Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

4.14 / 3+: oops in lan78xx_set_wol #2457

Closed
maxnet opened this issue Mar 21, 2018 · 7 comments
Closed

4.14 / 3+: oops in lan78xx_set_wol #2457

maxnet opened this issue Mar 21, 2018 · 7 comments

Comments

@maxnet
Copy link
Contributor

maxnet commented Mar 21, 2018

OpenELEC and clones try to enable WoL with an udev rule ( https://github.com/OpenELEC/OpenELEC.tv/blob/master/packages/network/ethtool/udev.d/99-wakeup-eth.rules )

This regulary -but not always- results in an Oops on 3+:

[   18.647311] Unable to handle kernel NULL pointer dereference at virtual address 000001b0
[   18.655625] pgd = b1a54000
[   18.658395] [000001b0] *pgd=2efde835, *pte=00000000, *ppte=00000000
[   18.664870] Internal error: Oops: 17 [#1] SMP ARM
[   18.669669] Modules linked in: fuse ipv6 sch_fq_codel brcmfmac brcmutil snd_bcm2835(C) cfg80211 rfkill snd_pcm snd_timer snd joydev uio_pdrv_genirq uio fixed
[   18.684060] CPU: 3 PID: 312 Comm: ethtool Tainted: G         C      4.14.26v7-aufs #1
[   18.684062] Hardware name: BCM2835
[   18.684067] task: b0098f00 task.stack: ae5ee000
[   18.684082] PC is at phy_ethtool_set_wol+0x14/0x3c
[   18.684090] LR is at lan78xx_set_wol+0xd4/0xe4
[   18.684093] pc : [<806c5610>]    lr : [<806cd71c>]    psr: 00070013
[   18.684096] sp : ae5efdd0  ip : ae5efde0  fp : ae5efddc
[   18.684098] r10: 00000000  r9 : 00000040  r8 : 00114809
[   18.684101] r7 : 00000000  r6 : ae5efe28  r5 : aeebe000  r4 : ae461800
[   18.684104] r3 : 00000001  r2 : aee33c00  r1 : ae5efe28  r0 : 00000000
[   18.684110] Flags: nzcv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment user
[   18.684112] Control: 10c5383d  Table: 31a5406a  DAC: 00000055
[   18.684116] Process ethtool (pid: 312, stack limit = 0xae5ee210)
[   18.684122] Stack: (0xae5efdd0 to 0xae5f0000)
[   18.684131] fdc0:                                     ae5efdfc ae5efde0 806cd71c 806c5608
[   18.684137] fde0: 7ebbc400 8095bbe4 7ebbca7c aeebe000 ae5efe84 ae5efe00 807f0ce8 806cd654
[   18.684142] fe00: 7ebbca7c ae5efe90 00114809 00000040 00000006 00000006 7ebbca7c ae5efe90
[   18.684147] fe20: 00000000 00000000 00000006 0000003f 00000020 00000000 00000000 807df494
[   18.684153] fe40: 00000004 80d80540 ae5efe6c 808f4710 ae5efe74 ae5efe60 808f4710 00008946
[   18.684158] fe60: 80d80540 7ebbca7c ae5efe90 00000000 00000000 00000000 ae5efedc ae5efe88
[   18.684167] fe80: 808095b8 807ef3d0 76e8b4d0 b301e540 30687465 00000000 00000000 00000000
[   18.684177] fea0: 7ebbc400 00000000 00000000 00000000 ae5efef4 00008946 fffffdfd 80d80540
[   18.684183] fec0: 7ebbca7c 7ebbca7c b47b48e0 00000000 ae5efefc ae5efee0 807c4dc0 80809080
[   18.684188] fee0: ae41d180 ae41d180 00008946 7ebbca7c ae5eff7c ae5eff00 8029d028 807c4bb0
[   18.684193] ff00: 802a15a4 80d048e0 00000003 ae5eff40 ae5eff3c ae5eff20 8028b3d8 802a9ef0
[   18.684198] ff20: ae5eff64 ae5eff30 802a9ef0 808f278c ae5eff74 ae5eff40 807c49f0 8028b384
[   18.684204] ff40: b4a03110 aead1990 00000000 802a9524 00000006 ae41d180 ae41d180 00008946
[   18.684216] ff60: 7ebbca7c 00000003 ae5ee000 00000000 ae5effa4 ae5eff80 8029d798 8029cf88
[   18.684226] ff80: 00000006 7ebbc400 00000000 00000020 00000036 80108a84 00000000 ae5effa8
[   18.684231] ffa0: 801088e0 8029d760 7ebbc400 00000000 00000003 00008946 7ebbca7c 00000006
[   18.684237] ffc0: 7ebbc400 00000000 00000020 00000036 00000002 7ebbcc30 00000002 7ebbca74
[   18.684242] ffe0: 0004441c 7ebbbe54 0001d2ac 76e8b4dc 60070010 00000003 e594005c e5941060
[   18.684274] [<806c5610>] (phy_ethtool_set_wol) from [<806cd71c>] (lan78xx_set_wol+0xd4/0xe4)
[   18.684289] [<806cd71c>] (lan78xx_set_wol) from [<807f0ce8>] (dev_ethtool+0x1924/0x2358)
[   18.684302] [<807f0ce8>] (dev_ethtool) from [<808095b8>] (dev_ioctl+0x544/0x8a4)
[   18.684315] [<808095b8>] (dev_ioctl) from [<807c4dc0>] (sock_ioctl+0x21c/0x2d8)
[   18.684326] [<807c4dc0>] (sock_ioctl) from [<8029d028>] (do_vfs_ioctl+0xac/0x7d8)
[   18.684337] [<8029d028>] (do_vfs_ioctl) from [<8029d798>] (SyS_ioctl+0x44/0x6c)
[   18.684350] [<8029d798>] (SyS_ioctl) from [<801088e0>] (ret_fast_syscall+0x0/0x28)
[   18.684360] Code: e92dd800 e24cb004 e52de004 e8bd4000 (e59031b0) 
[   18.684482] ---[ end trace 40cd6684fa300f3c ]---

