Re: Memory management, palloc - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Memory management, palloc
Date
Msg-id 25596.984064452@sss.pgh.pa.us
Whole thread Raw
In response to Memory management, palloc  (Gavin Sherry <swm@linuxworld.com.au>)
List pgsql-hackers
Gavin Sherry <swm@linuxworld.com.au> writes:
> I've been looking through the memory management system today.

> When a request is made for a memory memory chunk larger than
> ALLOC_CHUNK_LIMIT, AllocSetAlloc() uses malloc() to give the request its
> own block. The result is tested by AllocSetAlloc() to see if the memory
> was allocated.

> Irrespective of this, a chunk can be returned which has not had memory
> allocated to it. There is no testing of the return status of
> palloc() through out the code. 

What's your point?

palloc() does not have the same specification as malloc.  It guarantees
to return allocated memory, or elog trying.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Gavin Sherry
Date:
Subject: Re: Memory management, palloc
Next
From: Tom Lane
Date:
Subject: Re: WAL does not recover gracefully from out-of-disk-space