> > > > On 18/02/2019 08:42, Thierry Reding wrote: > > > On Sat, Feb 16, 2019 at 08:33:07AM -0800, Sowjanya Komatineni wrote: > > >> Tegra186 does not support multi-master mode and also there is no > > >> master fifo control register. > > >> > > >> This patch fixes supported features of Tegra186 and prevents > > >> crashing during boot as master fifo control register are not > > >> present on Tegra186 and prior. > > >> > > >> Signed-off-by: Sowjanya Komatineni <skomatin...@nvidia.com> > > >> --- > > >> drivers/i2c/busses/i2c-tegra.c | 4 ++-- > > >> 1 file changed, 2 insertions(+), 2 deletions(-) > > > > > > We really should've caught this earlier. Jon, let's think about ways > > > to make it easier to catch these things in the future on our test system. > > > > Indeed. This has fixed the boot regression I was seeing over the > > weekend on -next for Tegra186. However, I am bit confused here, > > because when I look at the Tegra186 TRM it states that the I2C > > supports multi-master mode which disagrees with this patch. > > Furthermore, it seems odd that > > Tegra210 and Tegra194 would support multi-master mode but Tegra186 > > does not. So is this really correct? > > > > I would expect at least the missing master FIFO registers to cause a crash > > on Tegra186. The internal architecture specification also says that > > multi-master is supported. > > > > Sowjanya: do you have other documents that suggest we don't support > > multi-master on Tegra186 specifically? Or is this to work around an issue > > specific to Tegra186? > > > > As it is this contradicts documentation, so we need to either fix the > > commit to remove only master FIFO support or we need to fix the docs to > > reflect reality. > > > > Thierry
Thierry/Jonathan, Design wise we support but due to known hw bugs we defeatured multi-master mode. TRM docs also says no multi master support for T210, T186 and multi-master is supported only on T194. T210 also has multi master enabled in driver. We should disable that also. Thanks sowjanya
