2022-09-13 1:14 GMT+09:00, Atte Heikkilä <atteh.mailbox@xxxxxxxxx>:
> On Mon, 12 Sep 2022 19:20:54 +0900, Namjae Jeon wrote:
>>2022-09-12 5:57 GMT+09:00, Atte Heikkilä <atteh.mailbox@xxxxxxxxx>:
>>Hi Atte,
>>
>>[snip]
>>> +static char *casefold_sharename(struct unicode_map *um, const char
>>> *name>)
>>> +{
>>> + char *cf_name;
>>> + int cf_len;
>>> +
>>> + cf_name = kzalloc(KSMBD_REQ_MAX_SHARE_NAME, GFP_KERNEL);
>>> + if (!cf_name)
>>> + return ERR_PTR(-ENOMEM);
>>> +
>>> + if (IS_ENABLED(CONFIG_UNICODE)) {
>>> + const struct qstr q_name = {.name = name, .len = strlen(name)};
>>> +
>>> + if (!um)
>>> + goto out_ascii;
>>Minor nit, Wouldn't it be simpler to change something like the one below?
>>
>>+ if (IS_ENABLED(CONFIG_UNICODE) && um) {
>
> This mailing list already has a v2 patch series. Please, reply to that one.
Okay, but please add cc me when you send the patch to the list.
> As for your suggestion, I thought to keep the statements separate since the
> block with the IS_ENABLED() macro is optimized away when CONFIG_UNICODE is
> not set. I understand that the behavior is the same with your suggestion.
When CONFIG_UNICODE is not set, um is not checked in my suggestion.
Please tell me why my suggestion is worse. if you are okay, I will
update it directly.(i.e. no need to send v3 patch). and please check
the use of strncasecmp() in __caseless_lookup() also.
And I need to do full test for this patches, I think it will take
about two days.
>
> Thank you.
>
>>
>>Thanks!
>>> +
>>> + cf_len = utf8_casefold(um, &q_name, cf_name,
>>> + KSMBD_REQ_MAX_SHARE_NAME);
>>> + if (cf_len < 0)
>>> + goto out_ascii;
>>> +
>>> + return cf_name;
>>> + }
>>> +
>>> +out_ascii:
>>> + cf_len = strscpy(cf_name, name, KSMBD_REQ_MAX_SHARE_NAME);
>>> + if (cf_len < 0)
>>> + return ERR_PTR(-E2BIG);
>>> +
>>> + for (; *cf_name; ++cf_name)
>>> + *cf_name = isascii(*cf_name) ? tolower(*cf_name) : *cf_name;
>>> + return cf_name - cf_len;
>>> +}
>>> +
>
