Re: [PATCH] pipe: don't block after data has been written

[Zan Lynx <zlynx@acm.org>]

On 11/5/09 9:37 AM, Eric Dumazet wrote:
> Max Kellermann a écrit :
>> On 2009/11/05 17:27, Eric Dumazet<eric.dumazet@gmail.com>  wrote:
>>> Your patch breaks many programs, that dont use poll()/select()
>>>
>>> char result[1000000];
>>> main()
>>> {
>>> 	computethings();
>>> 	write(1, buffer, 1000000);
>>> }
>>
>> Your code does not check the return value of write().  This is a bug.
>>
>
> Welcome to real world.

Yes in the real world there are bugs. The decision is to choose which 
bug you are going to expose. If it was my decision I would make the code 
work as documented, as Max wants to do.

I remember many years ago needing to fix some inetd-called server code 
that got unexpected partial writes on blocking sockets. It was either 
Solaris or HP/UX. So this is nothing new.

In fact I think that Linux will already do short writes if a signal is 
received without restart set for the handler. I found several bugs last 
year in glibc and libstdc++ fwrite and iostreams regarding that.
-- 
Zan Lynx
zlynx@acm.org

"Knowledge is Power.  Power Corrupts.  Study Hard.  Be Evil."