Re: [coLinux-devel] colinux/os/winnt/user/poll.c

[Nir P. <ni...@em...>]

Hi y'all,

Just a small comment - not really coLinux related, but it is a general 
fault-tolerance issue:

When using function calls that can fail (assuming co_os_malloc(), like 
malloc(), can run out of memory),
one should always handle the failures.
Of course, just returning an error here would not be enough - the higher 
levels must handle it too.

Before you shoot me, yes - I know such code appears in SO MANY places, 
and in so many projects, but that doesn't mean we want more...

Some people tend to say - "So what? if malloc() fails, I don't care, 
because nothing can work after it. Just buy more memory". But 
fault-tolerance is important to software. Especially to OS kernels.

Nir


Ballard, Jonathan wrote:

> Hello,
>
> A few suggestions that help portability:
>
>
> int co_os_poll_chain_wait(co_os_poll_chain_t chain)
> {
> //--    HANDLE wevents[chain->count];
> ++    HANDLE *wevents ;
>     DWORD result;
>     co_os_poll_t poll;
>     int i;
>
> ++    wevents = co_os_malloc( sizeof(HANDLE) * chain->count ) ;
>
>     i=0;
>     co_list_each_entry(poll, &chain->list, node) {
>         wevents[i] = poll->event;
>         i++;
>         }
>
>     result = MsgWaitForMultipleObjects(chain->count, (void * const 
> *)&wevents,
>             FALSE, INFINITE, QS_ALLINPUT);
>
> ++    co_os_free( wevents ) ;
>
>     if (result == WAIT_OBJECT_0 + chain->count)
>         return 0;
> .
>
> .
>
> .
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> coLinux-devel mailing list
> coL...@li...
> https://lists.sourceforge.net/lists/listinfo/colinux-devel