| Message ID | 20161111085707.GC16907@hardcore |
|---|---|
| State | New |
| Headers | show |
On 11/11/2016 02:57 AM, Jan Glauber wrote: > > we can reproduce the problem on our side, but I don't have direct access > to a MIPS system so I'm still just guessing what happens. If you want > you can try the attached patches. > > I'm trying to get rid of the polling around the interrupt altogether. > It works fine on Thunderx, sometimes I run into an interrupt timeout > after a lost-arbitration but recovery/retry is able to clean that up. > > Please also revert 70121f7 as before. The last patch adds some debugging > output to see if we run into timeouts or recovery. > Using your three patches without reverting 70121f7 fails on both 71xx and 78xx boards. Paul's "[PATCH 1/2] i2c: octeon: Fix register access" eliminates the i2c probe hang and both boards boot properly. Tested on top of Linus' v4.9-rc4 tag. Steve -- To unsubscribe from this list: send the line "unsubscribe linux-i2c" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 11/11/2016 12:51 PM, Steven J. Hill wrote: > On 11/11/2016 02:57 AM, Jan Glauber wrote: >> >> we can reproduce the problem on our side, but I don't have direct access >> to a MIPS system so I'm still just guessing what happens. If you want >> you can try the attached patches. >> >> I'm trying to get rid of the polling around the interrupt altogether. >> It works fine on Thunderx, sometimes I run into an interrupt timeout >> after a lost-arbitration but recovery/retry is able to clean that up. >> >> Please also revert 70121f7 as before. The last patch adds some debugging >> output to see if we run into timeouts or recovery. >> > Using your three patches without reverting 70121f7 fails on both 71xx > and 78xx boards. Paul's "[PATCH 1/2] i2c: octeon: Fix register access" > eliminates the i2c probe hang and both boards boot properly. Tested > on top of Linus' v4.9-rc4 tag. > We need a definitive set of patches to apply. If Paul's patches are not sufficient, work with Jan to create and test a set (perhaps including reversions) that will make it work on both OCTEON and Thunder, noting any dependencies on Paul's patches. Then get that posted ASAP so that I2C can work again David. -- To unsubscribe from this list: send the line "unsubscribe linux-i2c" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
From 45eae05b8d793f5652f77ab1d5faa62c30927a10 Mon Sep 17 00:00:00 2001 From: Jan Glauber <jglauber@cavium.com> Date: Fri, 11 Nov 2016 09:40:15 +0100 Subject: [PATCH 3/3] i2c: octeon: thunderx: Debug prints for timeout and recovery Signed-off-by: Jan Glauber <jglauber@cavium.com> --- drivers/i2c/busses/i2c-octeon-core.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/drivers/i2c/busses/i2c-octeon-core.c b/drivers/i2c/busses/i2c-octeon-core.c index 1d8775799..a2a92b6 100644 --- a/drivers/i2c/busses/i2c-octeon-core.c +++ b/drivers/i2c/busses/i2c-octeon-core.c @@ -72,9 +72,10 @@ static int octeon_i2c_wait(struct octeon_i2c *i2c) return 0; } - if (!time_left) + if (!time_left) { + pr_err("%s: timed out\n", __func__); return -ETIMEDOUT; - + } return 0; } @@ -169,8 +170,10 @@ static int octeon_i2c_hlc_wait(struct octeon_i2c *i2c) return 0; } - if (!time_left) + if (!time_left) { + pr_err("%s: timed out\n", __func__); return -ETIMEDOUT; + } return 0; } @@ -280,6 +283,7 @@ static int octeon_i2c_start(struct octeon_i2c *i2c) error: /* START failed, try to recover */ + pr_err("%s: try to recover from status: %d\n", __func__, stat); ret = octeon_i2c_recovery(i2c); return (ret) ? ret : -EAGAIN; } -- 1.9.1