Re: [PATCH] riscv: sifive_test: Allow 16-bit writes to memory region
Quoting Alistair Francis (2020-09-14 11:07:26) > On Fri, Sep 11, 2020 at 5:26 AM Michael Roth > wrote: > > > > Quoting Alistair Francis (2020-09-10 13:10:43) > > > On Thu, Sep 3, 2020 at 2:05 PM Michael Roth > > > wrote: > > > > > > > > Quoting Alistair Francis (2020-09-01 18:59:29) > > > > > On Mon, Aug 31, 2020 at 10:59 PM Nathan Chancellor > > > > > wrote: > > > > > > > > > > > > When shutting down the machine running a mainline Linux kernel, the > > > > > > following error happens: > > > > > > > > > > > > $ build/riscv64-softmmu/qemu-system-riscv64 -bios default -M virt \ > > > > > > -display none -initrd rootfs.cpio -kernel Image -m 512m \ > > > > > > -nodefaults -serial mon:stdio > > > > > > ... > > > > > > Requesting system poweroff > > > > > > [4.999630] reboot: Power down > > > > > > sbi_trap_error: hart0: trap handler failed (error -2) > > > > > > sbi_trap_error: hart0: mcause=0x0007 > > > > > > mtval=0x0010 > > > > > > sbi_trap_error: hart0: mepc=0x8000d4cc > > > > > > mstatus=0x1822 > > > > > > sbi_trap_error: hart0: ra=0x8000999e sp=0x80015c78 > > > > > > sbi_trap_error: hart0: gp=0xffe000e76610 tp=0xffe0081b89c0 > > > > > > sbi_trap_error: hart0: s0=0x80015c88 s1=0x0040 > > > > > > sbi_trap_error: hart0: a0=0x a1=0x80004024 > > > > > > sbi_trap_error: hart0: a2=0x80004024 a3=0x80004024 > > > > > > sbi_trap_error: hart0: a4=0x0010 a5=0x > > > > > > sbi_trap_error: hart0: a6=0x4024 a7=0x80011158 > > > > > > sbi_trap_error: hart0: s2=0x s3=0x80016000 > > > > > > sbi_trap_error: hart0: s4=0x s5=0x > > > > > > sbi_trap_error: hart0: s6=0x0001 s7=0x > > > > > > sbi_trap_error: hart0: s8=0x s9=0x > > > > > > sbi_trap_error: hart0: s10=0x s11=0x0008 > > > > > > sbi_trap_error: hart0: t0=0x t1=0x > > > > > > sbi_trap_error: hart0: t2=0x t3=0x > > > > > > sbi_trap_error: hart0: t4=0x t5=0x > > > > > > sbi_trap_error: hart0: t6=0x > > > > > > > > > > > > The kernel does a 16-bit write when powering off the machine, which > > > > > > was allowed before commit 5d971f9e67 ("memory: Revert "memory: > > > > > > accept > > > > > > mismatching sizes in memory_region_access_valid""). Make > > > > > > min_access_size > > > > > > match reality so that the machine can shut down properly now. > > > > > > > > > > > > Cc: qemu-sta...@nongnu.org > > > > > > Fixes: 88a07990fa ("SiFive RISC-V Test Finisher") > > > > > > Fixes: 5d971f9e67 ("memory: Revert "memory: accept mismatching > > > > > > sizes in memory_region_access_valid"") > > > > > > Signed-off-by: Nathan Chancellor > > > > > > > > > > Thanks! > > > > > > > > > > Applied to riscv-to-apply.next > > > > > > > > FYI, I'm hoping to pull this patch into the upcoming 5.0.1 stable > > > > release. The freeze is scheduled for 2020-09-20, I will apply this if it > > > > hits master before then. > > > > > > I just sent a PR with this patch. Is this still on track to make it into > > > 5.0.1? > > > > Since it's not likely to invalidate any testing on my end outside of the > > ones built into QEMU I can probably still slip it in if it hits master > > by Monday, or maybe just grab it from your tree. > > This is in master now, I hope that's enough time to make it in. Yes I pulled it in yesterday: https://github.com/mdroth/qemu/commits/stable-5.0-staging > > Let me know if you want me to do anything else. Thanks! > > Alistair > > > > > > > > > Alistair > > > > > > > > > > > > > > > > > Alistair > > > > > > > > > > > --- > > > > > > > > > > > > Please let me know if the tags are wrong or