Messages in this thread |  | | | Date | Sat, 16 May 2009 00:27:45 +0530 | | From | Balbir Singh <> | | Subject | Re: Misleading OOM messages |
| |
* Dave Hansen <dave@linux.vnet.ibm.com> [2009-05-15 11:23:27]:
> On Fri, 2009-05-15 at 13:58 -0400, Christoph Lameter wrote: > > On Thu, 14 May 2009, David Rientjes wrote: > > > > To me it at least adds the fact that more should be made *available* and > > > > not just that you're out of it. So, definitely not perfect, but better > > > > than "out". > > > > > > > > > > I think "no allowable memory" followed by information on what is and is > > > not allowed in that specific context would remove any ambiguity. > > > > Useful information to have. If a NUMA or cgroup restriction caused the > > failure then we should print that out. > > We get a wee bit of info out for the cgroups case at least: > > void mem_cgroup_out_of_memory(struct mem_cgroup *mem, gfp_t gfp_mask) > { > ... > if (oom_kill_process(p, gfp_mask, 0, points, mem, > "Memory cgroup out of memory")) > goto retry; > > That can surely be improved, but it's a decent start. >
Also look at mem_cgroup_print_oom_info().
-- Balbir
|  |