Re: [PATCH,HURD] hurdselect: Step1, code split preparations

[Svante Signell]

On Wed, 2013-01-23 at 00:52 +0100, Samuel Thibault wrote:
> Svante Signell, le Tue 22 Jan 2013 20:53:06 +0100, a écrit :
> > Updated patch attached (not reindented). Shall I continue sending
> > patches or not?
> 
> Well, this one doesn't actually do much, so we need to see the others to
> be able to say anything.

Since there was a lot of concerns about code duplication I introduced a
helper function _wait_for_replies(), splitting out the union definitions
and the _mach_msg() call.

The patch does the following:
1) Introduce the ispoll enum
2) Create the three cases: DELAY, POLL, SELECT
3) Create a helper function: _wait_for_replies() (good enough name?)
4) Prepare for simplifying the code by introducing the switch (ispoll)
cases.

The patch is made against the latest version in 2.13-38 with the three
added patches (as previous patch). A few remaining questions remains:

1) The struct dfd definition is outside the functions
_wait_for_replies() and _hurd_select(). Can this create problems in a
threaded application? If so how to modify the code in the best way?

2) The functionality of ppoll and pselect has not been checked, OK?

hurdselect_step1_2.patch
Description: Text Data