -> applied the patch given by you.
> -> changed the PHYS_OFFSET to 0x10000000 to match our memory offset.
> -> cache write back mode enabled
> 
> Still we face the same problem. It crashes at different points when it
> enters the user space. I have attached one of the logs.
> 
> But in cache write through mode it works.

Another thought: if it works with write through, but not write back, you
may have a device driver that isn't flushing/invalidating cache appropriately
before doing DMA. This is much more important on MIPS than many other
processors.

Is this a new platform or one that has been working for a while?

David VomLehn

Re: Linux port failing on MIPS32 24Kc

[Maciej W. Rozycki]

On Wed, 8 Jul 2009, David VomLehn wrote:

> > (Why do people use non-zero starting addresses for memory?  Handling of
> > cache error exceptions is hard enough as it is but with no memory in the
> > low 32k the design idea of the cache architecture that stores relative to
> > $zero can be used goes down the drain and (not considering platform-specific
> > solutions here) only be handled by burning the scarce resource of a TLB
> > entry for an extremly rare event ...)
> 
> (Because hardware people are, uh, insufficiently acquainted with the MIPS
> architecture and don't know what the conventions are. Or why they are that
> way. Then, you're stuck with that architecture forever. Sigh.)

 Many processor architectures have assumptions (i.e. power-up/reset 
defaults) about or even force the existence of RAM at physical address 0.  
With the MIPS architecture actually it has been pretty a recent addition 
in its long history that you are able to get away without.  Given these 
circumstances I am afraid your suspicion sounds like an understatement.

 With the MIPS64r2 architecture though and if you want more than 400+some 
MB of memory, placing RAM from the top of the physical address space seems 
more reasonable to me as you can avoid ROM popping up in the middle then.  

 For the sake of flexibility, including but not limited to ERL operation, 
the virtual-to-physical address translation for the CKSEG0 and CKSEG1 
space could be made configurable via an additional CP0 register.  I'm a 
little bit surprised nobody had thought about it at the time CP0.EBase was 
introduced.  The remapping has to be done anyway and it does not look to 
me as it would take a lot of power and/or silicon space to make it use a 
configurable set of latches rather than a predefined set of hardwired 
logical levels.

  Maciej

Re: Linux port failing on MIPS32 24Kc

[joe seb]

[-- Attachment #1: Type: text/plain, Size: 1956 bytes --]

Hi David, Ralf,

This is a new FPGA based platform which has MIPS 24k core and we are trying
to bring up linux on this. MIPS cpu is running at 50MHz.

Other than ram, we have UART available and its a 8250 compatible UART which
is connected to one of the HW interrupt line. The MIPS timer interrupt is
connected to HW4 interrupt line.  So in our platform file, we provide the
irq dispatch function for these two irq lines. There is no driver which uses
the DMA.

We did another experiment where we replaced init application with a memory
test code which malloc and run incremental pattern test on the allocated
memory and we see sometimes this test is failing. The failure is happening
for a cache line, when we dump ddr corresponding to that cache line, we see
that, that particular cache line is not flushed to ddr and all the other
cache lines are fine. So, not sure any issue with flushing the write-back.
We checked the linux code and did not see any place dcache is flushed for a
particular line, its always for a page(blast_dcache functions). Is our
finding correct?

Regards,
Joe

On Wed, Jul 8, 2009 at 11:59 PM, David VomLehn <dvomlehn@cisco.com> wrote:

> On Wed, Jul 08, 2009 at 06:17:42PM +0530, joe seb wrote:
> > Hi,
> >
> > We made the following changes and tried,
> >
> > -> applied the patch given by you.
> > -> changed the PHYS_OFFSET to 0x10000000 to match our memory offset.
> > -> cache write back mode enabled
> >
> > Still we face the same problem. It crashes at different points when it
> > enters the user space. I have attached one of the logs.
> >
> > But in cache write through mode it works.
>
> Another thought: if it works with write through, but not write back, you
> may have a device driver that isn't flushing/invalidating cache
> appropriately
> before doing DMA. This is much more important on MIPS than many other
> processors.
>
> Is this a new platform or one that has been working for a while?
>
> David VomLehn
>

[-- Attachment #2: Type: text/html, Size: 2460 bytes --]

Re: Linux port failing on MIPS32 24Kc

[Eugene Surovegin]

On Thu, Jul 09, 2009 at 04:35:30PM +0530, joe seb wrote:
> This is a new FPGA based platform which has MIPS 24k core and we are trying
> to bring up linux on this. MIPS cpu is running at 50MHz.

