[Patch v1] [f2fs] add support mkfs.f2fs fsck.f2fs

[Chris Ruehl]

On Tuesday, November 04, 2014 04:23 PM, walter harms wrote:
>
>
> Am 04.11.2014 02:14, schrieb Chris Ruehl:
>> On Tuesday, November 04, 2014 01:58 AM, Steven Honeyman wrote:
>>> On 3 November 2014 05:25, Chris Ruehl <chris.ruehl at gtsys.com.hk> wrote:
>>>> Hi,
>>>>
>>>> I was in need to have f2fs support in busybox.
>>>> This patchset based on f2fs-tools commit
>>>> 056e4b04fc44a006d5529bafbf87b1d73296c665
>>>>
>>>> [Patch v1.1][f2fs] add build support
>>>> [Patch v1.2][f2fs] add libs header
>>>> [Patch v1.3][f2fs] add libs c-files
>>>>
>>>>
>>>> Please feel free to use it.
>>>>
>>>> Chris
>>>
>>> Hi Chris,
>>>
>>> I can't give an answer for the "official" maintainers, but in case you
>>> noticed my unofficial testing branch and wondered why I haven't added
>>> your contribution:
>>>
>>> Rather than patching makefiles, there's a script that runs and adds
>>> new applets so that everything can be in one file together. There's
>>> some good documentation in docs/new-applet-HOWTO.txt which explains
>>> this.
>>> Another reason is it looks like (apologies if I'm wrong) it's some
>>> already written programs that don't fit with the busybox design, for
>>> example free() is called many times with no surrounding ifdefs; and a
>>> lot of functionality may be repeated from libbb, or even just the
>>> malloc -> xmalloc type of changes.
>>>
>>>
>>> Thanks,
>>> Steven
>>>
>>
>> Steven,
>>
>> Thank you get back to me. I pretty much appreciated your advice. I will get
>> to the HOWTO.
>>
>> In general the 2 functions mkfs and fsck working, so fix the design
>> issues should be a matter of a day or two.
>>
>> Yes, Im using libbb, changed already some function calls from original
>> code.
>> e.g. generate_uuid();  I dig a bit further..
>>
>> I'd had a quick look what the meaning of the x<functions>, and it makes
>> sense to me. About the ifdefs around free's I didn't get it!?
>
>
> bb is about code size.
> The xFUNC are wrapper around common function like malloc that bail out on error,
> so no need the check the return value.
>
> free() is a special case as that allocated mem is freed() on exit(). this does
> not work in every situation so it is common to warp the free in
>   if (ENABLE_FEATURE_CLEAN_UP) free(something);
>
>
> re,
>   wh
>
>>
>> I have some time between my production and work it out.
>> Chris

The x<function> I understand on the spot.
The if ( ) free.. is an eye opener!

Thanks
Chris