dfaure added a comment.
I don't really understand the relation to ksmserver, surely kioslaves are not
part of session management.
I also don't really understand what would happen in the new event loop
(select/poll) of the kioslave while the non-blocking I/O is happening. It's not
like the slave can handle other commands in parallel. Slaves are in a separate
process exactly so that they can block.
Regarding O_DIRECT, I have some doubts. The man page you linked to, says
In summary, O_DIRECT is a potentially powerful tool that should be used
with caution. It is recommended that applications treat use of O_DIRECT as a
performance option which is disabled by default.
"The thing that has always disturbed me about O_DIRECT is that the whole
interface is just stupid, and was probably designed by a deranged monkey on
some serious mind-controlling substances."--Linus"`
Are you sure you want to use that? :-)
More precisely, it says it will break with NFS (in some cases).
And it sounds like a bad buffer size with O_DIRECT will slow things down (too
many direct writes instead of buffering).
Overall, I'd say -- did you check what the `cp` source code does? That's what
we're competing against :-)
TASK DETAIL
https://phabricator.kde.org/T12641
To: meven, dfaure
Cc: apol, dfaure, #frameworks, #dolphin, ognarb, broulik, meven, pberestov,
iasensio, fprice, LeGast00n, cblack, MrPepe, fbampaloukas, alexde, GB_2,
Codezela, feverfew, michaelh, spoorun, navarromorales, firef, ngraham,
andrebarros, bruns, emmanuelp, mikesomov