One thing to check is that your FPGA guys used correct memory models 
for caches, etc. We had very "funny" Liux failures when 24K FPGA 
emulation build was using async memories for caches insted of sync 
ones.

-- 
Eugene

Re: Linux port failing on MIPS32 24Kc

[joe seb]

[-- Attachment #1: Type: text/plain, Size: 15340 bytes --]

Ralf,

We made changes to our MIPS FPGA to map the RAM to KSEG0 (0x80000000) start
address. We still see the issue with the write-back. Write-through is
working fine.

The failure log we got this time is given below:

Linux version 2.6.29.4 () (gcc version 4.3.2 (Sourcery G
++ Lite 4.3-51) ) #15 PREEMPT Mon Jul 13 13:22:21 IST 2009
CPU revision is: 0101937c (MIPS 24Kc)
Determined physical RAM map:
User-defined physical RAM map:
 memory: 10000000 @ 00000000 (usable)
Initrd not found or empty - disabling initrd
Zone PFN ranges:
  Normal   0x00000000 -> 0x00010000
Movable zone start PFN for each node
early_node_map[1] active PFN ranges
    0: 0x00000000 -> 0x00010000
Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 65024
Kernel command line: mem=256M console=ttyS0,9600 cca=3
Primary instruction cache 32kB, VIPT, 4-way, linesize 32 bytes.
Primary data cache 32kB, 4-way, PIPT, no aliases, linesize 32 bytes
Using cache attribute 3
Writing ErrCtl register=00000000
Readback ErrCtl register=00000000
PID hash table entries: 1024 (order: 10, 4096 bytes)
CPU frequency 50.00 MHz
console [ttyS0] enabled
Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
Memory: 257092k/262144k available (1211k kernel code, 4728k reserved, 234k
data,
 692k init, 0k highmem)
Calibrating delay loop... 33.17 BogoMIPS (lpj=165888)
Mount-cache hash table entries: 512
VFS: Disk quotas dquot_6.5.2
Dquot-cache hash table entries: 1024 (order 0, 4096 bytes)
msgmni has been set to 502
Serial: 8250/16550 driver, 2 ports, IRQ sharing enabled
serial8250: ttyS0 at MMIO 0xb1a30000 (irq = 188) is a 16550A
serial8250: ttyS1 at MMIO 0xb1a40000 (irq = 189) is a 16550A
Freeing unused kernel memory: 692k freed
Algorithmics/MIPS FPU Emulator v1.5
CPU 0 Unable to handle kernel paging request at virtual address ffffb32c,
epc ==
 8f15b32c, ra == 8f15b32c
CPU 0 Unable to handle kernel paging request at virtual address 0000001c,
epc ==
 800210e4, ra == 8002120c
Oops[#1]:
Cpu 0
$ 0   : 00000000 7fccc8e8 fffffff8 00000000
$ 4   : 8021dbf0 00000000 8f81fc60 80220000
$ 8   : 00000000 9b915d6d 8021dbc0 fffffffc
$12   : 0000000f 00000002 80136f7c 00000010
$16   : 00000000 00000000 8021dbc0 8f821eb0
$20   : 00000000 00000000 00000003 ffffffff
$24   : 00000000 0043e4b0
$28   : 8f820000 8f821e00 8021dbc0 8002120c
Hi    : 0098963a
Lo    : ebe5df80
epc   : 800210e4 set_next_entity+0x14/0x8c
    Not tainted
ra    : 8002120c pick_next_task_fair+0x90/0xe4
Status: 10004402    KERNEL EXL
Cause : 00800008
BadVA : 0000001c
PrId  : 0101937c (MIPS 24Kc)
Process init (pid: 1, threadinfo=8f820000, task=8f81fb08, tls=0053f470)
Stack : 00000000 800221a8 00000000 80021550 00000000 00000000 8021dbc0
8002120c
        8f81fb08 80020a90 8f81fb00 8f81fb08 80130000 8f81fb08 8f81fb08
80004924
        7fcccc04 800051a8 8f81fc60 8f168de0 8f81fc60 8f821eb4 8f81fc60
00000001
        8f81fb00 8f81fb08 8f81fbfc 8f821eb0 00000000 00000000 00000003
ffffffff
        00000006 8002c95c 8021dbc0 8f15b238 00000001 8f81fb08 00000003
00000000
        ...
Call Trace:
[<800210e4>] set_next_entity+0x14/0x8c
[<8002120c>] pick_next_task_fair+0x90/0xe4
[<80004924>] schedule+0x5e4/0x72c
[<8002c95c>] do_wait+0x31c/0x4a0
[<8002cbe0>] sys_wait4+0x100/0x174
[<80002398>] stack_done+0x20/0x3c

