| Message ID | b0b17697-f23e-8fa5-3757-604a86f3a095@nvidia.com |
|---|---|
| State | New |
| Headers | show |
| Series | [BUG] net: stmmac: Panic observed in stmmac_napi_poll_rx() | expand |
On Fri, May 14, 2021 at 03:24:58PM +0100, Jon Hunter wrote: > Hello! > > I have been looking into some random crashes that appear to stem from > the stmmac_napi_poll_rx() function. There are two different panics I > have observed which are ... [...] > The bug being triggered in skbuff.h is the following ... > > void *skb_pull(struct sk_buff *skb, unsigned int len); > static inline void *__skb_pull(struct sk_buff *skb, unsigned int len) > { > skb->len -= len; > BUG_ON(skb->len < skb->data_len); > return skb->data += len; > } > > Looking into the above panic triggered in skbuff.h, when this occurs > I have noticed that the value of skb->data_len is unusually large ... > > __skb_pull: len 1500 (14), data_len 4294967274 [...] The big value looks suspiciously similar to (unsigned)-EINVAL. > I then added some traces to stmmac_napi_poll_rx() and > stmmac_rx_buf2_len() to trace the values of various various variables > and when the problem occurs I see ... > > stmmac_napi_poll_rx: stmmac_rx: count 0, len 1518, buf1 66, buf2 1452 > stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 66, plen 1518 > stmmac_napi_poll_rx: stmmac_rx: count 1, len 1518, buf1 66, buf2 1452 > stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 66, plen 1536 > stmmac_napi_poll_rx: stmmac_rx: count 2, len 1602, buf1 66, buf2 1536 > stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 1602, plen 1518 > stmmac_napi_poll_rx: stmmac_rx: count 2, len 1518, buf1 0, buf2 4294967212 > stmmac_napi_poll_rx: stmmac_rx: dma_buf_sz 1536, buf1 0, buf2 4294967212 And this one to (unsigned)-EILSEQ. > Note I added the above BUG_ON to trap unusually large buffers. Let > me know if you have any thoughts. Do above ring any bell? Best Regards Michał Mirosław
On 14/05/2021 22:49, Michał Mirosław wrote: > On Fri, May 14, 2021 at 03:24:58PM +0100, Jon Hunter wrote: >> Hello! >> >> I have been looking into some random crashes that appear to stem from >> the stmmac_napi_poll_rx() function. There are two different panics I >> have observed which are ... > [...] >> The bug being triggered in skbuff.h is the following ... >> >> void *skb_pull(struct sk_buff *skb, unsigned int len); >> static inline void *__skb_pull(struct sk_buff *skb, unsigned int len) >> { >> skb->len -= len; >> BUG_ON(skb->len < skb->data_len); >> return skb->data += len; >> } >> >> Looking into the above panic triggered in skbuff.h, when this occurs >> I have noticed that the value of skb->data_len is unusually large ... >> >> __skb_pull: len 1500 (14), data_len 4294967274 > [...] > > The big value looks suspiciously similar to (unsigned)-EINVAL. Yes it does and at first, I thought it was being set to -EINVAL. However, from tracing the length variables I can see that this is not the case. >> I then added some traces to stmmac_napi_poll_rx() and >> stmmac_rx_buf2_len() to trace the values of various various variables >> and when the problem occurs I see ... >> >> stmmac_napi_poll_rx: stmmac_rx: count 0, len 1518, buf1 66, buf2 1452 >> stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 66, plen 1518 >> stmmac_napi_poll_rx: stmmac_rx: count 1, len 1518, buf1 66, buf2 1452 >> stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 66, plen 1536 >> stmmac_napi_poll_rx: stmmac_rx: count 2, len 1602, buf1 66, buf2 1536 >> stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 1602, plen 1518 >> stmmac_napi_poll_rx: stmmac_rx: count 2, len 1518, buf1 0, buf2 4294967212 >> stmmac_napi_poll_rx: stmmac_rx: dma_buf_sz 1536, buf1 0, buf2 4294967212 > > And this one to (unsigned)-EILSEQ. Yes but this simply comes from 1518-1602 = -84. So it is purely coincidence. Jon -- nvpublic
