Re: riscv64: icache flush using sysarch(2)

2021-09-13 Thread Theo de Raadt
Mark Kettenis wrote: > Really depends on how the future riscv64 instructions behave. A valid > concern so the XXX is appropriate. But I would pass the address and > size to pmap_proc_iflush(). Yes, let the pmap decide what it wants to do.

Re: riscv64: icache flush using sysarch(2)

2021-09-13 Thread Mark Kettenis
> From: Jeremie Courreges-Anglas > Date: Mon, 13 Sep 2021 21:35:33 +0200 > > On Thu, Sep 09 2021, Mark Kettenis wrote: > >> From: Jeremie Courreges-Anglas > >> Date: Sun, 05 Sep 2021 21:41:35 +0200 > >> > >> On Sat, Sep 04 2021, Jeremie Courreges-Anglas wrote: > >> > The first problem I was

Re: riscv64: icache flush using sysarch(2)

2021-09-13 Thread Jeremie Courreges-Anglas
On Thu, Sep 09 2021, Mark Kettenis wrote: >> From: Jeremie Courreges-Anglas >> Date: Sun, 05 Sep 2021 21:41:35 +0200 >> >> On Sat, Sep 04 2021, Jeremie Courreges-Anglas wrote: >> > The first problem I was able to diagnose using egdb on riscv was >> > lang/python/2.7 using libffi and aborting

Re: riscv64: icache flush using sysarch(2)

2021-09-09 Thread Mark Kettenis
> From: Jeremie Courreges-Anglas > Date: Sun, 05 Sep 2021 21:41:35 +0200 > > On Sat, Sep 04 2021, Jeremie Courreges-Anglas wrote: > > The first problem I was able to diagnose using egdb on riscv was > > lang/python/2.7 using libffi and aborting in libcompiler-rt (the > > compilerrt_abort() call

Re: riscv64: icache flush using sysarch(2)

2021-09-05 Thread Jeremie Courreges-Anglas
On Sat, Sep 04 2021, Jeremie Courreges-Anglas wrote: > The first problem I was able to diagnose using egdb on riscv was > lang/python/2.7 using libffi and aborting in libcompiler-rt (the > compilerrt_abort() call below). > > --8<-- > #elif defined(__riscv) && defined(__linux__) > #define