Re: [PATCH] ceph: optimize ceph_base64_encode() with block processing

[Luis Henriques <luis@igalia.com>]

On Wed, Sep 03 2025, Kuan-Wei Chiu wrote:

> On Wed, Sep 03, 2025 at 08:55:36AM +0100, Luis Henriques wrote:
>> On Tue, Sep 02, 2025 at 09:21:14PM +0000, Viacheslav Dubeyko wrote:
>> > On Wed, 2025-09-03 at 05:05 +0800, Kuan-Wei Chiu wrote:
>> > > On Tue, Sep 02, 2025 at 07:37:22PM +0000, Viacheslav Dubeyko wrote:
>> > > > On Sat, 2025-08-30 at 21:28 +0800, Guan-Chun Wu wrote:
>> > > > > Previously, ceph_base64_encode() used a bitstream approach, handling one
>> > > > > input byte at a time and performing extra bit operations. While correct,
>> > > > > this method was suboptimal.
>> > > > > 
>> > > > 
>> > > > Sounds interesting!
>> > > > 
>> > > > Is ceph_base64_decode() efficient then?
>> > > > Do we have something in crypto library of Linux kernel? Maybe we can use
>> > > > something efficient enough from there?
>> > > > 
>> > > Hi Viacheslav,
>> > > 
>> > > FYI, we already have base64 encode/decode implementations in
>> > > lib/base64.c. As discussed in another thread [1], I think we can put
>> > > the optimized version there and have users switch to call the library
>> > > functions.
>> > > 
>> > > [1]: https://lore.kernel.org/lkml/38753d95-8503-4b72-9590-cb129aa49a41@t-8ch.de/  
>> > > 
>> > > 
>> > 
>> > Sounds great! Generalized version of this algorithm is much better than
>> > supporting some implementation in Ceph code.
>> 
>> Please note that ceph can not use the default base64 implementation because
>> it uses the '_' character in the encoding, as explained in commit
>> 
>>   64e86f632bf1 ("ceph: add base64 endcoding routines for encrypted names")
>> 
>> That's why it implements it's own version according to an IMAP RFC, which
>> uses '+' and ',' instead of '-' and '_'.
>> 
> Perhaps we could modify the API to allow users to provide a custom
> base64 table or an extra parameter to specify which RFC standard to use
> for encoding/decoding?

Yes, sure.  That should work as well.  If I remember correctly, I didn't
bother doing that back then because ceph was the only place that needed a
custom base64.  But I not really sure, that was long ago.

Cheers,
-- 
Luís