inappropriate, this is > > > > > > my > > > > > > first QEMU patch. > > > > > > > > > > > > hw/riscv/sifive_test.c | 2 +- > > > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > > > > > diff --git a/hw/riscv/sifive_test.c b/hw/riscv/sifive_test.c > > > > > > index 0c78fb2c93..8c70dd69df 100644 > > > > > > --- a/hw/riscv/sifive_test.c > > > > > > +++ b/hw/riscv/sifive_test.c > > > > > > @@ -59,7 +59,7 @@ static const MemoryRegionOps sifive_test_ops = { > > > > > > .write = sifive_test_write, > > > > > > .endianness = DEVICE_NATIVE_ENDIAN, > > > > > > .valid = { > > > > > > -.min_access_size = 4, > > > > > > +.min_access_size = 2, > > > > > > .max_access_size = 4 > > > > > > } > > > > > > }; > > > > > > > > > > > > base-commit: 2f4c51c0f384d7888a04b4815861e6d5fd244d75 > > > > > > -- > > > > > > 2.28.0 > > > > > > > > > > > > > > > > >
Re: [PATCH] riscv: sifive_test: Allow 16-bit writes to memory region
On Fri, Sep 11, 2020 at 5:26 AM Michael Roth wrote: > > Quoting Alistair Francis (2020-09-10 13:10:43) > > On Thu, Sep 3, 2020 at 2:05 PM Michael Roth > > wrote: > > > > > > Quoting Alistair Francis (2020-09-01 18:59:29) > > > > On Mon, Aug 31, 2020 at 10:59 PM Nathan Chancellor > > > > wrote: > > > > > > > > > > When shutting down the machine running a mainline Linux kernel, the > > > > > following error happens: > > > > > > > > > > $ build/riscv64-softmmu/qemu-system-riscv64 -bios default -M virt \ > > > > > -display none -initrd rootfs.cpio -kernel Image -m 512m \ > > > > > -nodefaults -serial mon:stdio > > > > > ... > > > > > Requesting system poweroff > > > > > [4.999630] reboot: Power down > > > > > sbi_trap_error: hart0: trap handler failed (error -2) > > > > > sbi_trap_error: hart0: mcause=0x0007 > > > > > mtval=0x0010 > > > > > sbi_trap_error: hart0: mepc=0x8000d4cc > > > > > mstatus=0x1822 > > > > > sbi_trap_error: hart0: ra=0x8000999e sp=0x80015c78 > > > > > sbi_trap_error: hart0: gp=0xffe000e76610 tp=0xffe0081b89c0 > > > > > sbi_trap_error: hart0: s0=0x80015c88 s1=0x0040 > > > > > sbi_trap_error: hart0: a0=0x a1=0x80004024 > > > > > sbi_trap_error: hart0: a2=0x80004024 a3=0x80004024 > > > > > sbi_trap_error: hart0: a4=0x0010 a5=0x > > > > > sbi_trap_error: hart0: a6=0x4024 a7=0x80011158 > > > > > sbi_trap_error: hart0: s2=0x s3=0x80016000 > > > > > sbi_trap_error: hart0: s4=0x s5=0x > > > > > sbi_trap_error: hart0: s6=0x0001 s7=0x > > > > > sbi_trap_error: hart0: s8=0x s9=0x > > > > > sbi_trap_error: hart0: s10=0x s11=0x0008 > > > > > sbi_trap_error: hart0: t0=0x t1=0x > > > > > sbi_trap_error: hart0: t2=0x t3=0x > > > > > sbi_trap_error: hart0: t4=0x t5=0x > > > > > sbi_trap_error: hart0: t6=0x > > > > > > > > > > The kernel does a 16-bit write when powering off the machine, which > > > > > was allowed before commit 5d971f9e67 ("memory: Revert "memory: accept > > > > > mismatching sizes in memory_region_access_valid""). Make > > > > > min_access_size > > > > > match reality so that the machine can shut down properly now. > > > > > > > > > > Cc: qemu-sta...@nongnu.org > > > > > Fixes: 88a07990fa ("SiFive RISC-V Test Finisher") > > > > > Fixes: 5d971f9e67 ("memory: Revert "memory: accept mismatching sizes > > > > > in memory_region_access_valid"") > > > > > Signed-off-by: Nathan Chancellor > > > > > > > > Thanks! > > > > > > > > Applied to riscv-to-apply.next > > > > > > FYI, I'm hoping to pull this patch into the upcoming 5.0.1 stable > > > release. The freeze is scheduled for 2020-09-20, I will apply this if it > > > hits master before then. > > > > I just sent a PR with this patch. Is this