On Tue, Mar 22, 2016 at 10:30 AM, Yan, Zheng <zyan@xxxxxxxxxx> wrote:
>
>> On Mar 22, 2016, at 17:17, Ilya Dryomov <idryomov@xxxxxxxxx> wrote:
>>
>> On Tue, Mar 22, 2016 at 8:52 AM, Yan, Zheng <zyan@xxxxxxxxxx> wrote:
>>>
>>>> On Mar 22, 2016, at 14:11, Ilya Dryomov <idryomov@xxxxxxxxx> wrote:
[ snip ]
>>>> This ties into my question about how namespaces are going to be used
>>>> and how long the namespace name is allowed to be.
>>>>
>>>> CEPH_MAX_NAMESPACE_LEN is defined to 100 above, but that definition is
>>>> removed in patch 5. That needs fixing, and if the 100 char limit is
>>>> real, then buf can just be
>>>>
>>>> CEPH_MAX_OID_NAME_LEN + CEPH_MAX_NAMESPACE_LEN + 1
>>>>
>>>> with no need for a kmalloc().
>>>
>>> CEPH_MAX_NAMESPACE_LEN is a intermediate variable for splitting patches (make individual patch be able to compile). As I know there is no limitation on namespace length.
>>
>> To me, that's indication of a poorly structured series.
>>
>> I understand that it's just a std::string in userspace and so there
>> isn't a limit as such. Same goes for OIDs, but we do limit those in
>> the kernel client. Can we do the same for namespace names?
>
> We can. But it’s irrelevance for this series, I can squash this patch and patch 5.
On the contrary, it's very relevant. It answers whether embedding
namespace names by value into ceph_osd_request is feasible.
Thanks,
Ilya
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
