This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: simultaneous socket write/close causes panic?


On Wed, Oct 30, 2002 at 11:49:05AM +0000, Nick Garnett wrote:

> > Someday, in my spare time, I'll build an eCos binary with the
> > new stack and see if they have time to test it.
> 
> I take it from this that they are using the old stack.

Yes.

> Which API are they using -- the FILEIO package or the API that
> comes with the stack?

The API that comes with the stack.  We don't have the FILEIO
package in our configuration.

> The locking and reference counting in the FILEIO package should
> either delay the close() until after the write() completes, or
> cause the write() to get an EBADF error, depending on the order
> in which they happen. The stack's own API is less robust in
> this respect. The new stack only uses the FILEIO API, so it
> should be safer.

Ah.  So the new stack won't just drop in place of the old one.
That makes it a bit more difficult to switch.

> However, this only fixes this particular problem, it does not
> solve races in other concurrent operations.

-- 
Grant Edwards
grante@visi.com

-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]