linux/arch/arc, branch v3.9

ARC: Add implicit compiler barrier to raw_local_irq* functions

2013-04-08T23:10:26Z

ARC irqsave/restore macros were missing the compiler barrier, causing a stale load in irq-enabled region be used in irq-safe region, despite being changed, because the register holding the value was still live. The problem manifested as random crashes in timer code when stress testing ARCLinux (3.9-rc3) on a !SMP && !PREEMPT_COUNT Here's the exact sequence which caused this: (0). tv1[x] <----> t1 <---> t2 (1). mod_timer(t1) interrupted after it calls timer_pending() (2). mod_timer(t2) completes (3). mod_timer(t1) resumes but messes up the list (4). __runt_timers( ) uses bogus timer_list entry / crashes in timer->function Essentially mod_timer() was racing against itself and while the spinlock serialized the tv1[] timer link list, timer_pending() called outside the spinlock, cached timer link list element in a register. With low register pressure (and a deep register file), lack of barrier in raw_local_irqsave() as well as preempt_disable (!PREEMPT_COUNT version), there was nothing to force gcc to reload across the spinlock, causing a stale value in reg be used for link list manipulation - ensuing a corruption. ARcompact disassembly which shows the culprit generated code: mod_timer: push_s blink mov_s r13,r0 # timer, timer .. ###### timer_pending( ) ld_s r3,[r13] # <------ .entry.next LOADED brne r3, 0, @.L163 .L163: .. ###### spin_lock_irq( ) lr r5, [status32] # flags bic r4, r5, 6 # temp, flags, and.f 0, r5, 6 # flags, flag.nz r4 ###### detach_if_pending( ) begins tst_s r3,r3 <-------------- # timer_pending( ) checks timer->entry.next # r3 is NOT reloaded by gcc, using stale value beq.d @.L169 mov.eq r0,0 ##### detach_timer( ): __list_del( ) ld r4,[r13,4] # .entry.prev, D.31439 st r4,[r3,4] # .prev, D.31439 st r3,[r4] # .next, D.30246 We initially tried to fix this by adding barrier() to preempt_* macros for !PREEMPT_COUNT but Linus clarified that it was anything but wrong. http://www.spinics.net/lists/kernel/msg1512709.html [vgupta: updated commitlog] Reported-by/Signed-off-by: Christian Ruppert Cc: Christian Ruppert Cc: Pierrick Hascoet Debugged-by/Signed-off-by: Vineet Gupta Signed-off-by: Linus Torvalds

ARC: Fix the typo in event identifier flags used by ptrace

2013-03-20T13:15:45Z

orig_r8_IS_EXCPN and orig_r8_IS_BRKPT were same values due to a copy/paste error. Although it looks bad and is wrong, it really doesn't affect gdb working. orig_r8_IS_BRKPT is the one relevant to debugging (breakpoints), since it is used to provide EFA vs. ERET to a ptrace "stop_pc" request. So when gdb has inserted a breakpoint, orig_r8_IS_BRKPT is already set, and anything else (i.e. orig_r8_IS_EXCPN) becoming same as it, really doesn't hurt gdb. The corollary case, could be nasty but nobody uses the ptrace "stop_pc" request in that case Signed-off-by: Vineet Gupta

arc: fix dma_address assignment during dma_map_sg()

2013-03-19T10:04:53Z

Signed-off-by: Pierrick Hascoet Signed-off-by: Vineet Gupta

ARC: Remove SET_PERSONALITY (tracks cross-arch change)

2013-03-18T09:07:05Z

Tracks commit e72837e3e7b "default SET_PERSONALITY() in linux/elf.h" Signed-off-by: Vineet Gupta

ARC: ABIv3: fork/vfork wrappers not needed in "no-legacy-syscall" ABI

2013-03-11T13:31:10Z

When switching to clone() only ABI - I missed out pruning the low level asm syscall wrappers Signed-off-by: Vineet Gupta

ARC: ABIv3: Print the correct ABI ver

2013-03-11T13:31:09Z

Signed-off-by: Vineet Gupta

ARC: make allyesconfig build breakages

2013-03-11T13:31:09Z

CC drivers/mmc/host/mmc_spi.o drivers/mmc/host/mmc_spi.c:118: error: redefinition of 'struct scratch' make[3]: *** [drivers/mmc/host/mmc_spi.o] Error 1 make[2]: *** [drivers/mmc/host] Error 2 make[1]: *** [drivers/mmc] Error 2 make: *** [drivers] Error 2 CC arch/arc/kernel/kgdb.o In file included from include/linux/kgdb.h:20, from arch/arc/kernel/kgdb.c:11: /home/vineetg/arc/k.org/arc-port/arch/arc/include/asm/kgdb.h:34: warning: 'struct pt_regs' declared inside parameter list /home/vineetg/arc/k.org/arc-port/arch/arc/include/asm/kgdb.h:34: warning: its scope is only this definition or declaration, which is probably not what you want arch/arc/kernel/kgdb.c:172: error: conflicting types for 'kgdb_trap' CC arch/arc/kernel/kgdb.o arch/arc/kernel/kgdb.c: In function 'pt_regs_to_gdb_regs': arch/arc/kernel/kgdb.c:62: error: dereferencing pointer to incomplete type Signed-off-by: Vineet Gupta

ARC: split elf.h into uapi and export it for userspace

2013-02-27T14:30:26Z

Signed-off-by: Vineet Gupta

ARC: Fixup the current ABI version

2013-02-27T14:30:25Z

The upstream kernel ABI (v3) is different from current out-of-tree (v2): * no-legacy-syscalls * user_regs_struct layout has changed So we rev up the ABI version Signed-off-by: Vineet Gupta

ARC: gdbserver using regset interface possibly broken

2013-02-27T14:29:55Z

ptrace regset interface relies on ELF_NGREG for ceiling the size of user request. So any larger request (even if legit) would be clipped. The existing def of ELF_NGREG didn't use user_regs_struct and was technically one placeholder short (stop_pc) - although the current code would still work because pt_regs includes a bunch of extra fields, making ELF_NGREG >= sizeof(struct user_regs_struct)/sizeof(long) But we need to remove this ambiguity, specially since pt_regs should NOT be directly associated with with anything userspace-ish. Signed-off-by: Vineet Gupta