still on track to make it into > > 5.0.1? > > Since it's not likely to invalidate any testing on my end outside of the > ones built into QEMU I can probably still slip it in if it hits master > by Monday, or maybe just grab it from your tree. This is in master now, I hope that's enough time to make it in. Let me know if you want me to do anything else. Alistair > > > > > Alistair > > > > > > > > > > > > > Alistair > > > > > > > > > --- > > > > > > > > > > Please let me know if the tags are wrong or inappropriate, this is my > > > > > first QEMU patch. > > > > > > > > > > hw/riscv/sifive_test.c | 2 +- > > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > > > diff --git a/hw/riscv/sifive_test.c b/hw/riscv/sifive_test.c > > > > > index 0c78fb2c93..8c70dd69df 100644 > > > > > --- a/hw/riscv/sifive_test.c > > > > > +++ b/hw/riscv/sifive_test.c > > > > > @@ -59,7 +59,7 @@ static const MemoryRegionOps sifive_test_ops = { > > > > > .write = sifive_test_write, > > > > > .endianness = DEVICE_NATIVE_ENDIAN, > > > > > .valid = { > > > > > -.min_access_size = 4, > > > > > +.min_access_size = 2, > > > > > .max_access_size = 4 > > > > > } > > > > > }; > > > > > > > > > > base-commit: 2f4c51c0f384d7888a04b4815861e6d5fd244d75 > > > > > -- > > > > > 2.28.0 > > > > > > > > > > > > > >
Re: [PATCH] riscv: sifive_test: Allow 16-bit writes to memory region
Quoting Alistair Francis (2020-09-10 13:10:43) > On Thu, Sep 3, 2020 at 2:05 PM Michael Roth wrote: > > > > Quoting Alistair Francis (2020-09-01 18:59:29) > > > On Mon, Aug 31, 2020 at 10:59 PM Nathan Chancellor > > > wrote: > > > > > > > > When shutting down the machine running a mainline Linux kernel, the > > > > following error happens: > > > > > > > > $ build/riscv64-softmmu/qemu-system-riscv64 -bios default -M virt \ > > > > -display none -initrd rootfs.cpio -kernel Image -m 512m \ > > > > -nodefaults -serial mon:stdio > > > > ... > > > > Requesting system poweroff > > > > [4.999630] reboot: Power down > > > > sbi_trap_error: hart0: trap handler failed (error -2) > > > > sbi_trap_error: hart0: mcause=0x0007 > > > > mtval=0x0010 > > > > sbi_trap_error: hart0: mepc=0x8000d4cc > > > > mstatus=0x1822 > > > > sbi_trap_error: hart0: ra=0x8000999e sp=0x80015c78 > > > > sbi_trap_error: hart0: gp=0xffe000e76610 tp=0xffe0081b89c0 > > > > sbi_trap_error: hart0: s0=0x80015c88 s1=0x0040 > > > > sbi_trap_error: hart0: a0=0x a1=0x80004024 > > > > sbi_trap_error: hart0: a2=0x80004024 a3=0x80004024 > > > > sbi_trap_error: hart0: a4=0x0010 a5=0x > > > > sbi_trap_error: hart0: a6=0x4024 a7=0x80011158 > > > > sbi_trap_error: hart0: s2=0x s3=0x80016000 > > > > sbi_trap_error: hart0: s4=0x s5=0x > > > > sbi_trap_error: hart0: s6=0x0001 s7=0x > > > > sbi_trap_error: hart0: s8=0x s9=0x > > > > sbi_trap_error: hart0: s10=0x s11=0x0008 > > > > sbi_trap_error: hart0: t0=0x t1=0x > > > > sbi_trap_error: hart0: t2=0x t3=0x > > > > sbi_trap_error: hart0: t4=0x t5=0x > > > > sbi_trap_error: hart0: t6=0x > > > > > > > > The kernel does a 16-bit write when powering off the machine, which > > > > was allowed before commit 5d971f9e67 ("memory: Revert "memory: accept > > > > mismatching sizes in memory_region_access_valid""). Make min_access_size > > > > match reality so that the machine can shut down properly now. > > > > > > > > Cc: qemu-sta...