Code: afb10014  afbf001c  afb00010 <8ca2001c> 00a08821  10400008  00809021
8c82
0024  24b00008
note: init[1] exited with preempt_count 2
BUG: scheduling while atomic: init/1/0x10000003
Call Trace:
[<80003fd0>] dump_stack+0x8/0x38
[<800048a0>] schedule+0x560/0x72c
[<80021774>] __cond_resched+0x18/0x38
[<80004c20>] _cond_resched+0x50/0x58
[<80085874>] unmap_vmas+0x614/0x6d4
[<8008b260>] exit_mmap+0xe8/0x1f8
[<80025dd8>] mmput+0x9c/0x194
[<8002aa40>] exit_mm+0x15c/0x268
[<8002ce28>] do_exit+0xf4/0x88c
[<8000e110>] nmi_exception_handler+0x0/0x34
CPU 0 Unable to handle kernel paging request at virtual address 0000001c,
epc ==
 800210e4, ra == 8002120c
Oops[#2]:
Cpu 0
$ 0   : 00000000 00000001 fffffff8 00000000
$ 4   : 8021dbf0 00000000 8f81eef8 80220000
$ 8   : 00000000 9b915d6d 8021dbc0 ffff8db3
$12   : 00001b5c 80138ef0 80136f7c 00000010
$16   : 00000000 00000000 8021dbc0 00000000
$20   : 00000000 00000000 00000000 00000000
$24   : 00000019 80121a84
$28   : 8fb08000 8fb09ee8 8021dbc0 8002120c
Hi    : 0098963b
Lo    : 67e02780
epc   : 800210e4 set_next_entity+0x14/0x8c
    Tainted: G      D
ra    : 8002120c pick_next_task_fair+0x90/0xe4
Status: 10004402    KERNEL EXL
Cause : 00800008
BadVA : 0000001c
PrId  : 0101937c (MIPS 24Kc)
Process events/0 (pid: 4, threadinfo=8fb08000, task=8f81eda0, tls=00000000)
Stack : 00000000 800221a8 80160000 802637b4 00000000 00000000 8021dbc0
8002120c
        8f81eda0 80020a90 8faf4a88 8faf4a80 80130000 8faf4a80 8f81eda0
80004924
        8faf4a80 800959b0 8fb09f80 8fb09f80 8f81eef8 8faf4a88 8f81eef8
800436a0
        8faf4a88 8faf4a80 8fb09f80 00000000 00000000 00000000 00000000
00000000
        00000000 8003e4f8 8f81eef8 8015e000 8f81eef8 80217540 00000000
8f81eda0
        ...
Call Trace:
[<800210e4>] set_next_entity+0x14/0x8c
[<8002120c>] pick_next_task_fair+0x90/0xe4
[<80004924>] schedule+0x5e4/0x72c
[<8003e4f8>] worker_thread+0xc4/0xcc
[<80042d14>] kthread+0x58/0xa4
[<800095ec>] kernel_thread_helper+0x10/0x18

Code: afb10014  afbf001c  afb00010 <8ca2001c> 00a08821  10400008  00809021
8c82
0024  24b00008
note: events/0[4] exited with preempt_count 2
Any suggestions on debugging this?

