[PATCH 00/10] ARM: V7M: Support caches

[Alexandre Torgue]

Hi Vladimir,

On 06/13/2016 06:29 PM, Alexandre Torgue wrote:
> Hi Vladimir,
>
> On 06/13/2016 06:19 PM, Vladimir Murzin wrote:
>> Hi Alex,
>>
>> On 13/06/16 17:09, Alexandre Torgue wrote:
>>> Hi Vladimir,
>>>
>>> On 06/13/2016 05:02 PM, Vladimir Murzin wrote:
>>>> Hi,
>>>>
>>>> This patch set allows M-class cpus benefit of optional cache support.
>>>> It originaly was written by Jonny, I've been keeping it localy mainly
>>>> rebasing over Linux versions.
>>>>
>>>> The main idea behind patches is to reuse existing cache handling code
>>>> from v7A/R. In case v7M cache operations are provided via memory
>>>> mapped interface rather than co-processor instructions, so extra
>>>> macros have been introduced to factor out cache handling logic and
>>>> low-level operations.
>>>>
>>>> Along with the v7M cache support the first user (Cortex-M7) is
>>>> introduced.
>>>>
>>>> Patches were tested on MPS2 platform with Cortex-M3/M4/M7. The later
>>>> one showed significant boot speed-up.
>>>>
>>>> Based on 4.7-rc3.
>>>>
>>>> Thanks!
>>>> Vladimir
>>>>
>>>
>>> Thanks for the series.
>>> Did you change something in those patches compare to patches you sent in
>>> the RFC ?
>>
>> Only 04, 05, 06 have been changed (each has it's own changelog included).
>>
>> I did try to hammer patches because of your report of instability with
>> caches on, but all run fine... so nothing has been changed in regard of
>> your case.
>>
> Ok. Thanks for this input. From now, I will use this series to perform
> my tests. I will continue to debug my instabilities this week.
> I will let you know.
>
I fixed my instabilities and it was linked to my SDRAM timings :(.
I don't see behavior issues with your series (You can add my Tested-by: 
Alexandre TORGUE <alexandre.torgue at st.com>).

I have two others questions for you:

1- Maybe I didn't search deeper in the code but when you enable Dcache 
and Icache, what is the state of cache ? I mean we need to invalidate 
caches before enable it. It is done by kernel (with your patches) or it 
has to be done by a bootloader ?

2- I observe an issue (IMO not linked to your series) when I switch 
off/switch on my board. During the first boot after an hard reset I get 
a bus error  (BFSR part in CFSR register indicates an "A bus fault on an 
instruction prefetch has occurred."). After SRST reset through JTAG, the 
kernel boot correctly.
Do you already seen this kind of issue ?

I add the backtrace and registers info.

__invalid_entry () at arch/arm/kernel/entry-v7m.S:33
33	1:	b	1b
(gdb) bt
#0  __invalid_entry () at arch/arm/kernel/entry-v7m.S:33
#1  0xc000bc2a in unwind_exec_insn (ctrl=<optimized out>) at 
arch/arm/kernel/unwind.c:321
#2  unwind_frame (frame=0xc023d0c0 <reset_devices>) at 
arch/arm/kernel/unwind.c:449
#3  0xc000caec in ?? () at arch/arm/mm/cache-v7.S:69
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
(gdb) f 3
#3  0xc000caec in ?? () at arch/arm/mm/cache-v7.S:69
69	       ret     lr
(gdb) info reg
r0             0x1	1
r1             0x0	0
r2             0x1	1
r3             0x0	0
r4             0x0	0
r5             0x100	256
r6             0x0	0
r7             0xa	10
r8             0x40096	262294
r9             0x0	0
r10            0xc012b004	-1072517116
r11            0x0	0
r12            0x29	41
sp             0xc0230018	0xc0230018 <cpu_cache+4>
lr             0xc000caed	-1073689875
pc             0xc000caec	0xc000caec <v7_flush_icache_all>
xPSR           0x81000005	-2130706427


Thanks for your time.

Alex