@nongnu.org > > > > Fixes: 88a07990fa ("SiFive RISC-V Test Finisher") > > > > Fixes: 5d971f9e67 ("memory: Revert "memory: accept mismatching sizes in > > > > memory_region_access_valid"") > > > > Signed-off-by: Nathan Chancellor > > > > > > Thanks! > > > > > > Applied to riscv-to-apply.next > > > > FYI, I'm hoping to pull this patch into the upcoming 5.0.1 stable > > release. The freeze is scheduled for 2020-09-20, I will apply this if it > > hits master before then. > > I just sent a PR with this patch. Is this still on track to make it into > 5.0.1? Since it's not likely to invalidate any testing on my end outside of the ones built into QEMU I can probably still slip it in if it hits master by Monday, or maybe just grab it from your tree. > > Alistair > > > > > > > > > Alistair > > > > > > > --- > > > > > > > > Please let me know if the tags are wrong or inappropriate, this is my > > > > first QEMU patch. > > > > > > > > hw/riscv/sifive_test.c | 2 +- > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > diff --git a/hw/riscv/sifive_test.c b/hw/riscv/sifive_test.c > > > > index 0c78fb2c93..8c70dd69df 100644 > > > > --- a/hw/riscv/sifive_test.c > > > > +++ b/hw/riscv/sifive_test.c > > > > @@ -59,7 +59,7 @@ static const MemoryRegionOps sifive_test_ops = { > > > > .write = sifive_test_write, > > > > .endianness = DEVICE_NATIVE_ENDIAN, > > > > .valid = { > > > > -.min_access_size = 4, > > > > +.min_access_size = 2, > > > > .max_access_size = 4 > > > > } > > > > }; > > > > > > > > base-commit: 2f4c51c0f384d7888a04b4815861e6d5fd244d75 > > > > -- > > > > 2.28.0 > > > > > > > > > > >
Re: [PATCH] riscv: sifive_test: Allow 16-bit writes to memory region
On Thu, Sep 3, 2020 at 2:05 PM Michael Roth wrote: > > Quoting Alistair Francis (2020-09-01 18:59:29) > > On Mon, Aug 31, 2020 at 10:59 PM Nathan Chancellor > > wrote: > > > > > > When shutting down the machine running a mainline Linux kernel, the > > > following error happens: > > > > > > $ build/riscv64-softmmu/qemu-system-riscv64 -bios default -M virt \ > > > -display none -initrd rootfs.cpio -kernel Image -m 512m \ > > > -nodefaults -serial mon:stdio > > > ... > > > Requesting system poweroff > > > [4.999630] reboot: Power down > > > sbi_trap_error: hart0: trap handler failed (error -2) > > > sbi_trap_error: hart0: mcause=0x0007 mtval=0x0010 > > > sbi_trap_error: hart0: mepc=0x8000d4cc mstatus=0x1822 > > > sbi_trap_error: hart0: ra=0x8000999e sp=0x80015c78 > > > sbi_trap_error: hart0: gp=0xffe000e76610 tp=0xffe0081b89c0 > > > sbi_trap_error: hart0: s0=0x80015c88 s1=0x0040 > > > sbi_trap_error: hart0: a0=0x a1=0x80004024 > > > sbi_trap_error: hart0: a2=0x80004024 a3=0x80004024 > > > sbi_trap_error: hart0: a4=0x0010 a5=0x > > > sbi_trap_error: hart0: a6=0x4024 a7=0x80011158 > > > sbi_trap_error: hart0: s2=0x s3=0x80016000 > > > sbi_trap_error: hart0: s4=0x s5=0x > > > sbi_trap_error: hart0: s6=0x0001 s7=0x > > > sbi_trap_error: hart0: s8=0x s9=0x > > > sbi_trap_error: hart0: s10=0x s11=0x0008 > > > sbi_trap_error: hart0: t0=0x t1=0x > > > sbi_trap_error: hart0: t2=0x t3=0x > > > sbi_trap_error: hart0: t4=0x t5=0x > > > sbi_trap_error: hart0: t6=0x > > > > > > The kernel does a 16-bit write when powering off the machine, which > > > was allowed before commit 5d971f9e67 ("memory: Revert "memory: accept > > > mismatching sizes in memory_region_access_valid""). Make min_access_size > > > match reality so that the machine can shut down properly now. > > > > > > Cc: qemu-sta...