diff options
author | Stephen Warren <swarren@nvidia.com> | 2012-09-28 10:15:48 -0600 |
---|---|---|
committer | Mark Brown <broonie@opensource.wolfsonmicro.com> | 2012-10-01 13:29:49 +0100 |
commit | 536a53a300d0d40152796eefb0a9e6e36ca37f7d (patch) | |
tree | e644203945ee780d8a1a4cdc656d9b4984f5361d /drivers/mmc | |
parent | f4b81dd83e38277a3d1e6661f0fa8eea8cfe784b (diff) | |
download | linux-3.10-536a53a300d0d40152796eefb0a9e6e36ca37f7d.tar.gz linux-3.10-536a53a300d0d40152796eefb0a9e6e36ca37f7d.tar.bz2 linux-3.10-536a53a300d0d40152796eefb0a9e6e36ca37f7d.zip |
spi: remove completely broken Tegra driver
The current SPI driver has many issues. Examples are:
* Segfaulting on most transfers due to expecting all transfers to have
both RX and TX buffers.
* Hanging on TX transfers since the whole driver flow is driven by RX
DMA completion, but the HW is only told to enable RX for RX transfers.
* Use of clk_disable_unprepare() from atomic context.
* Once those and other minor issues are fixed, the driver still doesn't
actually work.
* The driver also implements a deprecated API to the SPI core.
For this reason, simply remove the driver completely. This has two
advantages:
1) This will remove the last use of Tegra's <mach/dma.h>, which will
allow that file to be removed, which is required for single zImage
work.
2) The downstream driver is significaly different from the current
code. I believe a patch to re-add the downstream driver (with
appropriate cleanup) will be much simpler to review if it's a new
file rather than randomly interspered with essentially unrelated
existing code.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Diffstat (limited to 'drivers/mmc')
0 files changed, 0 insertions, 0 deletions