Messages in this thread | | | Date | Mon, 10 Jun 2013 20:54:43 -0400 | From | Sasha Levin <> | Subject | Re: [PATCH] slab: prevent warnings when allocating with __GFP_NOWARN |
| |
On 06/10/2013 07:40 PM, Christoph Lameter wrote: > On Mon, 10 Jun 2013, Sasha Levin wrote: > >> [ 1691.807621] Call Trace: >> [ 1691.809473] [<ffffffff83ff4041>] dump_stack+0x4e/0x82 >> [ 1691.812783] [<ffffffff8111fe12>] warn_slowpath_common+0x82/0xb0 >> [ 1691.817011] [<ffffffff8111fe55>] warn_slowpath_null+0x15/0x20 >> [ 1691.819936] [<ffffffff81243dcf>] kmalloc_slab+0x2f/0xb0 >> [ 1691.824942] [<ffffffff81278d54>] __kmalloc+0x24/0x4b0 >> [ 1691.827285] [<ffffffff8196ffe3>] ? security_capable+0x13/0x20 >> [ 1691.829405] [<ffffffff812a26b7>] ? pipe_fcntl+0x107/0x210 >> [ 1691.831827] [<ffffffff812a26b7>] pipe_fcntl+0x107/0x210 >> [ 1691.833651] [<ffffffff812b7ea0>] ? fget_raw_light+0x130/0x3f0 >> [ 1691.835343] [<ffffffff812aa5fb>] SyS_fcntl+0x60b/0x6a0 >> [ 1691.837008] [<ffffffff8403ca98>] tracesys+0xe1/0xe6 >> >> The caller specifically sets __GFP_NOWARN presumably to avoid this warning on >> slub but I'm not sure if there's any other reason. > > There must be another reason. Lets fix this.
My, I feel silly now.
I was the one who added __GFP_NOFAIL in the first place in 2ccd4f4d ("pipe: fail cleanly when root tries F_SETPIPE_SZ with big size").
What happens is that root can go ahead and specify any size it wants to be used as buffer size - and the kernel will attempt to comply by allocation that buffer. Which fails if the size is too big.
Either way, even if we do end up doing something different, shouldn't we prevent slab from spewing a warning if __GFP_NOWARN is passed?
Thanks, Sasha
| |