amd756 driver error introduced with kernel-4.9.0-0.rc1.git0.2.fc26

Hello,
It appears that a driver_probe_device failure was introduced with kernel-4.9.0-0.rc1.git0.2.fc26. When booting this kernel, or any newer kernel, I see the following:

[   15.675485] ------------[ cut here ]------------
[   15.680129] WARNING: CPU: 4 PID: 528 at drivers/base/dd.c:343 driver_probe_device+0x353/0x430
[   15.688678] Modules linked in: i2c_amd756(+) pcc_cpufreq(-) k8temp fjes amd_rng acpi_cpufreq tpm_tis tpm_tis_core tpm nfsd auth_rpcgss nfs_acl lockd grace sunrpc xfs libcrc32c hpsa scsi_transport_sas ata_generic pata_acpi serio_raw cciss pata_amd
[   15.688681] CPU: 4 PID: 528 Comm: systemd-udevd Not tainted 4.9.0-0.rc1.git0.2.fc26.x86_64 #1
[   15.688690] Hardware name: HP ProLiant DL385 G1, BIOS A05 03/01/2006
[   15.688693]  ffffb5ddc062bb00 ffffffff953e93fd 0000000000000000 0000000000000000
[   15.688695]  ffffb5ddc062bb40 ffffffff950a181b 00000157bd43a0a0 ffff8f70bd43a0a0
[   15.688697]  ffffffffc0440468 ffffffffc0440468 000000000000000b 0000000000000001
[   15.688698] Call Trace:
[   15.688703]  [<ffffffff953e93fd>] dump_stack+0x63/0x86
[   15.688706]  [<ffffffff950a181b>] __warn+0xcb/0xf0
[   15.688708]  [<ffffffff950a194d>] warn_slowpath_null+0x1d/0x20
[   15.688709]  [<ffffffff955409a3>] driver_probe_device+0x353/0x430
[   15.688711]  [<ffffffff95540b5f>] __driver_attach+0xdf/0xf0
[   15.688713]  [<ffffffff95540a80>] ? driver_probe_device+0x430/0x430
[   15.688715]  [<ffffffff9553e25c>] bus_for_each_dev+0x6c/0xc0
[   15.688717]  [<ffffffff9553ffbe>] driver_attach+0x1e/0x20
[   15.688718]  [<ffffffff9553f9f0>] bus_add_driver+0x170/0x270
[   15.688720]  [<ffffffffc0407000>] ? 0xffffffffc0407000
[   15.688722]  [<ffffffff95541480>] driver_register+0x60/0xe0
[   15.688723]  [<ffffffffc0407000>] ? 0xffffffffc0407000
[   15.688725]  [<ffffffff9543e7dc>] __pci_register_driver+0x4c/0x50
[   15.688728]  [<ffffffffc040701e>] amd756_driver_init+0x1e/0x1000 [i2c_amd756]
[   15.688731]  [<ffffffff95002190>] do_one_initcall+0x50/0x180
[   15.688734]  [<ffffffff9520ab81>] ? __vunmap+0x81/0xd0
[   15.688737]  [<ffffffff9522ad29>] ? kmem_cache_alloc_trace+0x99/0x1b0
[   15.688739]  [<ffffffff951bb9aa>] ? do_init_module+0x27/0x1ef
[   15.688741]  [<ffffffff951bb9e2>] do_init_module+0x5f/0x1ef
[   15.688744]  [<ffffffff9512f581>] load_module+0x25b1/0x2980
[   15.688746]  [<ffffffff9512be60>] ? __symbol_put+0x60/0x60
[   15.688749]  [<ffffffff9512fac3>] SYSC_init_module+0x173/0x190
[   15.688751]  [<ffffffff9512fbfe>] SyS_init_module+0xe/0x10
[   15.688753]  [<ffffffff95003c17>] do_syscall_64+0x67/0x180
[   15.688757]  [<ffffffff958112ab>] entry_SYSCALL64_slow_path+0x25/0x25
[   15.688758] ---[ end trace 9920abb49d6753e2 ]---

The problem does not appear in kernel-4.9.0-0.rc0.git9.1.fc26, but it does appear in kernel-4.9.0-0.rc1.git0.2.fc26. It reproduces 100% of the time for me. Can you please look into this issue?

Thanks,
Johnny Bieren

Hi Johnny,

Are the failing kernels built with option DEBUG_TEST_DRIVER_REMOVE enabled?

Jean Delvare
SUSE L3 Support

Hey Jean,

No, that option is not enabled.

cat /boot/config-4.9.0-0.rc1.git0.2.fc26.x86_64 | grep DEBUG_TEST # CONFIG_DEBUG_TEST_DRIVER_REMOVE is not set

Thanks,
Johnny

Then I don't know. I can't see how the only recent change to the i2c-amd756 driver (ea734404f3daf1f6b5103171d848c9d4641fd96b) could cause this. The warning comes from:

	WARN_ON(!list_empty(&dev->devres_head));

which was added a decade ago so nothing new here.

I'm out of ideas. The only way to move forward is to bisect it to find out the exact commit which triggered the warning.

Jean Delvare
SUSE L3 Support