qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Bug 1850378] Re: RISC-V unreliable IPIs


From: tm
Subject: [Bug 1850378] Re: RISC-V unreliable IPIs
Date: Wed, 30 Oct 2019 09:44:14 -0000

** Description changed:

  I am working on a project with custom inter processor interrupts (IPIs) on 
the RISC-V virt machine.
  After upgrading from version 3.1.0 to 4.1.0 which fixes a related issue 
(https://github.com/riscv/riscv-qemu/issues/132) I am able to use the CPU 
hotplug feature.
  
  However, if I try to use IPIs for communication between two cores, the
  wfi instruction behaves strangely. Either it does not return, or it
  returns on timer interrupts, even though they are disabled. The code, I
  use on one core to wait for an interrupt is the following.
  
-       csr_clear(sie, SIE_SEIE | SIE_STIE);
-       do {
-               wait_for_interrupt();
-               sipval = csr_read(sip);
-               sieval = csr_read(sie);
-               scauseval = csr_read(scause) & 0xFF;
-       /* only break if wfi returns for an software interrupt */
-       } while ((sipval & sieval) == 0 && scauseval != 1);
-       csr_set(sie, SIE_SEIE | SIE_STIE);
+  csr_clear(sie, SIE_SEIE | SIE_STIE);
+  do {
+   wait_for_interrupt();
+   sipval = csr_read(sip);
+   sieval = csr_read(sie);
+   scauseval = csr_read(scause) & 0xFF;
+  /* only break if wfi returns for an software interrupt */
+  } while ((sipval & sieval) == 0 && scauseval != 1);
+  csr_set(sie, SIE_SEIE | SIE_STIE);
  
  Since the resulting sequence does not seem to be deterministic, my guess
  is, that it has something to do with the communication of qemu's threads
  for the different cores.
+ 
+ Update:
+ The exact same setup works fine in spike (the actual sim, not the qemu 
board), which might give a hint, that it is related to the interrupt controller 
implementation.

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1850378

Title:
  RISC-V unreliable IPIs

Status in QEMU:
  New

Bug description:
  I am working on a project with custom inter processor interrupts (IPIs) on 
the RISC-V virt machine.
  After upgrading from version 3.1.0 to 4.1.0 which fixes a related issue 
(https://github.com/riscv/riscv-qemu/issues/132) I am able to use the CPU 
hotplug feature.

  However, if I try to use IPIs for communication between two cores, the
  wfi instruction behaves strangely. Either it does not return, or it
  returns on timer interrupts, even though they are disabled. The code,
  I use on one core to wait for an interrupt is the following.

   csr_clear(sie, SIE_SEIE | SIE_STIE);
   do {
    wait_for_interrupt();
    sipval = csr_read(sip);
    sieval = csr_read(sie);
    scauseval = csr_read(scause) & 0xFF;
   /* only break if wfi returns for an software interrupt */
   } while ((sipval & sieval) == 0 && scauseval != 1);
   csr_set(sie, SIE_SEIE | SIE_STIE);

  Since the resulting sequence does not seem to be deterministic, my
  guess is, that it has something to do with the communication of qemu's
  threads for the different cores.

  Update:
  The exact same setup works fine in spike (the actual sim, not the qemu 
board), which might give a hint, that it is related to the interrupt controller 
implementation.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1850378/+subscriptions



reply via email to

[Prev in Thread] Current Thread [Next in Thread]