Re: [2.6.35-rc1] page alloc failure order:1, mode:0x4020

[Eric Dumazet <eric.dumazet@gmail.com>]

Le mercredi 25 août 2010 à 13:25 -0700, Nikhil Sethi (निखिल सेठी) a
écrit :
> Hello.
> 
> I am running 2.5.35.3 with the above patch, and I still get these
> failures. Though they are much less often than without the patch. A
> snippet from dmesg below.
> 
> Please let me know what other details I should provide. Thanks
> 
> skbuff alloc of size 3872 failed
> java: page allocation failure. order:1, mode:0x4020
> Pid: 11464, comm: java Not tainted 2.6.35.3 #3
> Call Trace:
>  [<c0243d26>] ? __alloc_pages_nodemask+0x3e6/0x513
>  [<c025c293>] ? __slab_alloc+0x2d7/0x2eb
>  [<c025c946>] ? __kmalloc_track_caller+0x74/0x95
>  [<d09e801a>] ? ath_rxbuf_alloc+0x1a/0x78 [ath]
>  [<d09e801a>] ? ath_rxbuf_alloc+0x1a/0x78 [ath]
>  [<c0334097>] ? __alloc_skb+0x57/0x100
>  [<d09e801a>] ? ath_rxbuf_alloc+0x1a/0x78 [ath]
>  [<d0af4100>] ? ath_rx_tasklet+0x2fb/0x808 [ath9k]
>  [<d0cbc89f>] ? br_handle_frame+0x1b3/0x1c3 [bridge]
>  [<c033d18a>] ? __netif_receive_skb+0x141/0x25f
>  [<d0af23c1>] ? ath9k_tasklet+0xcc/0x107 [ath9k]
>  [<c02195bf>] ? tasklet_action+0x5f/0x65
>  [<c0219873>] ? __do_softirq+0x60/0xc6
>  [<c0219907>] ? do_softirq+0x2e/0x30
>  [<c02199f9>] ? irq_exit+0x53/0x55
>  [<c020392c>] ? do_IRQ+0x3a/0x72
>  [<c0202be9>] ? common_interrupt+0x29/0x30
> Mem-Info:
> DMA per-cpu:
> CPU    0: hi:    0, btch:   1 usd:   0
> Normal per-cpu:
> CPU    0: hi:   90, btch:  15 usd:  32
> active_anon:6158 inactive_anon:15748 isolated_anon:0
>  active_file:11477 inactive_file:22926 isolated_file:0
>  unevictable:468 dirty:5788 writeback:0 unstable:0
>  free:990 slab_reclaimable:3091 slab_unreclaimable:2349
>  mapped:1410 shmem:5 pagetables:227 bounce:0
> DMA free:1000kB min:124kB low:152kB high:184kB active_anon:1304kB
> inactive_anon:2304kB active_file:2584kB inactive_file:4900kB
> unevictable:0kB isolated(anon):0kB isolated(file):0kB present:15864kB
> mlocked:0kB dirty:0kB writeback:0kB mapped:284kB shmem:0kB
> slab_reclaimable:648kB slab_unreclaimable:752kB kernel_stack:152kB
> pagetables:0kB unstable:0kB bounce:0kB writeback_tmp:0kB
> pages_scanned:0 all_unreclaimable? no
> lowmem_reserve[]: 0 238 238
> Normal free:2960kB min:1908kB low:2384kB high:2860kB
> active_anon:23328kB inactive_anon:60688kB active_file:43324kB
> inactive_file:86804kB unevictable:1872kB isolated(anon):0kB
> isolated(file):0kB present:243840kB mlocked:1872kB dirty:23152kB
> writeback:0kB mapped:5356kB shmem:20kB slab_reclaimable:11716kB
> slab_unreclaimable:8644kB kernel_stack:888kB pagetables:908kB
> unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:0
> all_unreclaimable? no
> lowmem_reserve[]: 0 0 0
> DMA: 76*4kB 29*8kB 5*16kB 0*32kB 0*64kB 1*128kB 1*256kB 0*512kB
> 0*1024kB 0*2048kB 0*4096kB = 1000kB
> Normal: 650*4kB 9*8kB 2*16kB 8*32kB 0*64kB 0*128kB 0*256kB 0*512kB
> 0*1024kB 0*2048kB 0*4096kB = 2960kB
> 37964 total pagecache pages
> 3160 pages in swap cache
> Swap cache stats: add 164954, delete 161794, find 97931/115609
> Free swap  = 937284kB
> Total swap = 963896kB
> 65535 pages RAM
> 1237 pages reserved
> 29042 pages shared
> 38802 pages non-shared
> SLUB: Unable to allocate memory on node -1 (gfp=0x20)
>   cache: kmalloc-8192, object size: 8192, buffer size: 8192, default
> order: 3, min order: 1
>   node 0: slabs: 0, objs: 0, free: 0
> skbuff alloc of size 3872 failed
