diff options
| author | Stephen Warren <swarren@nvidia.com> | 2012-10-01 21:33:20 (GMT) |
|---|---|---|
| committer | Stephen Warren <swarren@nvidia.com> | 2012-11-16 19:22:17 (GMT) |
| commit | 1a6d3da8bc45502940efb19604d4c94c470ef6be (patch) | |
| tree | b0fcdf1c6950019c2eabed1ebd0f8513127418c8 /fs/jffs2 | |
| parent | adc18315889b3e67a4a2a9717fbbfab5cd2d7b57 (diff) | |
| download | linux-fsl-qoriq-1a6d3da8bc45502940efb19604d4c94c470ef6be.tar.xz | |
ARM: tegra: decouple uncompress.h and debug-macro.S
Prior to this change, Tegra's debug-macro.S relied on uncompress.h having
determined which UART to use, and whether it was safe to use the UART
(i.e. is it not in reset, and is clocked). This determination was
communicated from uncompress.h to debug-macro.S using a few bytes of
Tegra's IRAM (an on-SoC RAM). This had the disadvantage that uncompress.h
was a required part of the kernel boot process; booting a non-compressed
kernel would not allow earlyprintk to operate.
This change duplicates the UART selection and validation logic into
debug-macro.S so that the reliance on uncompress.h is removed.
This also helps out with single-zImage work, since there is currently no
support for using any uncompress.h with single-zImage.
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Diffstat (limited to 'fs/jffs2')
0 files changed, 0 insertions, 0 deletions