On 5/17/21 1:39 PM, Jon Hunter wrote: > > On 14/05/2021 22:49, Michał Mirosław wrote: >> On Fri, May 14, 2021 at 03:24:58PM +0100, Jon Hunter wrote: >>> Hello! >>> >>> I have been looking into some random crashes that appear to stem from >>> the stmmac_napi_poll_rx() function. There are two different panics I >>> have observed which are ... >> [...] >>> The bug being triggered in skbuff.h is the following ... >>> >>> void *skb_pull(struct sk_buff *skb, unsigned int len); >>> static inline void *__skb_pull(struct sk_buff *skb, unsigned int len) >>> { >>> skb->len -= len; >>> BUG_ON(skb->len < skb->data_len); >>> return skb->data += len; >>> } >>> >>> Looking into the above panic triggered in skbuff.h, when this occurs >>> I have noticed that the value of skb->data_len is unusually large ... >>> >>> __skb_pull: len 1500 (14), data_len 4294967274 >> [...] >> >> The big value looks suspiciously similar to (unsigned)-EINVAL. > > Yes it does and at first, I thought it was being set to -EINVAL. > However, from tracing the length variables I can see that this is not > the case. > >>> I then added some traces to stmmac_napi_poll_rx() and >>> stmmac_rx_buf2_len() to trace the values of various various variables >>> and when the problem occurs I see ... >>> >>> stmmac_napi_poll_rx: stmmac_rx: count 0, len 1518, buf1 66, buf2 1452 >>> stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 66, plen 1518 >>> stmmac_napi_poll_rx: stmmac_rx: count 1, len 1518, buf1 66, buf2 1452 >>> stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 66, plen 1536 >>> stmmac_napi_poll_rx: stmmac_rx: count 2, len 1602, buf1 66, buf2 1536 >>> stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 1602, plen 1518 >>> stmmac_napi_poll_rx: stmmac_rx: count 2, len 1518, buf1 0, buf2 4294967212 >>> stmmac_napi_poll_rx: stmmac_rx: dma_buf_sz 1536, buf1 0, buf2 4294967212 >> >> And this one to (unsigned)-EILSEQ. > > Yes but this simply comes from 1518-1602 = -84. So it is purely > coincidence. > > Jon > I dug around this a little bit. It looks like the issue occurs when we get (pardon my terminology, I haven't dealt with networking stuff much) a split packet. What happens is we first process the first frame, growing 'len'. buf1_len, I think, hits the "First descriptor, get split header length" case and the length is 66. buf2_len hits the rx_not_ls case and the length is 1536. In total 1602. Then the condition 'likely(status & rx_not_ls)' passes and we goto back to 'read_again', and read the next frame. Here we eventually get to buf2_len again. stmmac_get_rx_frame_len returns 1518 for this frame which sounds reasonable, that's what we normally get for non-split frames. So what we get is 1518 - 1602 which overflows. I can dig around a bit more but it would be nice if someone with a bit more knowledge of the hardware could comment on the above. Thanks, Mikko
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c index 4749bd0af160..26336de80fe8 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c @@ -3728,7 +3728,7 @@ static unsigned int stmmac_rx_buf2_len(struct stmmac_priv *priv, /* Not last descriptor */ if (status & rx_not_ls) - return priv->dma_buf_sz; + return priv->dma_buf_sz - len; plen = stmmac_get_rx_frame_len(priv, p, coe); @@ -3848,7 +3848,7 @@ static int stmmac_rx(struct stmmac_priv *priv, int limit, u32 queue) buf1_len = stmmac_rx_buf1_len(priv, p, status, len); len += buf1_len; - buf2_len = stmmac_rx_buf2_len(priv, p, status, len); + buf2_len = stmmac_rx_buf2_len(priv, p, status, buf1_len); len += buf2_len; The patch I added to trace the various variable is as follows ... diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c index 4749bd0af160..975059bb3165 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c @@ -3723,15 +3723,20 @@ static unsigned int stmmac_rx_buf2_len(struct stmmac_priv *priv, unsigned int plen = 0; /* Not split header, buffer is not available */ - if (!priv->sph) + if (!priv->sph) { + trace_printk("%s: len %u, plen %u\n", __func__, len, priv->dma_buf_sz); return 0; + } /* Not last