@nongnu.org > > > Fixes: 88a07990fa ("SiFive RISC-V Test Finisher") > > > Fixes: 5d971f9e67 ("memory: Revert "memory: accept mismatching sizes in > > > memory_region_access_valid"") > > > Signed-off-by: Nathan Chancellor > > > > Thanks! > > > > Applied to riscv-to-apply.next > > FYI, I'm hoping to pull this patch into the upcoming 5.0.1 stable > release. The freeze is scheduled for 2020-09-20, I will apply this if it > hits master before then. I just sent a PR with this patch. Is this still on track to make it into 5.0.1? Alistair > > > > > Alistair > > > > > --- > > > > > > Please let me know if the tags are wrong or inappropriate, this is my > > > first QEMU patch. > > > > > > hw/riscv/sifive_test.c | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/hw/riscv/sifive_test.c b/hw/riscv/sifive_test.c > > > index 0c78fb2c93..8c70dd69df 100644 > > > --- a/hw/riscv/sifive_test.c > > > +++ b/hw/riscv/sifive_test.c > > > @@ -59,7 +59,7 @@ static const MemoryRegionOps sifive_test_ops = { > > > .write = sifive_test_write, > > > .endianness = DEVICE_NATIVE_ENDIAN, > > > .valid = { > > > -.min_access_size = 4, > > > +.min_access_size = 2, > > > .max_access_size = 4 > > > } > > > }; > > > > > > base-commit: 2f4c51c0f384d7888a04b4815861e6d5fd244d75 > > > -- > > > 2.28.0 > > > > > > > >
Re: [PATCH] riscv: sifive_test: Allow 16-bit writes to memory region
Quoting Alistair Francis (2020-09-01 18:59:29) > On Mon, Aug 31, 2020 at 10:59 PM Nathan Chancellor > wrote: > > > > When shutting down the machine running a mainline Linux kernel, the > > following error happens: > > > > $ build/riscv64-softmmu/qemu-system-riscv64 -bios default -M virt \ > > -display none -initrd rootfs.cpio -kernel Image -m 512m \ > > -nodefaults -serial mon:stdio > > ... > > Requesting system poweroff > > [4.999630] reboot: Power down > > sbi_trap_error: hart0: trap handler failed (error -2) > > sbi_trap_error: hart0: mcause=0x0007 mtval=0x0010 > > sbi_trap_error: hart0: mepc=0x8000d4cc mstatus=0x1822 > > sbi_trap_error: hart0: ra=0x8000999e sp=0x80015c78 > > sbi_trap_error: hart0: gp=0xffe000e76610 tp=0xffe0081b89c0 > > sbi_trap_error: hart0: s0=0x80015c88 s1=0x0040 > > sbi_trap_error: hart0: a0=0x a1=0x80004024 > > sbi_trap_error: hart0: a2=0x80004024 a3=0x80004024 > > sbi_trap_error: hart0: a4=0x0010 a5=0x > > sbi_trap_error: hart0: a6=0x4024 a7=0x80011158 > > sbi_trap_error: hart0: s2=0x s3=0x80016000 > > sbi_trap_error: hart0: s4=0x s5=0x > > sbi_trap_error: hart0: s6=0x0001 s7=0x > > sbi_trap_error: hart0: s8=0x s9=0x > > sbi_trap_error: hart0: s10=0x s11=0x0008 > > sbi_trap_error: hart0: t0=0x t1=0x > > sbi_trap_error: hart0: t2=0x t3=0x > > sbi_trap_error: hart0: t4=0x t5=0x > > sbi_trap_error: hart0: t6=0x > > > > The kernel does a 16-bit write when powering off the machine, which > > was allowed before commit 5d971f9e67 ("memory: Revert "memory: accept > > mismatching sizes in memory_region_access_valid""). Make min_access_size > > match reality so that the machine can shut down properly now. > > > > Cc: qemu-sta...@nongnu.org > > Fixes: 88a07990fa ("SiFive RISC-V Test Finisher") > > Fixes: 5d971f9e67 ("memory: Revert "memory: accept mismatching sizes in > > memory_region_access_valid"") > > Signed-off-by: Nathan Chancellor > > Thanks! > > Applied to riscv-to-apply.next FYI, I'm hoping to pull this patch into the upcoming 5.0.1 stable release. The freeze is scheduled for 2020-09-20, I will apply this if it hits master before then. > > Alistair > > > --- > > > > Please let me know if the tags are wrong or inappropriate, this is my > > first QEMU patch. > > > > hw/riscv/sifive_test.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/hw/riscv/sifive_test.c b/hw/riscv/sifive_test.c > > index 0c78fb2c93..8c70dd69df 100644 > > --- a/hw/riscv/sifive_test.c > > +++ b/hw/riscv/sifive_test.c > > @@ -59,7 +59,7 @@ static const MemoryRegionOps sifive_test_ops = { > > .write = sifive_test_write, > > .endianness = DEVICE_NATIVE_ENDIAN, > > .valid = { > > -.min_access_size = 4, > > +.min_access_size = 2, > > .max_access_size = 4 > > } > > }; > > > > base-commit: 2f4c51c0f384d7888a04b4815861e6d5fd244d75 > > -- > > 2.28.0 > > > > >