> java: page allocation failure. order:1, mode:0x4020
> Pid: 11464, comm: java Not tainted 2.6.35.3 #3
> Call Trace:
>  [<c0243d26>] ? __alloc_pages_nodemask+0x3e6/0x513
>  [<c025c293>] ? __slab_alloc+0x2d7/0x2eb
>  [<c025c946>] ? __kmalloc_track_caller+0x74/0x95
>  [<d09e801a>] ? ath_rxbuf_alloc+0x1a/0x78 [ath]
>  [<d09e801a>] ? ath_rxbuf_alloc+0x1a/0x78 [ath]
>  [<c0334097>] ? __alloc_skb+0x57/0x100
>  [<d09e801a>] ? ath_rxbuf_alloc+0x1a/0x78 [ath]
>  [<d0af4100>] ? ath_rx_tasklet+0x2fb/0x808 [ath9k]
>  [<d0cbc89f>] ? br_handle_frame+0x1b3/0x1c3 [bridge]
>  [<c033d18a>] ? __netif_receive_skb+0x141/0x25f
>  [<d0af23c1>] ? ath9k_tasklet+0xcc/0x107 [ath9k]
>  [<c02195bf>] ? tasklet_action+0x5f/0x65
>  [<c0219873>] ? __do_softirq+0x60/0xc6
>  [<c0219907>] ? do_softirq+0x2e/0x30
>  [<c02199f9>] ? irq_exit+0x53/0x55
>  [<c020392c>] ? do_IRQ+0x3a/0x72
>  [<c0202be9>] ? common_interrupt+0x29/0x30
> Mem-Info:
> DMA per-cpu:
> CPU    0: hi:    0, btch:   1 usd:   0
> Normal per-cpu:
> CPU    0: hi:   90, btch:  15 usd:  32
> active_anon:6158 inactive_anon:15748 isolated_anon:0
>  active_file:11477 inactive_file:22926 isolated_file:0
>  unevictable:468 dirty:5788 writeback:0 unstable:0
>  free:990 slab_reclaimable:3091 slab_unreclaimable:2349
>  mapped:1410 shmem:5 pagetables:227 bounce:0
> DMA free:1000kB min:124kB low:152kB high:184kB active_anon:1304kB
> inactive_anon:2304kB active_file:2584kB inactive_file:4900kB
> unevictable:0kB isolated(anon):0kB isolated(file):0kB present:15864kB
> mlocked:0kB dirty:0kB writeback:0kB mapped:284kB shmem:0kB
> slab_reclaimable:648kB slab_unreclaimable:752kB kernel_stack:152kB
> pagetables:0kB unstable:0kB bounce:0kB writeback_tmp:0kB
> pages_scanned:0 all_unreclaimable? no
> lowmem_reserve[]: 0 238 238
> Normal free:2960kB min:1908kB low:2384kB high:2860kB
> active_anon:23328kB inactive_anon:60688kB active_file:43324kB
> inactive_file:86804kB unevictable:1872kB isolated(anon):0kB
> isolated(file):0kB present:243840kB mlocked:1872kB dirty:23152kB
> writeback:0kB mapped:5356kB shmem:20kB slab_reclaimable:11716kB
> slab_unreclaimable:8644kB kernel_stack:888kB pagetables:908kB
> unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:0
> all_unreclaimable? no
> lowmem_reserve[]: 0 0 0
> DMA: 76*4kB 29*8kB 5*16kB 0*32kB 0*64kB 1*128kB 1*256kB 0*512kB
> 0*1024kB 0*2048kB 0*4096kB = 1000kB
> Normal: 650*4kB 9*8kB 2*16kB 8*32kB 0*64kB 0*128kB 0*256kB 0*512kB
> 0*1024kB 0*2048kB 0*4096kB = 2960kB
> 37964 total pagecache pages
> 3160 pages in swap cache
> 
> On Sun, Jun 6, 2010 at 3:42 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> >
> > Le dimanche 06 juin 2010 à 11:56 +0200, Michael Guntsche a écrit :
> > > On 2010.06.04 18:16:44 , Michael Guntsche wrote:
> > > > I applied the patch recompiled and run it on the routerboard, trying
> > > > to trigger the bug again.
> > >
> > > Hi Eric,
> > >
> > > Up to now I was not able to reproduce the bug, do you think this patch
> > > can be pushed to mainline or is there a "better"/other  fix for it?
> > >
> > > Kind regards,
> > > Michael
> > >
> > >
> >
> > Thanks Michael for testing.
> >
> > I'll submit ASAP an official patch, sent to all people involved in this
> > driver to get their Ack (or Nack).
> >
> > IEEE80211_MAX_MPDU_LEN being 3840 + somebits is suspect, since it doesnt
> > match 802.11 specs.
> >
> > It should be more close of 2304 + MAC header (32bytes) + FCS (4 bytes) ?
> >
> >
> >
> > --