maxnet/berryboot#484
@MilhouseVH
Copy link

Yes, we see this too with LibreELEC, although we just disabled 78xx WOL as a temporary solution (there were other priorities at the time).

@maxnet
Copy link
Contributor Author

maxnet commented Mar 21, 2018

Problem isn't limited to WOL though, but seems to happen in all functions in which net->phydev is used, without checking if it is not null.

E.g. can also reproduce it with:

sudo ifconfig eth0 down
sudo ethtool --show-eee eth0

[   60.285843] Unable to handle kernel NULL pointer dereference at virtual address 000001b0
[   60.294113] pgd = b1a70000
[   60.296857] [000001b0] *pgd=2fe1e835, *pte=00000000, *ppte=00000000
[   60.303261] Internal error: Oops: 17 [#1] SMP ARM
[   60.308041] Modules linked in: fuse rfcomm cmac bnep hci_uart bluetooth ecdh_generic i2c_dev ip_tables x_tables ipv6 brcmfmac brcmutil snd_bcm2835(C) joydev cfg80211 snd_pcm rfkill snd_timer snd uio_pdrv_genirq uio fixed
[   60.327992] CPU: 3 PID: 966 Comm: ethtool Tainted: G         C      4.14.26v7-aufs #1
[   60.335935] Hardware name: BCM2835
[   60.339385] task: b1af2d00 task.stack: b1a08000
[   60.343990] PC is at phy_ethtool_get_eee+0x14/0x130
[   60.348943] LR is at lan78xx_get_eee+0x40/0xd4
[   60.353450] pc : [<806c5900>]    lr : [<806cd2e8>]    psr: 60000013
[   60.359807] sp : b1a09dc0  ip : b1a09dd8  fp : b1a09dd4
[   60.365106] r10: 00000000  r9 : 00000040  r8 : 00114809
[   60.370408] r7 : aee0c000  r6 : 00000000  r5 : b1a09e28  r4 : aee0c000
[   60.377031] r3 : 806cd2a8  r2 : 00000000  r1 : b1a09e28  r0 : 00000000
[   60.383652] Flags: nZCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment user
[   60.390884] Control: 10c5383d  Table: 31a7006a  DAC: 00000055
[   60.396706] Process ethtool (pid: 966, stack limit = 0xb1a08210)
[   60.402793] Stack: (0xb1a09dc0 to 0xb1a0a000)
[   60.407207] 9dc0: aee0c000 b1a09e28 b1a09dfc b1a09dd8 806cd2e8 806c58f8 00000000 b1a09e10
[   60.415500] 9de0: 00000004 7ea41bc4 8095bbe4 7ea41c1c b1a09e84 b1a09e00 807efa4c 806cd2b4
[   60.423792] 9e00: 1cd36835 b1a09e40 00114809 00000040 00000044 00000044 b1a09ea4 b1a09e28
[   60.432083] 9e20: 802591e4 8021e3b8 00000044 00000000 00000000 00000000 00000000 00000000
[   60.440375] 9e40: 00000000 00000000 00000000 00000000 b1a09e74 b1a09e60 808f4710 00008946
[   60.448667] 9e60: 80d80540 7ea41c1c b1a09e90 00000000 00000000 00000000 b1a09edc b1a09e88
[   60.456959] 9e80: 808095b8 807ef3d0 00000000 b4a2b780 30687465 00000000 00000000 00000000