Re: [PATCH] riscv: sifive_test: Allow 16-bit writes to memory region
On Mon, Aug 31, 2020 at 10:59 PM Nathan Chancellor wrote: > > When shutting down the machine running a mainline Linux kernel, the > following error happens: > > $ build/riscv64-softmmu/qemu-system-riscv64 -bios default -M virt \ > -display none -initrd rootfs.cpio -kernel Image -m 512m \ > -nodefaults -serial mon:stdio > ... > Requesting system poweroff > [4.999630] reboot: Power down > sbi_trap_error: hart0: trap handler failed (error -2) > sbi_trap_error: hart0: mcause=0x0007 mtval=0x0010 > sbi_trap_error: hart0: mepc=0x8000d4cc mstatus=0x1822 > sbi_trap_error: hart0: ra=0x8000999e sp=0x80015c78 > sbi_trap_error: hart0: gp=0xffe000e76610 tp=0xffe0081b89c0 > sbi_trap_error: hart0: s0=0x80015c88 s1=0x0040 > sbi_trap_error: hart0: a0=0x a1=0x80004024 > sbi_trap_error: hart0: a2=0x80004024 a3=0x80004024 > sbi_trap_error: hart0: a4=0x0010 a5=0x > sbi_trap_error: hart0: a6=0x4024 a7=0x80011158 > sbi_trap_error: hart0: s2=0x s3=0x80016000 > sbi_trap_error: hart0: s4=0x s5=0x > sbi_trap_error: hart0: s6=0x0001 s7=0x > sbi_trap_error: hart0: s8=0x s9=0x > sbi_trap_error: hart0: s10=0x s11=0x0008 > sbi_trap_error: hart0: t0=0x t1=0x > sbi_trap_error: hart0: t2=0x t3=0x > sbi_trap_error: hart0: t4=0x t5=0x > sbi_trap_error: hart0: t6=0x > > The kernel does a 16-bit write when powering off the machine, which > was allowed before commit 5d971f9e67 ("memory: Revert "memory: accept > mismatching sizes in memory_region_access_valid""). Make min_access_size > match reality so that the machine can shut down properly now. > > Cc: qemu-sta...@nongnu.org > Fixes: 88a07990fa ("SiFive RISC-V Test Finisher") > Fixes: 5d971f9e67 ("memory: Revert "memory: accept mismatching sizes in > memory_region_access_valid"") > Signed-off-by: Nathan Chancellor Thanks! Applied to riscv-to-apply.next Alistair > --- > > Please let me know if the tags are wrong or inappropriate, this is my > first QEMU patch. > > hw/riscv/sifive_test.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/riscv/sifive_test.c b/hw/riscv/sifive_test.c > index 0c78fb2c93..8c70dd69df 100644 > --- a/hw/riscv/sifive_test.c > +++ b/hw/riscv/sifive_test.c > @@ -59,7 +59,7 @@ static const MemoryRegionOps sifive_test_ops = { > .write = sifive_test_write, > .endianness = DEVICE_NATIVE_ENDIAN, > .valid = { > -.min_access_size = 4, > +.min_access_size = 2, > .max_access_size = 4 > } > }; > > base-commit: 2f4c51c0f384d7888a04b4815861e6d5fd244d75 > -- > 2.28.0 > >
Re: [PATCH] riscv: sifive_test: Allow 16-bit writes to memory region
On Mon, Aug 31, 2020 at 10:59 PM Nathan Chancellor wrote: > > When shutting down the machine running a mainline Linux kernel, the > following error happens: > > $ build/riscv64-softmmu/qemu-system-riscv64 -bios default -M virt \ > -display none -initrd rootfs.cpio -kernel Image -m 512m \ > -nodefaults -serial mon:stdio > ... > Requesting system poweroff > [4.999630] reboot: Power down > sbi_trap_error: hart0: trap handler failed (error -2) > sbi_trap_error: hart0: mcause=0x0007 mtval=0x0010 > sbi_trap_error: hart0: mepc=0x8000d4cc mstatus=0x1822 > sbi_trap_error: hart0: ra=0x8000999e sp=0x80015c78 > sbi_trap_error: hart0: gp=0xffe000e76610 tp=0xffe0081b89c0 > sbi_trap_error: hart0: s0=0x80015c88 