Thanks and Regards,
Joe

On Wed, Jul 8, 2009 at 4:07 PM, Ralf Baechle <ralf@linux-mips.org> wrote:

>  On Wed, Jul 08, 2009 at 01:37:42PM +0530, joe seb wrote:
>
> > We are trying to port linux 2.6.29.4 version of the kernel from
> > linux-mips.org site to our MIPS 24K based platform and we see issues
> when we
> > use the cache in write-back mode. Cache with write-through configuration
> > works fine.
> > We use:
> > Linux kernel - 2.6.29.4
> > GNU cross tools - 4.3.2
> > Busybox - 1.14.1
> > U-boot - 2009.03
> >
> > Our platform has 256MB of RAM and its mapped to second 256 MB of the
> KSEG0
> > (0x90000000 - 0x9FFFFFFF) and KSEG1 (0xB0000000 - 0xBFFFFFFF), and we
> > specify that "mem=16M@256M" as boot parameter (we just want to use the
> first
> > 16MB by the kernel). The cache initialization for the KSEG0 is done in
> > u-boot.
> >
> > The error we get when cache is configured as write-back is given below:
> >
> > --------------------
> > Linux version 2.6.29.4 (gcc version 4.3.2 (Sourcery G
> > ++ Lite 4.3-51) ) #11 PREEMPT Tue Jul 7 21:16:00 IST 2009
> > CPU revision is: 0101937c (MIPS 24Kc)
> > Determined physical RAM map:
> > User-defined physical RAM map:
> >  memory: 01000000 @ 10000000 (usable)
> > Wasting 2097152 bytes for tracking 65536 unused pages
> > Initrd not found or empty - disabling initrd
> > Zone PFN ranges:
> >   Normal   0x00010000 -> 0x00011000
> > Movable zone start PFN for each node
> > early_node_map[1] active PFN ranges
> >     0: 0x00010000 -> 0x00011000
> > Built 1 zonelists in Zone order, mobility grouping off.  Total pages:
> 4064
> > Kernel command line: mem=16M@256M console=ttyS0,9600 cca=3
> > Primary instruction cache 32kB, VIPT, 4-way, linesize 32 bytes.
> > Primary data cache 32kB, 4-way, PIPT, no aliases, linesize 32 bytes
> > Using cache attribute 3
> > Writing ErrCtl register=00000000
> > Readback ErrCtl register=00000000
> > PID hash table entries: 64 (order: 6, 256 bytes)
> > CPU frequency 50.00 MHz
> > console [ttyS0] enabled
> > Dentry cache hash table entries: 2048 (order: 1, 8192 bytes)
> > Inode-cache hash table entries: 1024 (order: 0, 4096 bytes)
> > Memory: 13776k/16384k available (1210k kernel code, 2608k reserved, 234k
> > data, 6
> > 76k init, 0k highmem)
> > Calibrating delay loop... 33.17 BogoMIPS (lpj=165888)
> > Mount-cache hash table entries: 512
> > VFS: Disk quotas dquot_6.5.2
> > Dquot-cache hash table entries: 1024 (order 0, 4096 bytes)
> > msgmni has been set to 26
> > Serial: 8250/16550 driver, 2 ports, IRQ sharing enabled
> > serial8250: ttyS0 at MMIO 0xa1a30000 (irq = 188) is a 16550A
> > serial8250: ttyS1 at MMIO 0xa1a40000 (irq = 189) is a 16550A
> > Freeing unused kernel memory: 676k freed
> > Algorithmics/MIPS FPU Emulator v1.5
> > CPU 0 Unable to handle kernel paging request at virtual address cccccccc,
> > epc ==
> >  cccccccc, ra == cccccccc
> > Oops[#1]:
> > Cpu 0
> > $ 0   : 00000000 0053934a 00000000 00000001
> > $ 4   : 00000001 00000000 00000000 0000009a
> > $ 8   : 00000010 900038c0 00532730 00532730
> > $12   : 00000018 90284880 00020000 00000034
> > $16   : cccccccc 9016c320 9016aae0 90213520
> > $20   : 00000000 9016a960 00000001 90fa9680
> > $24   : 00000000 9009a414
> > $28   : 90176000 90177d58 9016aae0 cccccccc
> > Hi    : 0000000d
> > Lo    : 0000004a
