diff options
author | Michal Simek | 2011-01-31 15:10:04 +0100 |
---|---|---|
committer | Michal Simek | 2011-03-09 08:09:54 +0100 |
commit | 6e83557c38b40d6e9d1c82ad0ae59d8e5db9c50c (patch) | |
tree | 856d8cf13273c2f8a696473ff9237cdc17c46632 /arch/microblaze/include/asm/entry.h | |
parent | microblaze: Do not use r0_ram space for syscall debugging (diff) | |
download | kernel-qcow2-linux-6e83557c38b40d6e9d1c82ad0ae59d8e5db9c50c.tar.gz kernel-qcow2-linux-6e83557c38b40d6e9d1c82ad0ae59d8e5db9c50c.tar.xz kernel-qcow2-linux-6e83557c38b40d6e9d1c82ad0ae59d8e5db9c50c.zip |
microblaze: Remove r0_ram pointer and PTO alignment
r0_ram pool was used for saving/restoring register
content if hw exception happen. This poll was replaced by
pt_pool_space with PT_SIZE size.
Based on this change SAVE_STATE_ARG_SPACE was removed which
caused that PTO offset is zero that's why is also removed.
r0_ram space was used as scratchpad by v850. In early
Microblaze Linux developing phase was this part of code
blindly copied.
Signed-off-by: Michal Simek <monstr@monstr.eu>
Diffstat (limited to 'arch/microblaze/include/asm/entry.h')
-rw-r--r-- | arch/microblaze/include/asm/entry.h | 36 |
1 files changed, 0 insertions, 36 deletions
diff --git a/arch/microblaze/include/asm/entry.h b/arch/microblaze/include/asm/entry.h index ec89f2ad0fe1..af0144b91b79 100644 --- a/arch/microblaze/include/asm/entry.h +++ b/arch/microblaze/include/asm/entry.h @@ -31,40 +31,4 @@ DECLARE_PER_CPU(unsigned int, R11_SAVE); /* Temp variable for entry */ DECLARE_PER_CPU(unsigned int, CURRENT_SAVE); /* Saved current pointer */ # endif /* __ASSEMBLY__ */ -#ifndef CONFIG_MMU - -/* noMMU hasn't any space for args */ -# define STATE_SAVE_ARG_SPACE (0) - -#else /* CONFIG_MMU */ - -/* If true, system calls save and restore all registers (except result - * registers, of course). If false, then `call clobbered' registers - * will not be preserved, on the theory that system calls are basically - * function calls anyway, and the caller should be able to deal with it. - * This is a security risk, of course, as `internal' values may leak out - * after a system call, but that certainly doesn't matter very much for - * a processor with no MMU protection! For a protected-mode kernel, it - * would be faster to just zero those registers before returning. - * - * I can not rely on the glibc implementation. If you turn it off make - * sure that r11/r12 is saved in user-space. --KAA - * - * These are special variables using by the kernel trap/interrupt code - * to save registers in, at a time when there are no spare registers we - * can use to do so, and we can't depend on the value of the stack - * pointer. This means that they must be within a signed 16-bit - * displacement of 0x00000000. - */ - -/* A `state save frame' is a struct pt_regs preceded by some extra space - * suitable for a function call stack frame. */ - -/* Amount of room on the stack reserved for arguments and to satisfy the - * C calling conventions, in addition to the space used by the struct - * pt_regs that actually holds saved values. */ -#define STATE_SAVE_ARG_SPACE (6*4) /* Up to six arguments */ - -#endif /* CONFIG_MMU */ - #endif /* _ASM_MICROBLAZE_ENTRY_H */ |