[   60.465251] 9ea0: 7ea41bc4 00000000 00000000 00000000 b1a09ef4 00008946 fffffdfd 80d80540
[   60.473543] 9ec0: 7ea41c1c 7ea41c1c 9c827520 00000000 b1a09efc b1a09ee0 807c4dc0 80809080
[   60.481834] 9ee0: 9cd1ea80 9cd1ea80 00008946 7ea41c1c b1a09f7c b1a09f00 8029d028 807c4bb0
[   60.490126] 9f00: 802a15a4 80d048e0 00000003 b1a09f40 b1a09f3c b1a09f20 8028b3d8 802a9ef0
[   60.498418] 9f20: b1a09f64 b1a09f30 802a9ef0 808f278c b1a09f74 b1a09f40 807c49f0 8028b384
[   60.506710] 9f40: b4a03110 9c865330 00000000 802a9524 00000044 9cd1ea80 9cd1ea80 00008946
[   60.515002] 9f60: 7ea41c1c 00000003 b1a08000 00000000 b1a09fa4 b1a09f80 8029d798 8029cf88
[   60.523294] 9f80: 00000044 7ea41c14 7ea41bc4 0005f430 00000036 80108a84 00000000 b1a09fa8
[   60.531585] 9fa0: 801088e0 8029d760 7ea41c14 7ea41bc4 00000003 00008946 7ea41c1c 00000044
[   60.539877] 9fc0: 7ea41c14 7ea41bc4 0005f430 00000036 7ea41dd0 0002dcc8 00031514 00000000
[   60.548168] 9fe0: 76e048a0 7ea41bbc 00018010 76e048ac 60000010 00000003 00000000 00000000
[   60.556469] [<806c5900>] (phy_ethtool_get_eee) from [<806cd2e8>] (lan78xx_get_eee+0x40/0xd4)
[   60.565030] [<806cd2e8>] (lan78xx_get_eee) from [<807efa4c>] (dev_ethtool+0x688/0x2358)
[   60.573152] [<807efa4c>] (dev_ethtool) from [<808095b8>] (dev_ioctl+0x544/0x8a4)
[   60.580653] [<808095b8>] (dev_ioctl) from [<807c4dc0>] (sock_ioctl+0x21c/0x2d8)
[   60.588067] [<807c4dc0>] (sock_ioctl) from [<8029d028>] (do_vfs_ioctl+0xac/0x7d8)
[   60.595656] [<8029d028>] (do_vfs_ioctl) from [<8029d798>] (SyS_ioctl+0x44/0x6c)
[   60.603071] [<8029d798>] (SyS_ioctl) from [<801088e0>] (ret_fast_syscall+0x0/0x28)
[   60.610749] Code: e92dd830 e24cb004 e52de004 e8bd4000 (e59031b0) 
[   60.616957] ---[ end trace fdd7ee79ef0a2e9f ]---

@macmpi
Copy link

macmpi commented Mar 21, 2018

related? #2442

@penfold42
Copy link

Yep

pelwell pushed a commit that referenced this issue Apr 4, 2018
lan78xx: Don't reset the interface on open
b5b6bb9 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix added a commit to raspberrypi/firmware that referenced this issue Apr 4, 2018
@popcornmix
kernel: net: lan78xx: Allow for VLAN headers in timeout
0dff9ec 
See: raspberrypi/linux#2458

kernel: Revert lan78xx: Simple patch to prevent some crashes
kernel: lan78xx: Connect phy early
kernel: lan78xx: Don't reset the interface on open
See: raspberrypi/linux#2437
See: raspberrypi/linux#2442
See: raspberrypi/linux#2457

