Messages in this thread | | | Date | Thu, 21 Mar 2013 00:43:08 +0800 | Subject | Re: [PATCH 01/10] workqueue: add wq_numa_tbl_len and wq_numa_possible_cpumask[] | From | Lai Jiangshan <> |
| |
On Wed, Mar 20, 2013 at 11:48 PM, Tejun Heo <tj@kernel.org> wrote: > On Wed, Mar 20, 2013 at 11:43:30PM +0800, Lai Jiangshan wrote: >> > + for_each_node(node) >> > + BUG_ON(!alloc_cpumask_var_node(&wq_numa_possible_cpumask[node], >> > + GFP_KERNEL, node)); >> > + for_each_possible_cpu(cpu) { >> > + node = cpu_to_node(cpu); >> > + if (WARN_ON(node == NUMA_NO_NODE)) { >> > + pr_err("workqueue: NUMA node mapping not available for cpu%d, disabling NUMA support\n", cpu); >> >> since you used WARN_ON(), not BUG_ON(), I think you need to free >> allocated memory here. > > I don't know. Is it necessary? It shouldn't happen sans bugs in arch > code and we're vomiting warning message with full stack trace. The > system will function but something is seriously screwed. I don't > think it matters whether we free the puny number of bytes there or > not and I don't wanna nest deeper there. :P
I agree with you, but many people use semantic analysis tools to do hunting in the kernel. they may interrupt you again.
> > -- > tejun
| |