> Attached is a suggested patch for the i2c-nforce2 driver, applying on
> top of your own cleanup patch. Other drivers (i2c-amd8111 and the not
> yet merged i2c-acpi-ec) would need something similar. That being said,
> I doubt this will fix the hangs I was observing, as they seem to be
> caused by the hardware itself and not the i2c-nforce2 driver.
Hm, here it is:
The ACPI specification says that the length of SMBus block transactions
is stored on 5 bits. This suggests that the maximum value of 32 is encoded
as 0, so we must modify the i2c-nforce2 driver to handle this special case
properly.
Signed-off-by: Jean Delvare <khali at linux-fr.org>
---
drivers/i2c/busses/i2c-nforce2.c | 11 ++++++++---
1 file changed, 8 insertions(+), 3 deletions(-)
--- linux-2.6.17-rc6.orig/drivers/i2c/busses/i2c-nforce2.c 2006-06-15 09:16:16.000000000 +0200
+++ linux-2.6.17-rc6/drivers/i2c/busses/i2c-nforce2.c 2006-06-15 09:27:51.000000000 +0200
@@ -149,7 +149,9 @@
outb_p(command, NVIDIA_SMB_CMD);
if (read_write == I2C_SMBUS_WRITE) {
len = min_t(u8, data->block[0], 32);
- outb_p(len, NVIDIA_SMB_BCNT);
+ /* Block length is stored on only 5 bits
+ internally, 0 means 32 */
+ outb_p(len & 0x1f, NVIDIA_SMB_BCNT);
for (i = 0; i < len; i++)
outb_p(data->block[i + 1], NVIDIA_SMB_DATA+i);
}
@@ -195,8 +197,11 @@
break;
case I2C_SMBUS_BLOCK_DATA:
- len = inb_p(NVIDIA_SMB_BCNT);
- len = min_t(u8, len, 32);
+ /* Block length is stored on only 5 bits internally,
+ 0 means 32 */
+ len = inb_p(NVIDIA_SMB_BCNT) & 0x1f;
+ if (len == 0)
+ len = 32;
for (i = 0; i < len; i++)
data->block[i+1] = inb_p(NVIDIA_SMB_DATA + i);
data->block[0] = len;
--
Jean Delvare
