| Message ID | 20240807000652.1417776-2-debug@rivosinc.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | riscv support for control flow integrity extensions | expand |
On 8/7/24 10:06, Deepak Gupta wrote: > commit 16ad9788 [1] restricted icount to qemu-system only. Although > assert in `cpu_loop_exec_tb` is on `icount_enabled()` which is 0 when > its qemu-user and debug build starts asserting. > Move assert for qemu-system. > > [1] - https://lists.gnu.org/archive/html/qemu-riscv/2024-01/msg00608.html > > Signed-off-by: Deepak Gupta <debug@rivosinc.com> > --- > accel/tcg/cpu-exec.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c > index 245fd6327d..8cc2a6104f 100644 > --- a/accel/tcg/cpu-exec.c > +++ b/accel/tcg/cpu-exec.c > @@ -927,9 +927,9 @@ static inline void cpu_loop_exec_tb(CPUState *cpu, TranslationBlock *tb, > return; > } > > +#ifndef CONFIG_USER_ONLY > /* Instruction counter expired. */ > assert(icount_enabled()); > -#ifndef CONFIG_USER_ONLY > /* Ensure global icount has gone forward */ > icount_update(cpu); > /* Refill decrementer and continue execution. */ No, this is a real bug. Just above we handled (1) exit for tcg (non-)chaining (!= TB_EXIT_REQUESTED), (2) exit for exception/interrupt (cpu_loop_exit_requested). The only thing that is left is exit for icount expired. And for that we *must* have icount enabled. How did you encounter this? r~
On Wed, Aug 07, 2024 at 10:48:56AM +1000, Richard Henderson wrote: >On 8/7/24 10:06, Deepak Gupta wrote: >>commit 16ad9788 [1] restricted icount to qemu-system only. Although >>assert in `cpu_loop_exec_tb` is on `icount_enabled()` which is 0 when >>its qemu-user and debug build starts asserting. >>Move assert for qemu-system. >> >>[1] - https://lists.gnu.org/archive/html/qemu-riscv/2024-01/msg00608.html >> >>Signed-off-by: Deepak Gupta <debug@rivosinc.com> >>--- >> accel/tcg/cpu-exec.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >>diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c >>index 245fd6327d..8cc2a6104f 100644 >>--- a/accel/tcg/cpu-exec.c >>+++ b/accel/tcg/cpu-exec.c >>@@ -927,9 +927,9 @@ static inline void cpu_loop_exec_tb(CPUState *cpu, TranslationBlock *tb, >> return; >> } >>+#ifndef CONFIG_USER_ONLY >> /* Instruction counter expired. */ >> assert(icount_enabled()); >>-#ifndef CONFIG_USER_ONLY >> /* Ensure global icount has gone forward */ >> icount_update(cpu); >> /* Refill decrementer and continue execution. */ > >No, this is a real bug. > >Just above we handled > > (1) exit for tcg (non-)chaining (!= TB_EXIT_REQUESTED), > (2) exit for exception/interrupt (cpu_loop_exit_requested). > >The only thing that is left is exit for icount expired. >And for that we *must* have icount enabled. > >How did you encounter this? hmm I was experimenting with reducing TB flags (i.e. not using two different TB flags for zicfilp). But I swear, I didn't see it go away (for qemu-user) when I had switched to two TB flags. Now that you've pointed it out specifically, I tried again. Its not asserting at this place anymore for qemu-ser I'll remove this patch in next version. And if I encounter this again, will dig a little bit more deep into it and try to get repro steps. > > >r~
On Wed, Aug 07, 2024 at 10:48:56AM +1000, Richard Henderson wrote: >On 8/7/24 10:06, Deepak Gupta wrote: >>commit 16ad9788 [1] restricted icount to qemu-system only. Although >>assert in `cpu_loop_exec_tb` is on `icount_enabled()` which is 0 when >>its qemu-user and debug build starts asserting. >>Move assert for qemu-system. >> >>[1] - https://lists.gnu.org/archive/html/qemu-riscv/2024-01/msg00608.html >> >>Signed-off-by: Deepak Gupta <debug@rivosinc.com> >>--- >> accel/tcg/cpu-exec.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >>diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c >>index 245fd6327d..8cc2a6104f 100644 >>--- a/accel/tcg/cpu-exec.c >>+++ b/accel/tcg/cpu-exec.c >>@@ -927,9 +927,9 @@ static inline void cpu_loop_exec_tb(CPUState *cpu, TranslationBlock *tb, >> return; >> } >>+#ifndef CONFIG_USER_ONLY >> /* Instruction counter expired. */ >> assert(icount_enabled()); >>-#ifndef CONFIG_USER_ONLY >> /* Ensure global icount has gone forward */ >> icount_update(cpu); >> /* Refill decrementer and continue execution. */ > >No, this is a real bug. > >Just above we handled > > (1) exit for tcg (non-)chaining (!= TB_EXIT_REQUESTED), > (2) exit for exception/interrupt (cpu_loop_exit_requested). > >The only thing that is left is exit for icount expired. >And for that we *must* have icount enabled. > >How did you encounter this? I spent last week incorporating your suggestions. And during the flux of it, I started seeing this issue again. As soon as I switch to branch from where I sent the patches out, this icount assert issue disappears. So something definitley is triggering the issue. It happens only when zicfilp and zicfiss are enabled. I am still trying to root cause and in a fog right now. Although I am not very well versed with tcg internals. So any pointer here which could help me debug it faster would be well appreciated. Thanks. > > >r~
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c index 245fd6327d..8cc2a6104f 100644 --- a/accel/tcg/cpu-exec.c +++ b/accel/tcg/cpu-exec.c @@ -927,9 +927,9 @@ static inline void cpu_loop_exec_tb(CPUState *cpu, TranslationBlock *tb, return; } +#ifndef CONFIG_USER_ONLY /* Instruction counter expired. */ assert(icount_enabled()); -#ifndef CONFIG_USER_ONLY /* Ensure global icount has gone forward */ icount_update(cpu); /* Refill decrementer and continue execution. */
commit 16ad9788 [1] restricted icount to qemu-system only. Although assert in `cpu_loop_exec_tb` is on `icount_enabled()` which is 0 when its qemu-user and debug build starts asserting. Move assert for qemu-system. [1] - https://lists.gnu.org/archive/html/qemu-riscv/2024-01/msg00608.html Signed-off-by: Deepak Gupta <debug@rivosinc.com> --- accel/tcg/cpu-exec.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)