> > epc   : cccccccc 0xcccccccc
> >     Not tainted
> > ra    : cccccccc 0xcccccccc
> > Status: 10004403    KERNEL EXL IE
> > Cause : 10800008
> > BadVA : cccccccc
> > PrId  : 0101937c (MIPS 24Kc)
> > Process rcS (pid: 11, threadinfo=90176000, task=90200500, tls=0053f470)
> > Stack : cccccccc cccccccc 90200500 9016aae0 90213520 00000000 90213520
> > 00000000
> >         9016aae0 90213520 00000000 90025c28 90200500 90faf9c0 90200500
> > 00000107
> >         9016aae0 00000107 9016aae0 900a04c8 00000003 902a4a40 902a4a48
> > 00000000
> >         902a4a44 900b3ad8 90fb1880 90fa9680 00000000 00000003 90fb1880
> > 90fa9680
> >         90177f30 90da7b20 cccccccc cccccccc cccccccc cccccccc cccccccc
> > cccccccc
> >         ...
> > Call Trace:
> > [<90025c28>] mmput+0x9c/0x194
> > [<900a04c8>] flush_old_exec+0x47c/0x988
> > [<900b3ad8>] alloc_fd+0x9c/0x1a4
> > [<90086c88>] handle_mm_fault+0x9a8/0x107c
> > [<9002f7c4>] do_softirq+0xc8/0xd0
> > [<900cc60c>] load_elf_binary+0x0/0x1410
> > [<9009fd9c>] search_binary_handler+0xa0/0x2bc
> > [<900a138c>] do_execve+0x298/0x300
> > [<900a4c60>] getname+0x28/0xc8
> > [<9000c714>] sys_execve+0x4c/0x78
> > [<90002398>] stack_done+0x20/0x3c
> >
> > Code: (Bad address in epc)
> > do_cpu invoked from kernel context![#2]:
> > Cpu 0
> > $ 0   : 00000000 90210000 9016a98c 00000001
> > $ 4   : 00000002 00000003 90168468 00000000
> > $ 8   : 000007c4 00000004 9016846c 00000001
> > $12   : ffffff80 00000000 90136f7c 00000010
> > $16   : 00000000 00000000 90200500 90213520
> > $20   : 9016a994 90177ca8 00000000 90fa9680
> > $24   : 00000000 90121648
> > $28   : 90176000 90177b48 9016aae0 90fa9680
> > Hi    : 0098963b
> > Lo    : 38c9b600
> > epc   : 90fa9680 0x90fa9680
> >     Tainted: G      D
> > ra    : 90fa9680 0x90fa9680
> > Status: 10004403    KERNEL EXL IE
> > Cause : 1080002c
> > PrId  : 0101937c (MIPS 24Kc)
> > Process rcS (pid: 11, threadinfo=90176000, task=90200500, tls=0053f470)
> > Stack : 9016aae0 900041c4 00000000 00000000 90177ca8 90177ca8 0000000b
> > 90200500
> >         90200500 cccccccc 00000000 9002cc78 90200658 9016a960 9020065c
> > 90213520
> >         90152d44 90177ca8 00000001 cccccccc 00000000 90177ca8 90152d44
> > 90177ca8
> >         90200500 cccccccc 00000000 90177ca8 00000000 90fa9680 9016aae0
> > 9000e0d4
> >         cccccccc 90220000 ffffffff 00000e89 90177bec cccccccc 9016a960
> > 90010f58
> >         ...
> > Call Trace:
> > [<900041c4>] printk+0x24/0x30
> > [<9002cc78>] do_exit+0xf4/0x88c
> > [<9000e0d4>] nmi_exception_handler+0x0/0x34
> > [<90010f58>] do_page_fault+0x2e0/0x350
> > [<90070234>] rmqueue_bulk+0x54/0xd8
> > [<900b0d48>] touch_atime+0xf8/0x174
> > [<9006c7e8>] generic_file_aio_read+0x4d8/0x8d8
> > [<90000404>] ret_from_exception+0x0/0x10
> > [<900038c0>] __bzero+0xc4/0x164
> > [<9009a414>] do_sync_read+0x0/0x168
> > [<90025c28>] mmput+0x9c/0x194
> > [<900a04c8>] flush_old_exec+0x47c/0x988
> > [<900b3ad8>] alloc_fd+0x9c/0x1a4
> > [<90086c88>] handle_mm_fault+0x9a8/0x107c
> > [<9002f7c4>] do_softirq+0xc8/0xd0
> > [<900cc60c>] load_elf_binary+0x0/0x1410
> > [<9009fd9c>] search_binary_handler+0xa0/0x2bc
> > [<900a138c>] do_execve+0x298/0x300
> > [<900a4c60>] getname+0x28/0xc8
> > [<9000c714>] sys_execve+0x4c/0x78
> > [<90002398>] stack_done+0x20/0x3c
> >
> > Code: 040a001a  e5a8b400  4018e618 <464c457f> 00010101  00000000