firmware: clockman: Don't use OSC for pixel clock
See: https://www.raspberrypi.org/forums/viewtopic.php?f=29&t=24679&start=150#p1297298
popcornmix added a commit to Hexxeh/rpi-firmware that referenced this issue Apr 4, 2018
@popcornmix
kernel: net: lan78xx: Allow for VLAN headers in timeout
0382fa9 
See: raspberrypi/linux#2458

kernel: Revert lan78xx: Simple patch to prevent some crashes
kernel: lan78xx: Connect phy early
kernel: lan78xx: Don't reset the interface on open
See: raspberrypi/linux#2437
See: raspberrypi/linux#2442
See: raspberrypi/linux#2457

firmware: clockman: Don't use OSC for pixel clock
See: https://www.raspberrypi.org/forums/viewtopic.php?f=29&t=24679&start=150#p1297298
@popcornmix
Copy link
Collaborator

Latest rpi-update kernel has a potential fix for this issue. Please test.

@MilhouseVH
Copy link

MilhouseVH commented Apr 5, 2018 •
edited
Loading

With LibreELEC and latest 4.14.32 b5b6bb9, the initial WOL panic and the following eee related panic are both no longer an issue:

sudo ifconfig eth0 down
sudo ethtool --show-eee eth0

Thanks!

@maxnet
Copy link
Contributor Author

maxnet commented Apr 5, 2018

Seems to work

@maxnet maxnet closed this as completed Apr 5, 2018
popcornmix pushed a commit that referenced this issue Apr 8, 2018
@popcornmix
lan78xx: Don't reset the interface on open
4d926e8 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
pelwell pushed a commit that referenced this issue Apr 17, 2018
lan78xx: Don't reset the interface on open
d235661 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
pelwell pushed a commit that referenced this issue Apr 17, 2018
lan78xx: Don't reset the interface on open
ee0f3b1 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue Apr 26, 2018
@popcornmix
lan78xx: Don't reset the interface on open
c445a42 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue May 5, 2018
@popcornmix
lan78xx: Don't reset the interface on open
24544b9 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue May 5, 2018
@popcornmix
lan78xx: Don't reset the interface on open
b4c8ab3 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue May 17, 2018
@popcornmix
lan78xx: Don't reset the interface on open
499a690 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
pelwell pushed a commit that referenced this issue May 20, 2018
lan78xx: Don't reset the interface on open
c891414 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
pelwell pushed a commit that referenced this issue May 20, 2018
lan78xx: Don't reset the interface on open
adfb274 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue May 21, 2018
@popcornmix
lan78xx: Don't reset the interface on open
c44827d 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue May 25, 2018
@popcornmix
lan78xx: Don't reset the interface on open
099c7b8 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue May 29, 2018
@popcornmix
lan78xx: Don't reset the interface on open
ceb25a0 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue Jun 5, 2018
@popcornmix
lan78xx: Don't reset the interface on open
ef10500 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue Jun 11, 2018
@popcornmix
lan78xx: Don't reset the interface on open
123f453 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue Jun 13, 2018
@popcornmix
lan78xx: Don't reset the interface on open
73ba259 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue Jun 18, 2018
@popcornmix
lan78xx: Don't reset the interface on open
0ee21d6 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
popcornmix pushed a commit that referenced this issue Jun 27, 2018
@popcornmix
lan78xx: Don't reset the interface on open
3ad3b8a 
With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: #2437
     #2442
     #2457
APokorny pushed a commit to ubports/ubuntu_kernel_xenial that referenced this issue Oct 11, 2018
@ksacilotto
lan78xx: Don't reset the interface on open
d2936e3 
BugLink: http://bugs.launchpad.net/bugs/1784025

With Alexander Graf's patch ("lan78xx: Connect phy early") applied,
the call to lan78xx_reset within lan78xx_open prevents the phy
interrupt from being generated (even though the link is up).

Avoid this issue by removing the lan78xx_reset call.

See: raspberrypi/linux#2437
     raspberrypi/linux#2442
     raspberrypi/linux#2457

Fixes: 92571a1aae40 ("lan78xx: Connect phy early")
Signed-off-by: Phil Elwell <phil@raspberrypi.org>
(backported from commit 47b998653fea4ef69e3e89574956386f262bccca)
Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>
Acked-by: Stefan Bader <stefan.bader@canonical.com>
Acked-by: Kleber Sacilotto de Souza <kleber.souza@canonical.com>
Signed-off-by: Kleber Sacilotto de Souza <kleber.souza@canonical.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
5 participants
@MilhouseVH @popcornmix @maxnet @penfold42 @macmpi