reading my former patch, I believe I mistakenly used a max_t() instead
of a min_() macro :

diff --git a/drivers/net/wireless/ath/ath9k/recv.c b/drivers/net/wireless/ath/ath9k/recv.c
index ca6065b..0a0dc3a 100644
--- a/drivers/net/wireless/ath/ath9k/recv.c
+++ b/drivers/net/wireless/ath/ath9k/recv.c
@@ -226,10 +226,10 @@ static int ath_rx_edma_init(struct ath_softc *sc, int nbufs)
        u32 size;
 
 
-       common->rx_bufsize = roundup(IEEE80211_MAX_MPDU_LEN +
-                                    ah->caps.rx_status_len,
-                                    min(common->cachelsz, (u16)64));
-
+       size = roundup(IEEE80211_MAX_MPDU_LEN + ah->caps.rx_status_len,
+                      min(common->cachelsz, (u16)64));
+       common->rx_bufsize = max_t(u32, size,
+                                  SKB_MAX_ORDER(NET_SKB_PAD + common->cachelsz, 0));
        ath9k_hw_set_rx_bufsize(ah, common->rx_bufsize -
                                    ah->caps.rx_status_len);
 


Please use instead this one :

diff --git a/drivers/net/wireless/ath/ath9k/recv.c b/drivers/net/wireless/ath/ath9k/recv.c
index ca6065b..0a0dc3a 100644
--- a/drivers/net/wireless/ath/ath9k/recv.c
+++ b/drivers/net/wireless/ath/ath9k/recv.c
@@ -226,10 +226,10 @@ static int ath_rx_edma_init(struct ath_softc *sc, int nbufs)
        u32 size;
 
 
-       common->rx_bufsize = roundup(IEEE80211_MAX_MPDU_LEN +
-                                    ah->caps.rx_status_len,
-                                    min(common->cachelsz, (u16)64));
-
+       size = roundup(IEEE80211_MAX_MPDU_LEN + ah->caps.rx_status_len,
+                      min(common->cachelsz, (u16)64));
+       common->rx_bufsize = min_t(u32, size,
+                                  SKB_MAX_ORDER(NET_SKB_PAD + common->cachelsz, 0));
        ath9k_hw_set_rx_bufsize(ah, common->rx_bufsize -
                                    ah->caps.rx_status_len);