descriptor */ - if (status & rx_not_ls) + if (status & rx_not_ls) { + trace_printk("%s: len %u, plen %u\n", __func__, len, priv->dma_buf_sz); return priv->dma_buf_sz; + } plen = stmmac_get_rx_frame_len(priv, p, coe); + trace_printk("%s: len %u, plen %u\n", __func__, len, plen); /* Last descriptor */ return plen - len; } @@ -3850,7 +3855,13 @@ static int stmmac_rx(struct stmmac_priv *priv, int limit, u32 queue) len += buf1_len; buf2_len = stmmac_rx_buf2_len(priv, p, status, len); len += buf2_len; - + trace_printk("%s: count %u, len %u, buf1 %u, buf2 %u\n", __func__, + count, len, buf1_len, buf2_len); + if (buf1_len > priv->dma_buf_sz || buf2_len > priv->dma_buf_sz) { + trace_printk("%s: dma_buf_sz %u, buf1 %u, buf2 %u\n", + __func__, priv->dma_buf_sz, buf1_len, buf2_len); + BUG_ON(1); + } /* ACS is set; GMAC core strips PAD/FCS for IEEE 802.3 * Type frames (LLC/LLC-SNAP) *
Hello! I have been looking into some random crashes that appear to stem from the stmmac_napi_poll_rx() function. There are two different panics I have observed which are ... [ 19.591967] ------------[ cut here ]------------ [ 19.596701] kernel BUG at /home/jonathanh/kernel/include/linux/skbuff.h:2297! [ 19.606273] Internal error: Oops - BUG: 0 [#1] PREEMPT SMP [ 19.611847] Modules linked in: [ 19.615146] CPU: 0 PID: 9 Comm: ksoftirqd/0 Not tainted 5.10.0 #1 [ 19.621357] Hardware name: NVIDIA Jetson AGX Xavier Developer Kit (DT) [ 19.628010] pstate: 00c00009 (nzcv daif +PAN +UAO -TCO BTYPE=--) [ 19.634134] pc : eth_type_trans+0xf8/0x1a0 [ 19.638326] lr : stmmac_napi_poll_rx+0x340/0xbd0 [ 19.643123] sp : ffff800011f63ba0 [ 19.646544] x29: ffff800011f63ba0 x28: 00000000000005ea [ 19.651995] x27: ffff0003d4fe48c0 x26: 00000000000000d3 [ 19.657532] x25: ffff0003d4fe48c0 x24: 0000000000000000 [ 19.662961] x23: ffff0000828d1a40 x22: ffff000083497f00 [ 19.668503] x21: ffff000084b4d6c0 x20: ffff800011b79000 [ 19.673934] x19: 0000000000000000 x18: 0000000000000000 [ 19.679472] x17: 0000000000000000 x16: ffff00008095e291 [ 19.684953] x15: 0000000000000001 x14: 4f77dff9f5ddb24a [ 19.690425] x13: 3ecd4adcbed3184e x12: bca593c98536d7f5 [ 19.695860] x11: 62327b117d535683 x10: 69f6607bf62388d6 [ 19.701297] x9 : a65753eedd84defc x8 : ffff000084b4dcd0 [ 19.706836] x7 : ffff000084b4dcc0 x6 : 0000000000000000 [ 19.712271] x5 : 0000000000001000 x4 : 0000008000000000 [ 19.717810] x3 : 00000000ffffffea x2 : 00000000000005dc [ 19.723241] x1 : ffff0003d4fe4000 x0 : ffff000083497f00 [ 19.728777] Call trace: [ 19.731322] eth_type_trans+0xf8/0x1a0 [ 19.735163] stmmac_napi_poll_rx+0x340/0xbd0 [ 19.739540] net_rx_action+0x220/0x368 [ 19.743387] __do_softirq+0x10c/0x264 [ 19.747145] run_ksoftirqd+0x44/0x58 [ 19.750840] smpboot_thread_fn+0x1d0/0x298 [ 19.755025] kthread+0x120/0x150 [ 19.758484] ret_from_fork+0x10/0x1c [ 19.762235] Code: 52800063 33000862 39020002 17ffffe0 (d4210000) [ 19.768472] ---[ end trace 8e68a50b77108dd8 ]--- [ 19.773159] Kernel panic - not syncing: Oops - BUG: Fatal exception in interrupt [ 19.780636] SMP: stopping secondary CPUs [ 19.785053] Kernel Offset: disabled [ 19.788691] CPU features: 0x0240002,2300aa30 [ 19.793044] Memory Limit: none [ 19.796183] ---[ end Kernel panic - not syncing: Oops - BUG: Fatal exception in interrupt ]--- ... or ... [ 14.914607] Unable to handle kernel write to read-only memory at virtual address ffff000087ab3000 [ 14.924064] Mem abort info: [ 14.927201] ESR = 0x9600014f [ 14.930578] EC = 0x25: DABT (current EL), IL = 32 bits [ 14.936484] SET = 0, FnV = 0 [ 14.939890] EA = 0, S1PTW = 0 [ 14.943603] Data abort info: [ 14.946781] ISV = 0, ISS = 0x0000014f [ 14.950995] CM = 1, WnR = 1 [ 14.954334] swapper pgtable: 4k pages, 48-bit VAs, pgdp=00000004505ae000 [ 14.961377] [ffff000087ab3000] pgd=000000047fff8003, p4d=000000047fff8003, pud=000000047fc6b003, pmd=000000047fc2d003, pte=0060000107ab3787 [ 14.974932] Internal error: Oops: 9600014f [#1] PREEMPT SMP [ 14.980660] Modules linked in: tegra210_adma