s1=0x0040 > sbi_trap_error: hart0: a0=0x a1=0x80004024 > sbi_trap_error: hart0: a2=0x80004024 a3=0x80004024 > sbi_trap_error: hart0: a4=0x0010 a5=0x > sbi_trap_error: hart0: a6=0x4024 a7=0x80011158 > sbi_trap_error: hart0: s2=0x s3=0x80016000 > sbi_trap_error: hart0: s4=0x s5=0x > sbi_trap_error: hart0: s6=0x0001 s7=0x > sbi_trap_error: hart0: s8=0x s9=0x > sbi_trap_error: hart0: s10=0x s11=0x0008 > sbi_trap_error: hart0: t0=0x t1=0x > sbi_trap_error: hart0: t2=0x t3=0x > sbi_trap_error: hart0: t4=0x t5=0x > sbi_trap_error: hart0: t6=0x > > The kernel does a 16-bit write when powering off the machine, which > was allowed before commit 5d971f9e67 ("memory: Revert "memory: accept > mismatching sizes in memory_region_access_valid""). Make min_access_size > match reality so that the machine can shut down properly now. > > Cc: qemu-sta...@nongnu.org > Fixes: 88a07990fa ("SiFive RISC-V Test Finisher") > Fixes: 5d971f9e67 ("memory: Revert "memory: accept mismatching sizes in > memory_region_access_valid"") > Signed-off-by: Nathan Chancellor Reviewed-by: Alistair Francis Alistair > --- > > Please let me know if the tags are wrong or inappropriate, this is my > first QEMU patch. > > hw/riscv/sifive_test.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/riscv/sifive_test.c b/hw/riscv/sifive_test.c > index 0c78fb2c93..8c70dd69df 100644 > --- a/hw/riscv/sifive_test.c > +++ b/hw/riscv/sifive_test.c > @@ -59,7 +59,7 @@ static const MemoryRegionOps sifive_test_ops = { > .write = sifive_test_write, > .endianness = DEVICE_NATIVE_ENDIAN, > .valid = { > -.min_access_size = 4, > +.min_access_size = 2, > .max_access_size = 4 > } > }; > > base-commit: 2f4c51c0f384d7888a04b4815861e6d5fd244d75 > -- > 2.28.0 > >
Re: [PATCH] riscv: sifive_test: Allow 16-bit writes to memory region
On Tue, Sep 01, 2020 at 07:08:34AM -0400, Michael S. Tsirkin wrote: > On Mon, Aug 31, 2020 at 10:58:23PM -0700, Nathan Chancellor wrote: > > When shutting down the machine running a mainline Linux kernel, the > > following error happens: > > > > $ build/riscv64-softmmu/qemu-system-riscv64 -bios default -M virt \ > > -display none -initrd rootfs.cpio -kernel Image -m 512m \ > > -nodefaults -serial mon:stdio > > ... > > Requesting system poweroff > > [4.999630] reboot: Power down > > sbi_trap_error: hart0: trap handler failed (error -2) > > sbi_trap_error: hart0: mcause=0x0007 mtval=0x0010 > > sbi_trap_error: hart0: mepc=0x8000d4cc mstatus=0x1822 > > sbi_trap_error: hart0: ra=0x8000999e sp=0x80015c78 > > sbi_trap_error: hart0: gp=0xffe000e76610 tp=0xffe0081b89c0 > > sbi_trap_error: hart0: s0=0x80015c88 s1=0x0040 > > sbi_trap_error: hart0: a0=0x a1=0x80004024 > > sbi_trap_error: hart0: a2=0x80004024 a3=0x80004024 > > sbi_trap_error: hart0: a4=0x0010 a5=0x > > sbi_trap_error: hart0: a6=0x4024 a7=0x80011158 > > sbi_trap_error: hart0: s2=0x s3=0x80016000 > > sbi_trap_error: hart0: s4=0x s5=0x > > sbi_trap_error: hart0: s6=0x0001 s7=0x > > sbi_trap_error: hart0: s8=0x s9=0x > > sbi_trap_error: hart0: s10=0x s11=0x0008 > > sbi_trap_error: hart0: t0=0x t1=0x > > sbi_trap_error: hart0: t2=0x t3=0x > > sbi_trap_error: hart0: t4=0x t5=0x > > sbi_trap_error: hart0: t6=0x > > > > The kernel does a 16-bit write when powering off the machine, which > > was allowed before commit 5d971f9e67 ("memory: Revert "memory: accept > > mismatching sizes in memory_region_access_valid""). Make min_access_size > > match reality so that the machine can shut down properly now. > > > > Cc: qemu-sta...