[CRIU] [PATCH] x86: syscalls -- Rework building procedure

Pavel Emelyanov xemul at parallels.com
Fri Apr 10 03:53:12 PDT 2015


On 04/10/2015 01:51 PM, Cyrill Gorcunov wrote:
> On Fri, Apr 10, 2015 at 01:39:31PM +0300, Pavel Emelyanov wrote:
>> On 04/09/2015 11:45 AM, Cyrill Gorcunov wrote:
>>>  - Get rid of shell script, we can do everything via make itself in parallel mode
>>
>> What's the benefit of in-Make code vs shell script?
> 
> - no need to carry additional file
> - the built can be done in paralles, in shell script we read the file one by one

Build of what will be done in parallel?

>>
>>>  - Collect syscall related data into syscalls subdirectory (we gonna implement
>>>    32 bit mode soon)
>>
>> What else is going to be in this directory?
> 
>  - syscall table for 32 bits
>  - syscall wrappers for 32 bits
> 
>>
>>>  - We can't drop off __NR_ constants because we're using them in parasite code
>>>    (when we inject dumper and for "criu exec" mode)
>>
>> OK
>>
>>
>>> -$(obj)/$(SYS-EXEC-TBL): $(obj)/$(SYS-GEN) $(obj)/$(SYS-DEF)
>>> +	$(Q) echo "/* Autogenerated, don't edit */"										>  $@
>>> +	$(Q) echo "#ifndef __ASM_CR_SYSCALL_CODES_H__"										>> $@
>>> +	$(Q) echo "#define __ASM_CR_SYSCALL_CODES_H__"										>> $@
>>> +	$(Q) cat $< | egrep -v '^#' | sed -e 's/\t\{1,\}/|/g' | awk -F '|' '{print "#define " $$1 " " $$2 }'			>> $@
>>> +	$(Q) echo "#endif /* __ASM_CR_SYSCALL_CODES_H__ */"									>> $@
>>
>> Can we fix these long lines? They don't fit even in my wide screen ;)
> 
> Sure. Actually, drop this one, I'll resend it later.
> .
> 



More information about the CRIU mailing list