>  00000000
> > 0008
> > 0002  00000001
> > Fixing recursive fault but reboot is needed!
> > -------------------------------
> >
> > We get crashes at different places and the above crash is one of them.
> > Do you think this failure is due to the wrong cache configuration or
> related
> > to the d-cache aliasing problem?
> >
> > The cache details of our platform:
> > D-cache: 32KB, 4-way, 32B line size, virtually indexed and physically
> > tagged, Config7[AR] bit is set (alias is removed by the hardware).
>
> Aliases disabled in hardware, yet you suspect aliases?  Doesn't make sense.
>
> > I-cache:  32KB, 4-way, 32B line size,virtually indexed and physically
> tagged
> >
> >
> > Is there any similar 24k platform supported in linux kernel which we are
> > refer for the configurations?
>
> There has been a kernel bug for a while which on platforms with a non-zero
> memory start address would effectively disable part of of the cache code.
> Your description above, including the changed behaviour between
> write-though
> and write-back caches is consistent with that bug.  Commit
> 67227819d6dd07f6ec225ea59c67aff3ba936e25 fixes this issue.  For your
> convenience I append it below.
>
> I'd appreciate feedback on your test results with this patch.
>
> (Why do people use non-zero starting addresses for memory?  Handling of
> cache error exceptions is hard enough as it is but with no memory in the
> low 32k the design idea of the cache architecture that stores relative to
> $zero can be used goes down the drain and (not considering
> platform-specific
> solutions here) only be handled by burning the scarce resource of a TLB
> entry for an extremly rare event ...)
>
>  Ralf
>
> From 67227819d6dd07f6ec225ea59c67aff3ba936e25 Mon Sep 17 00:00:00 2001
> From: Ralf Baechle <ralf@linux-mips.org>
> Date: Fri, 3 Jul 2009 07:11:15 +0100
> Subject: [PATCH] MIPS: Fix pfn_valid()
>
> For systems which do not define PHYS_OFFSET as 0 pfn_valid() may falsely
> have returned 0 on most configurations.  Bug introduced by commit
> 752fbeb2e3555c0d236e992f1195fd7ce30e728d (linux-mips.org) rsp.
> 6f284a2ce7b8bc49cb8455b1763357897a899abb (kernel.org) titled "[MIPS]
> FLATMEM: introduce PHYS_OFFSET."
>
> Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
>
> diff --git a/arch/mips/include/asm/page.h b/arch/mips/include/asm/page.h
> index dc0eaa7..96a14a4 100644
> --- a/arch/mips/include/asm/page.h
> +++ b/arch/mips/include/asm/page.h
> @@ -165,7 +165,14 @@ typedef struct { unsigned long pgprot; } pgprot_t;
>
>  #ifdef CONFIG_FLATMEM
>
> -#define pfn_valid(pfn)         ((pfn) >= ARCH_PFN_OFFSET && (pfn) <
> max_mapnr)
> +#define pfn_valid(pfn)                                                 \
> +({                                                                     \
> +       unsigned long __pfn = (pfn);                                    \
> +       /* avoid <linux/bootmem.h> include hell */                      \
> +       extern unsigned long min_low_pfn;                               \
> +                                                                       \
> +       __pfn >= min_low_pfn && __pfn < max_mapnr;                      \
> +})
>
>  #elif defined(CONFIG_SPARSEMEM)
>
>

[-- Attachment #2: Type: text/html, Size: 17801 bytes --]