On Mon, Jun 15, 2020 at 9:44 AM Hao Luo <haoluo@xxxxxxxxxx> wrote:
>
> Andrii,
>
> Thanks for this patch, it looks very nice! Decoupling kconfig from generic externs is much needed.
>
> On Fri, Jun 12, 2020 at 3:34 PM Andrii Nakryiko <andriin@xxxxxx> wrote:
>>
>> Switch existing Kconfig externs to be just one of few possible kinds of more
>> generic externs. This refactoring is in preparation for ksymbol extern
>> support, added in the follow up patch. There are no functional changes
>> intended.
>>
>> Signed-off-by: Andrii Nakryiko <andriin@xxxxxx>
>> ---
>> tools/lib/bpf/libbpf.c | 332 ++++++++++++++++++++++++-----------------
>> 1 file changed, 199 insertions(+), 133 deletions(-)
>>
[...]
>> @@ -1443,12 +1454,12 @@ static int set_ext_value_tri(struct extern_desc *ext, void *ext_val,
>> else /* value == 'n' */
>> *(enum libbpf_tristate *)ext_val = TRI_NO;
>> break;
>> - case EXT_CHAR:
>> + case KCFG_CHAR:
>> *(char *)ext_val = value;
>> break;
>> - case EXT_UNKNOWN:
>> - case EXT_INT:
>> - case EXT_CHAR_ARR:
>> + case KCFG_UNKNOWN:
>> + case KCFG_INT:
>> + case KCFG_CHAR_ARR:
>> default:
>> pr_warn("extern %s=%c should be bool, tristate, or char\n",
>> ext->name, value);
>
>
> Very minor: pr_warn("kconfig extern ..."); I noticed you have one similar message changed below.
>
yeah, good catch, I'll update
>>
>> @@ -1458,12 +1469,12 @@ static int set_ext_value_tri(struct extern_desc *ext, void *ext_val,
>> return 0;
>> }
>>
for the future, please cut irrelevant parts of the patch, makes it
easier to see where your replies are
[...]
