cadence: force nonlinear buffers to be cloned

In my test setup, I had a SAMA5D27 device configured with ip forwarding, and
second device with usb ethernet (r8152) sending ICMP packets.  If the packet
was larger than about 220 bytes, the SAMA5 device would "oops" with the
following trace:

kernel BUG at net/core/skbuff.c:1863!
Internal error: Oops - BUG: 0 [#1] ARM
Modules linked in: xt_MASQUERADE ppp_async ppp_generic slhc iptable_nat xt_nat nf_nat nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 can_raw can bridge stp llc ipt_REJECT nf_reject_ipv4 sd_mod cdc_ether usbnet usb_storage r8152 scsi_mod mii o
ption usb_wwan usbserial micrel macb at91_sama5d2_adc phylink gpio_sama5d2_piobu m_can_platform m_can industrialio_triggered_buffer kfifo_buf of_mdio can_dev fixed_phy sdhci_of_at91 sdhci_pltfm libphy sdhci mmc_core ohci_at91 ehci_atmel o
hci_hcd iio_rescale industrialio sch_fq_codel spidev prox2_hal(O)
CPU: 0 PID: 0 Comm: swapper Tainted: G           O      5.9.1-prox2+ #1
Hardware name: Atmel SAMA5
PC is at skb_put+0x3c/0x50
LR is at macb_start_xmit+0x134/0xad0 [macb]
pc : [<c05258cc>]    lr : [<bf0ea5b8>]    psr: 20070113
sp : c0d01a60  ip : c07232c0  fp : c4250000
r10: c0d03cc8  r9 : 00000000  r8 : c0d038c0
r7 : 00000000  r6 : 00000008  r5 : c59b66c0  r4 : 0000002a
r3 : 8f659eff  r2 : c59e9eea  r1 : 00000001  r0 : c59b66c0
Flags: nzCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment none
Control: 10c53c7d  Table: 2640c059  DAC: 00000051
Process swapper (pid: 0, stack limit = 0x75002d81)

<snipped stack>

[<c05258cc>] (skb_put) from [<bf0ea5b8>] (macb_start_xmit+0x134/0xad0 [macb])
[<bf0ea5b8>] (macb_start_xmit [macb]) from [<c053e504>] (dev_hard_start_xmit+0x90/0x11c)
[<c053e504>] (dev_hard_start_xmit) from [<c0571180>] (sch_direct_xmit+0x124/0x260)
[<c0571180>] (sch_direct_xmit) from [<c053eae4>] (__dev_queue_xmit+0x4b0/0x6d0)
[<c053eae4>] (__dev_queue_xmit) from [<c05a5650>] (ip_finish_output2+0x350/0x580)
[<c05a5650>] (ip_finish_output2) from [<c05a7e24>] (ip_output+0xb4/0x13c)
[<c05a7e24>] (ip_output) from [<c05a39d0>] (ip_forward+0x474/0x500)
[<c05a39d0>] (ip_forward) from [<c05a13d8>] (ip_sublist_rcv_finish+0x3c/0x50)
[<c05a13d8>] (ip_sublist_rcv_finish) from [<c05a19b8>] (ip_sublist_rcv+0x11c/0x188)
[<c05a19b8>] (ip_sublist_rcv) from [<c05a2494>] (ip_list_rcv+0xf8/0x124)
[<c05a2494>] (ip_list_rcv) from [<c05403c4>] (__netif_receive_skb_list_core+0x1a0/0x20c)
[<c05403c4>] (__netif_receive_skb_list_core) from [<c05405c4>] (netif_receive_skb_list_internal+0x194/0x230)
[<c05405c4>] (netif_receive_skb_list_internal) from [<c0540684>] (gro_normal_list.part.0+0x14/0x28)
[<c0540684>] (gro_normal_list.part.0) from [<c0541280>] (napi_complete_done+0x16c/0x210)
[<c0541280>] (napi_complete_done) from [<bf14c1c0>] (r8152_poll+0x684/0x708 [r8152])
[<bf14c1c0>] (r8152_poll [r8152]) from [<c0541424>] (net_rx_action+0x100/0x328)
[<c0541424>] (net_rx_action) from [<c01012ec>] (__do_softirq+0xec/0x274)
[<c01012ec>] (__do_softirq) from [<c012d6d4>] (irq_exit+0xcc/0xd0)
[<c012d6d4>] (irq_exit) from [<c0160960>] (__handle_domain_irq+0x58/0xa4)
[<c0160960>] (__handle_domain_irq) from [<c0100b0c>] (__irq_svc+0x6c/0x90)
Exception stack(0xc0d01ef0 to 0xc0d01f38)
1ee0:                                     00000000 0000003d 0c31f383 c0d0fa00
1f00: c0d2eb80 00000000 c0d2e630 4dad8c49 4da967b0 0000003d 0000003d 00000000
1f20: fffffff5 c0d01f40 c04e0f88 c04e0f8c 30070013 ffffffff
[<c0100b0c>] (__irq_svc) from [<c04e0f8c>] (cpuidle_enter_state+0x7c/0x378)
[<c04e0f8c>] (cpuidle_enter_state) from [<c04e12c4>] (cpuidle_enter+0x28/0x38)
[<c04e12c4>] (cpuidle_enter) from [<c014f710>] (do_idle+0x194/0x214)
[<c014f710>] (do_idle) from [<c014fa50>] (cpu_startup_entry+0xc/0x14)
[<c014fa50>] (cpu_startup_entry) from [<c0a00dc8>] (start_kernel+0x46c/0x4a0)
Code: e580c054 8a000002 e1a00002 e8bd8070 (e7f001f2)
---[ end trace 146c8a334115490c ]---

The solution was to force nonlinear buffers to be cloned.  This was previously
reported by Klaus Doth (https://www.spinics.net/lists/netdev/msg556937.html)
but never formally submitted as a patch.

This is the third revision, hopefully the formatting is correct this time!
Suggested-by: Klaus Doth <krnl@doth.eu>
Fixes: 653e92a9 ("net: macb: add support for padding and fcs computation")
Signed-off-by: Mark Deneen <mdeneen@saucontech.com>
Link: https://lore.kernel.org/r/20201030155814.622831-1-mdeneen@saucontech.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

drivers/net/ethernet/cadence/macb_main.c

@@ -1929,7 +1929,8 @@ static inline int macb_clear_csum(struct sk_buff *skb)
 
 static int macb_pad_and_fcs(struct sk_buff **skb, struct net_device *ndev)
 {
-	bool cloned = skb_cloned(*skb) || skb_header_cloned(*skb);
+	bool cloned = skb_cloned(*skb) || skb_header_cloned(*skb) ||
+		      skb_is_nonlinear(*skb);
 	int padlen = ETH_ZLEN - (*skb)->len;
 	int headroom = skb_headroom(*skb);
 	int tailroom = skb_tailroom(*skb);