snd_hda_codec_hdmi snd_hda_tegra snd_hda_codec host1x snd_hda_core crct10dif_ce tegra_aconnect pwm_fan lm90 pwm_tegra tegra_bpmp_thermal phy_tegra194_p2u pcie_tegra194 at24 ip_tables x_tables ipv6 [ 15.003003] CPU: 0 PID: 9 Comm: ksoftirqd/0 Not tainted 5.10.0 #1 [ 15.009381] Hardware name: NVIDIA Jetson AGX Xavier Developer Kit (DT) [ 15.016058] pstate: 80c00009 (Nzcv daif +PAN +UAO -TCO BTYPE=--) [ 15.022364] pc : __dma_inv_area+0x40/0x58 [ 15.026660] lr : arch_sync_dma_for_cpu+0x1c/0x28 [ 15.031368] sp : ffff800011f63ba0 [ 15.035035] x29: ffff800011f63ba0 x28: 0000000000000000 [ 15.040658] x27: ffff0003d4fb88c0 x26: 00000000000000be [ 15.046253] x25: ffff0003d4fb88c0 x24: 00000000000005ea [ 15.051960] x23: ffff0000828bd7a0 x22: 00000000ffffffa8 [ 15.057544] x21: 0000000000000002 x20: ffff000080929010 [ 15.063251] x19: 0000000107a3d000 x18: 0000000000000001 [ 15.068957] x17: 0000000000000000 x16: 0000000000000000 [ 15.074517] x15: 0000000000000068 x14: b17d6301a8460a08 [ 15.080185] x13: 010100003bd8701b x12: 1080e9f16ff0a040 [ 15.086033] x11: 3bb9990301080263 x10: a8c00163a8c0ccd5 [ 15.091856] x9 : 06400040fb17dc05 x8 : ffff000083e108c0 [ 15.097401] x7 : ffff000083e108c0 x6 : 00000000ffffffa8 [ 15.102924] x5 : ffff8000109ca6e0 x4 : 0000000000000000 [ 15.108612] x3 : 000000000000003f x2 : 0000000000000040 [ 15.114302] x1 : ffff000187a3cf80 x0 : ffff000087ab3000 [ 15.119952] Call trace: [ 15.122543] __dma_inv_area+0x40/0x58 [ 15.126521] dma_sync_single_for_cpu+0xac/0x100 [ 15.131195] stmmac_napi_poll_rx+0x490/0xbd0 [ 15.135786] net_rx_action+0x220/0x368 [ 15.139663] __do_softirq+0x10c/0x264 [ 15.143609] run_ksoftirqd+0x44/0x58 [ 15.147452] smpboot_thread_fn+0x1d0/0x298 [ 15.151678] kthread+0x120/0x150 [ 15.155042] ret_from_fork+0x10/0x1c [ 15.158772] Code: 8a230000 54000060 d50b7e20 14000002 (d5087620) [ 15.165228] ---[ end trace 23cff873d17a8509 ]--- This is seen on our Tegra194 Jetson AGX Xavier and it seems that it is more likely to occur when the CPU is operating at slower frequencies. For instance if I set the CPU speed to the min of 115.2MHz it occurs more often but not 100%. Please note that although the above panics logs are from 5.10, it is also seen with the current mainline. The bug being triggered in skbuff.h is the following ... void *skb_pull(struct sk_buff *skb, unsigned int len); static inline void *__skb_pull(struct sk_buff *skb, unsigned int len) { skb->len -= len; BUG_ON(skb->len < skb->data_len); return skb->data += len; } Looking into the above panic triggered in skbuff.h, when this occurs I have noticed that the value of skb->data_len is unusually large ... __skb_pull: len 1500 (14), data_len 4294967274 I then added some traces to stmmac_napi_poll_rx() and stmmac_rx_buf2_len() to trace the values of various various variables and when the problem occurs I see ... stmmac_napi_poll_rx: stmmac_rx: count 0, len 1518, buf1 66, buf2 1452 stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 66, plen 1518 stmmac_napi_poll_rx: stmmac_rx: count 1, len 1518, buf1 66, buf2 1452 stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 66, plen 1536 stmmac_napi_poll_rx: stmmac_rx: count 2, len 1602, buf1 66, buf2 1536 stmmac_napi_poll_rx: stmmac_rx_buf2_len: len 1602, plen 1518 stmmac_napi_poll_rx: stmmac_rx: count 2, len 1518, buf1 0, buf2 4294967212 stmmac_napi_poll_rx: stmmac_rx: dma_buf_sz 1536, buf1 0, buf2 4294967212 The above shows that occasionally the value of 'len' passed to stmmac_rx_buf2_len() is greater than the value returned by stmmac_get_rx_frame_len() and stored in 'plen', which then causes stmmac_rx_buf2_len() to return a negative value but is then treated as an unsigned value. So it is clear to me that there is overflow happening occasionally, but I am not sure how to fix it. I tried the following but this appears to be causing other issues ... Note I added the above BUG_ON to trap unusually large buffers. Let me know if you have any thoughts. Thanks! Jon