1. 16 2月, 2013 5 次提交
    • V
      ARC: SMP support · 41195d23
      Vineet Gupta 提交于
      ARC common code to enable a SMP system + ISS provided SMP extensions.
      
      ARC700 natively lacks SMP support, hence some of the core features are
      are only enabled if SoCs have the necessary h/w pixie-dust. This
      includes:
      -Inter Processor Interrupts (IPI)
      -Cache coherency
      -load-locked/store-conditional
      ...
      
      The low level exception handling would be completely broken in SMP
      because we don't have hardware assisted stack switching. Thus a fair bit
      of this code is repurposing the MMU_SCRATCH reg for event handler
      prologues to keep them re-entrant.
      
      Many thanks to Rajeshwar Ranga for his initial "major" contributions to
      SMP Port (back in 2008), and to Noam Camus and Gilad Ben-Yossef for help
      with resurrecting that in 3.2 kernel (2012).
      
      Note that this platform code is again singleton design pattern - so
      multiple SMP platforms won't build at the moment - this deficiency is
      addressed in subsequent patches within this series.
      Signed-off-by: NVineet Gupta <vgupta@synopsys.com>
      Cc: Arnd Bergmann <arnd@arndb.de>
      Cc: Thomas Gleixner <tglx@linutronix.de>
      Cc: Rajeshwar Ranga <rajeshwar.ranga@gmail.com>
      Cc: Noam Camus <noamc@ezchip.com>
      Cc: Gilad Ben-Yossef <gilad@benyossef.com>
      41195d23
    • V
      ARC: Support for high priority interrupts in the in-core intc · 4788a594
      Vineet Gupta 提交于
      There is a bit of hack/kludge right now where we disable preemption if a
      L2 (High prio) IRQ is taken while L1 (Low prio) is active.
      
      Need to revisit this
      Signed-off-by: NVineet Gupta <vgupta@synopsys.com>
      4788a594
    • V
      ARC: [optim] Cache "current" in Register r25 · 080c3747
      Vineet Gupta 提交于
      Signed-off-by: NVineet Gupta <vgupta@synopsys.com>
      080c3747
    • V
      ARC: [Review] Preparing to fix incorrect syscall restarts due to signals · 5c39c0ab
      Vineet Gupta 提交于
      To avoid multiple syscall restarts (multiple signals) or no restart at
      all (sigreturn), we need just an extra bit of state "literally 1 bit" in
      struct pt_regs. orig_r8 is the best place to do this, however given the
      way it is encoded currently, we can't add anything simplistically.
      
      Current orig_r8:
      * syscalls   -> 1 to NR_SYSCALLS
      * Exceptions -> NR_SYSCALLS + 1
      * Break-point-> NR_SYSCALLS + 2
      
      In new scheme it is a bit-field
      * lower short word contains the  exact event type (and a new bit to represent
         restart semantics : if syscall was already / can't be restarted)
      * upper short word optionally containing the syscall num - needed by
        likes of tracehooks etc
      
      This patch only changes how orig_r8 is organised and nothing should
      change behaviourily.
      Signed-off-by: NVineet Gupta <vgupta@synopsys.com>
      Cc: Al Viro <viro@ZenIV.linux.org.uk>
      5c39c0ab
    • V
      ARC: Signal handling · c3581039
      Vineet Gupta 提交于
      Includes following fixes courtesy review by Al-Viro
      
      * Tracer poke to Callee-regs were lost
      
        Before going off into do_signal( ) we save the user-mode callee regs
        (as they are not saved by default as part of pt_regs). This is to make
        sure that that a Tracer (if tracing related signal) is able to do likes
        of PEEKUSR(callee-reg).
      
        However in return path we were simply discarding the user-mode callee
        regs, which would break a POKEUSR(callee-reg) from a tracer.
      
      * Issue related to multiple syscall restarts are addressed in next patch
      Signed-off-by: NVineet Gupta <vgupta@synopsys.com>
      Cc: Al Viro <viro@ZenIV.linux.org.uk>
      Acked-by: NJonas Bonn <jonas@southpole.se>
      c3581039
  2. 11 2月, 2013 1 次提交