From 07500b0d855b7f3f47ca263b21b6397d743d45d2 Mon Sep 17 00:00:00 2001 From: Ralf Baechle Date: Tue, 30 Oct 2007 17:25:26 +0000 Subject: [PATCH] [MIPS] Fix context DSP context / TLS pointer switching bug for new threads. A new born thread starts execution not in schedule but rather in ret_from_fork which results in it bypassing the part of the code to load a new context written in C which are the DSP context and the userlocal register which Linux uses for the TLS pointer. Frequently we were just getting away with this bug for a number of reasons: o Real world application scenarios are very unlikely to use clone or fork in blocks of DSP code. o Linux by default runs the child process right after the fork, so the child by luck will find all the right context in the DSP and userlocal registers. o So far the rdhwr instruction was emulated on all hardware so userlocal wasn't getting referenced at all and the emulation wasn't suffering from the issue since it gets it's value straight from the thread's thread_info. Fixed by moving the code to load the context from switch_to() to finish_arch_switch which will be called by newborn and old threads. Signed-off-by: Ralf Baechle --- include/asm-mips/system.h | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/include/asm-mips/system.h b/include/asm-mips/system.h index 90e4b403f531..1030562d6ea6 100644 --- a/include/asm-mips/system.h +++ b/include/asm-mips/system.h @@ -68,11 +68,15 @@ do { \ if (cpu_has_dsp) \ __save_dsp(prev); \ (last) = resume(prev, next, task_thread_info(next)); \ +} while (0) + +#define finish_arch_switch(prev) \ +do { \ if (cpu_has_dsp) \ __restore_dsp(current); \ if (cpu_has_userlocal) \ - write_c0_userlocal(task_thread_info(current)->tp_value);\ -} while(0) + write_c0_userlocal(current_thread_info()->tp_value); \ +} while (0) static inline unsigned long __xchg_u32(volatile int * m, unsigned int val) { -- GitLab