From: Domen Puncer <domen.puncer at ultra.si> i2c-au1550: Fix timeout problem Fix from Jordan Crouse: If the transmit and recieve FIFOS are not empty, forceably flush them rather then waiting for them to drain on their own. This solves at least a problem reported by Clem Taylor: http://www.linux-mips.org/archives/linux-mips/2006-05/msg00240.html (1% of I2C transactions would timeout) Signed-off-by: Domen Puncer <domen.puncer at ultra.si> Signed-off-by: Jean Delvare <khali at linux-fr.org> Signed-off-by: Greg Kroah-Hartman <gregkh at suse.de> --- drivers/i2c/busses/i2c-au1550.c | 14 ++++++++++---- 1 files changed, 10 insertions(+), 4 deletions(-) diff --git a/drivers/i2c/busses/i2c-au1550.c b/drivers/i2c/busses/i2c-au1550.c index d06edce..02a359e 100644 --- a/drivers/i2c/busses/i2c-au1550.c +++ b/drivers/i2c/busses/i2c-au1550.c @@ -118,13 +118,19 @@ do_address(struct i2c_au1550_data *adap, /* Reset the FIFOs, clear events. */ - sp->psc_smbpcr = PSC_SMBPCR_DC; + stat = sp->psc_smbstat; sp->psc_smbevnt = PSC_SMBEVNT_ALLCLR; au_sync(); - do { - stat = sp->psc_smbpcr; + + if (!(stat & PSC_SMBSTAT_TE) || !(stat & PSC_SMBSTAT_RE)) { + sp->psc_smbpcr = PSC_SMBPCR_DC; au_sync(); - } while ((stat & PSC_SMBPCR_DC) != 0); + do { + stat = sp->psc_smbpcr; + au_sync(); + } while ((stat & PSC_SMBPCR_DC) != 0); + udelay(50); + } /* Write out the i2c chip address and specify operation */ -- 1.4.2.1
