user/sven/linux.git/arch/m68k/kernel/head.S, branch v5.2.2

m68k: Fix boot regression on machines with RAM at non-zero

2014-07-10T07:58:26Z

My enhancement to store the initial mapping size for later reuse in commit 486df8bc4627bdfc032d11bedcd056cc5343ee62 ("m68k: Increase initial mapping to 8 or 16 MiB if possible") broke booting on machines where RAM doesn't start at address zero. Use pc-relative addressing to fix this. Reported-by: Andreas Schwab Signed-off-by: Geert Uytterhoeven Tested-by: Andreas Schwab

m68k: Increase initial mapping to 8 or 16 MiB if possible

2014-05-28T08:11:54Z

If the size of the first memory chunk is at least 8 or 16 MiB increase the initial mapping to 8 resp. 16 MiB instead of 4 MiB. This makes it possible to 1. Map more memory in the first node without running out of space for the page tables, 2. Boot kernels that don't fit in 4 MiB (e.g. multi_defconfig). Signed-off-by: Andreas Schwab - Add support for 8 MiB, - Store initial mapping size in head.S for later reuse, - Add comment about large kernels. Signed-off-by: Geert Uytterhoeven

m68k/atari: fix SCC initialization for debug console

2014-05-28T08:11:53Z

Fix SCC initialization for Atari as was previously fixed for Mac. It's probably not practical to share more code but some attempt is made to align the Mac and Atari variants. Signed-off-by: Finn Thain Signed-off-by: Geert Uytterhoeven

m68k/mvme16x: Adopt common boot console

2014-05-28T08:11:53Z

In a multi-platform kernel binary we only need one early console instance. The difficulty here is that the common early console is started by early_param(), whereas the MVME16x instance is started later by config_mvme16x(). That means some interrupt setup must be done earlier. Signed-off-by: Finn Thain Tested-by: Stephen N Chivers [Geert] Tag debug_cons_write() with __ref to kill section mismatch warning Signed-off-by: Geert Uytterhoeven

m68k: Multi-platform EARLY_PRINTK

2014-05-28T08:10:04Z

Make the boot console available to more m68k platforms by leveraging the head.S debug console. The boot console is enabled by the "earlyprintk" command line argument which is how most other architectures do this. This is a change of behaviour for the Mac but does not negatively impact the common use-case which is not debugging. This is also a change of behaviour for other platforms because it means the serial port stays quiet when CONFIG_EARLY_PRINTK is not enabled. This is also an improvement for the common use-case. Signed-off-by: Finn Thain Tested-by: Stephen N Chivers [Geert: CONSOLE_DEBUG should depend on CONFIG_FONT_SUPPORT] Reported-by: kbuild test robot Signed-off-by: Geert Uytterhoeven

m68k: Toward platform agnostic framebuffer debug logging

2014-05-26T20:41:25Z

Code subject to #ifdef CONSOLE is made more generic, as was apparently intended by the original author. Remove console_put_stats() routine. If it should be somehow useful, it should also be useful on platforms without framebuffer debug logging. The present implementation is only built #if defined CONFIG_MAC && defined CONSOLE even though puts() works everywhere. Signed-off-by: Finn Thain Tested-by: Stephen N Chivers Signed-off-by: Geert Uytterhoeven

m68k: head.S - Remove bogus L prefix in comment

2014-03-10T20:15:08Z

Signed-off-by: Geert Uytterhoeven

m68k: Remove dead code

2014-03-10T20:15:07Z

Signed-off-by: Finn Thain Signed-off-by: Geert Uytterhoeven

m68k: Remove CONSOLE_PENGUIN macro, adopt CONFIG_LOGO

2014-03-10T20:15:06Z

Allow CONFIG_LOGO to enable/disable the head.S penguin logo as well as the framebuffer console logo. This should save a few bytes. It also gets rid of the obscure CONSOLE_PENGUIN macro. Signed-off-by: Finn Thain Signed-off-by: Geert Uytterhoeven

m68k/mac: Make SCC reset work more reliably

2014-01-19T10:53:22Z

For SCC initialization we cannot assume that the control register is in the correct state to accept a register pointer. So first read from the control register in order to "sync" up. Signed-off-by: Finn Thain Signed-off-by: Geert Uytterhoeven