Bug#853122: [Regression] Changes to i2c-piix4.c initialisation prevent loading of sp5100_tco watchdog driver on AMD SB800 chipset

[Paul Menzel <paulepanter@users.sourceforge.net>]

[-- Attachment #1: Type: text/plain, Size: 2182 bytes --]

Dear Linux folks,


Unfortunately, commit 2fee61d22e (i2c: piix4: Add support for
multiplexed main adapter in SB800) [1] caused a regression. Tim
reported that to the Linux Kernel Bugtracker as bug #170741 last
September [2], but it looks like the affected subsystems don’t use it.

So I just copy his report in here, and put all the people in CC
mentioned in the commit, the bug report, and in the subsystems
I2C/SMBUS CONTROLLER DRIVERS FOR PC and WATCHDOG DEVICE DRIVERS
mentioned in the file `MAINTAINERS`.

> On the AMD Turion N40L and other related SoCs, the i2c-piix4 driver
> now claims the 0xcd6 ioport, preventing the sp5100_tco watchdog
> driver from loading:
> 
> piix4_smbus 0000:00:14.0: SMBus Host Controller at 0xb00, revision 0
> piix4_smbus 0000:00:14.0: Using register 0x2c for SMBus port selection
> piix4_smbus 0000:00:14.0: Auxiliary SMBus Host Controller at 0xb20
> sp5100_tco: SP5100/SB800 TCO WatchDog Timer Driver v0.05
> sp5100_tco: PCI Vendor ID: 0x1002, Device ID: 0x4385, Revision ID: 0x42
> sp5100_tco: I/O address 0x0cd6 already in use
> 
> 
> This breaks watchdog operation on existing systems on upgrade and new
> deployments unless the i2c-piix4 driver is blacklisted.
> 
> See:
> 
> drivers/watchdog/sp5100_tco.c
> 
> tco_timer_enable(void)
> 
> (SB800_IO_PM_INDEX_REG is defined in drivers/watchdog/sp5100_tco.h)
> 
> This is the commit which prevents the watchdog driver from loading:
> 
> http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2fee61d22e606fc99ade9079fda15fdee83ec33e
> 
> See also AMD docs
> 
> 45483_sb800_bdg_pub_3.03
> 
> Perhaps a fix is to switch both drivers from static to dynamic
> allocation of the IO ports in question, since the watchdog driver
> only accesses the port during initialisation (with backoff/retry
> maybe to avoid races?).

Is there somebody having the resources to implement the dynamic
allocation to solve this regression?


Thanks,

Paul


[1] https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2fee61d22e606fc99ade9079fda15fdee83ec33e
[2] https://bugzilla.kernel.org/show_bug.cgi?id=170741

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 195 bytes --]