Messages in this thread | | | Date | Wed, 18 Jun 2003 17:33:27 -0700 | From | (John Myers) | Subject | Re: [PATCH 2.5.71-mm1] aio process hang on EINVAL |
| |
Joel Becker wrote:
>But POSIX specifes that all detection that can happen before queuing should. > The kernel would have to be substantially more complex to report all errors that could possibly be detected during queuing. The kernel could even detect success during queuing if it really tried.
This is not a reasonable requirement. A correctly written program has to be able to handle errors reported asynchronously. Why else would they be using an asynchronous API?
> You also now require a poll round and aditional system call to >see the errors. In addition, you waste resources until you pick up the >errorneous call. > So? That is a miniscule amount of resources used by an extremely rare condition. Such a picayune optimization hardly justifies the necessary increase in complexity.
[unhandled content-type:application/x-pkcs7-signature] | |