Re: [PATCH 2/3] rhashtable-test: retry insert operations in threads

[Phil Sutter <phil@nwl.cc>]

On Tue, Sep 01, 2015 at 09:00:57PM +0800, Herbert Xu wrote:
> On Tue, Sep 01, 2015 at 02:46:48PM +0200, Phil Sutter wrote:
> >
> > This is not an inherent behaviour of the implementation but general
> > agreement. The insertion may fail non-permanently (returning -EBUSY),
> > users are expected to handle this by retrying the operation.
> 
> Absolutely not.  The only reason for an insertion to fail is if we
> can't allocate enough memory.  Unless the user is also looping its
> kmalloc calls it definitely shouldn't be retrying the insert.

rhashtable_insert_fast() returns -EBUSY if the table is full
(rht_grow_above_100() returns true) and an asynchronous rehash operation
is active. AFAICT, this is not necessarily caused by memory pressure.

> If an expansion fails it means either that the system is suffering
> a catastrophic memory shortage, or the user of rhashtable is doing
> something wrong.

Hmm. Since memory allocation is first tried with GFP_ATOMIC set and upon
failure retried in background, this seems like a situation which might
happen during normal use. If that already indicates a severe problem,
why retry in background at all?

Cheers, Phil