@nongnu.org > > Fixes: 88a07990fa ("SiFive RISC-V Test Finisher") > > Fixes: 5d971f9e67 ("memory: Revert "memory: accept mismatching sizes in > > memory_region_access_valid"") > > Signed-off-by: Nathan Chancellor > > --- > > > > Please let me know if the tags are wrong or inappropriate, this is my > > first QEMU patch. > > > > hw/riscv/sifive_test.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/hw/riscv/sifive_test.c b/hw/riscv/sifive_test.c > > index 0c78fb2c93..8c70dd69df 100644 > > --- a/hw/riscv/sifive_test.c > > +++ b/hw/riscv/sifive_test.c > > @@ -59,7 +59,7 @@ static const MemoryRegionOps sifive_test_ops = { > > .write = sifive_test_write, > > .endianness = DEVICE_NATIVE_ENDIAN, > > > Apropos, why is this native endian? > The write handler seems to ignore target endian-ness, > looks like a bug ... patch itself looks fine Acked-by: Michael S. Tsirkin > > .valid = { > > -.min_access_size = 4, > > +.min_access_size = 2, > > .max_access_size = 4 > > } > > }; > > > > base-commit: 2f4c51c0f384d7888a04b4815861e6d5fd244d75 > > -- > > 2.28.0
Re: [PATCH] riscv: sifive_test: Allow 16-bit writes to memory region
On Mon, Aug 31, 2020 at 10:58:23PM -0700, Nathan Chancellor wrote: > When shutting down the machine running a mainline Linux kernel, the > following error happens: > > $ build/riscv64-softmmu/qemu-system-riscv64 -bios default -M virt \ > -display none -initrd rootfs.cpio -kernel Image -m 512m \ > -nodefaults -serial mon:stdio > ... > Requesting system poweroff > [4.999630] reboot: Power down > sbi_trap_error: hart0: trap handler failed (error -2) > sbi_trap_error: hart0: mcause=0x0007 mtval=0x0010 > sbi_trap_error: hart0: mepc=0x8000d4cc mstatus=0x1822 > sbi_trap_error: hart0: ra=0x8000999e sp=0x80015c78 > sbi_trap_error: hart0: gp=0xffe000e76610 tp=0xffe0081b89c0 > sbi_trap_error: hart0: s0=0x80015c88 s1=0x0040 > sbi_trap_error: hart0: a0=0x a1=0x80004024 > sbi_trap_error: hart0: a2=0x80004024 a3=0x80004024 > sbi_trap_error: hart0: a4=0x0010 a5=0x > sbi_trap_error: hart0: a6=0x4024 a7=0x80011158 > sbi_trap_error: hart0: s2=0x s3=0x80016000 > sbi_trap_error: hart0: s4=0x s5=0x > sbi_trap_error: hart0: s6=0x0001 s7=0x > sbi_trap_error: hart0: s8=0x s9=0x > sbi_trap_error: hart0: s10=0x s11=0x0008 > sbi_trap_error: hart0: t0=0x t1=0x > sbi_trap_error: hart0: t2=0x t3=0x > sbi_trap_error: hart0: t4=0x t5=0x > sbi_trap_error: hart0: t6=0x > > The kernel does a 16-bit write when powering off the machine, which > was allowed before commit 5d971f9e67 ("memory: Revert "memory: accept > mismatching sizes in memory_region_access_valid""). Make min_access_size > match reality so that the machine can shut down properly now. > > Cc: qemu-sta...@nongnu.org > Fixes: 88a07990fa ("SiFive RISC-V Test Finisher") > Fixes: 5d971f9e67 ("memory: Revert "memory: accept mismatching sizes in > memory_region_access_valid"") > Signed-off-by: Nathan Chancellor > --- > > Please let me know if the tags are wrong or inappropriate, this is my > first QEMU patch. > > hw/riscv/sifive_test.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/hw/riscv/sifive_test.c b/hw/riscv/sifive_test.c > index 0c78fb2c93..8c70dd69df 100644 > --- a/hw/riscv/sifive_test.c > +++ b/hw/riscv/sifive_test.c > @@ -59,7 +59,7 @@ static const MemoryRegionOps sifive_test_ops = { > .write = sifive_test_write, > .endianness = DEVICE_NATIVE_ENDIAN, Apropos, why is this native endian? The write handler seems to ignore target endian-ness, looks like a bug ... > .valid = { > -.min_access_size = 4, > +.min_access_size = 2, > .max_access_size = 4 > } > }; > > base-commit: 2f4c51c0f384d7888a04b4815861e6d5fd244d75 > -- > 2.28.0