[2.6.6-rc1-bk1][SMBUS/I2C PIIX4] - SMBus hangs during high CPU temperature report - Bug in SMBus driver or buggy hardware (IBM)?

[Shawn Starr]

After managing to get the lm80 sensor chip working on my IBM 300PL 6892-N2U
machine, I've reached problems with the SMBus hanging and not being able to
reset.

Below is some extra debugging I've added to try and force reset, then kill
the bus but it seems to be ignoring the requests. Is there a particular
reason the i2c-piix4 driver initially disables the SMBus aside from any
possible EEPROM corruption (not observed on this board, different Atmel
chip).

It would be great if someone from IBM could tell me if the SMBus on the
board is buggy. If I do a warm reboot, the bus returns to normal function.
I'm confused as why its not resetting or killing transactions on the bus
(perhaps my code is wrong?)


-- CODE --

        /* Make sure the SMBus host is ready to start transmitting */
        if ((temp = inb_p(SMBHSTSTS)) != 0x00) {
                dev_dbg(&piix4_adapter.dev, "SMBus busy (%02x), "
                        "waiting... \n", temp);
                do {
                        i2c_delay(1);
                        temp = inb_p(SMBHSTSTS);
                        dev_dbg(&piix4_adapter.dev, "SMBus busy: Wait State:
%x\n", temp);
                        if (temp & 0x04) goto reset;

                } while ((temp & 0x01) && (timeout++ < MAX_TIMEOUT));

                /* If the SMBus is still busy, try to reset */
                if (timeout >= MAX_TIMEOUT) {
reset:
                        dev_err(&piix4_adapter.dev, "SMBus Timeout! "
                        "Trying to reset...\n");
                        /* outb_p(temp, SMBHSTSTS); */
                        i2c_delay(1);
                        outb_p(inb(SMBHSTSTS) | ~0x01e, SMBHSTSTS);
                        i2c_delay(1);
                        temp = inb_p(SMBHSTSTS);

                        outb_p(inb(SMBHSTCNT) | 0x02, SMBHSTCNT);
                        i2c_delay(1);
                        temp = inb_p(SMBHSTCNT);

                        dev_dbg(&piix4_adapter.dev, "SMBHSTCNT Status\n");
                        dev_dbg(&piix4_adapter.dev, "Status Bits:
KILLED=%x\n", (temp & 0x02 ? 1 : 0));
                        dev_dbg(&piix4_adapter.dev, "Status Bits:
INTERRUPT=%x\n", (temp & 0x01 ? 1 : 0));


Kernel log:

piix4-smbus 0000:00:02.3: Found 0000:00:02.3 device
piix4-smbus 0000:00:02.3: Using Interrupt SMI# for SMBus.
piix4-smbus 0000:00:02.3: SMBREV = 0x0
piix4-smbus 0000:00:02.3: SMBA = 0xFE00
i2c_adapter i2c-17: Registered as minor 17
i2c_adapter i2c-17: registered as adapter #17
i2c-core: driver lm80 registered.
i2c_adapter i2c-17: found normal i2c_range entry for adapter 17, addr 0028
i2c_adapter i2c-17: Transaction (pre): CNT=00, CMD=02, ADD=50, DAT0=01,
DAT1=00
i2c_adapter i2c-17: SMBus busy (01), waiting...
i2c_adapter i2c-17: SMBus busy: Wait State: 1
i2c_adapter i2c-17: SMBus Timeout! Trying to reset...
i2c_adapter i2c-17: SMBus busy: Wait State: 1
i2c_adapter i2c-17: Status Bits: FAILED=0
i2c_adapter i2c-17: Status Bits: COLLISION=0
i2c_adapter i2c-17: Status Bits: ERROR=0
i2c_adapter i2c-17: Status Bits: INTERRUPT=0
i2c_adapter i2c-17: Status Bits: BUSY=1
i2c_adapter i2c-17: piix4-i2c: Attempting to KILL transactions on SMBus!
i2c_adapter i2c-17: SMBHSTCNT Status
i2c_adapter i2c-17: Status Bits: KILLED=1
i2c_adapter i2c-17: Status Bits: INTERRUPT=0

i2c_adapter i2c-17: Status Bits: FAILED=0
i2c_adapter i2c-17: Status Bits: COLLISION=0
i2c_adapter i2c-17: Status Bits: ERROR=0
i2c_adapter i2c-17: Status Bits: INTERRUPT=0
i2c_adapter i2c-17: Status Bits: BUSY=1
i2c_adapter i2c-17: Failed! (01)

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/