[PATCH 0/2] riscv: Fix the console of the Spike machine on big endian hosts

[PATCH 0/2] riscv: Fix the console of the Spike machine on big endian hosts

[Thomas Huth]

The tests/avocado/riscv_opensbi.py avocado test is currently failing
on big endian hosts since the console of the Spike machine is not
working there. With two small patches, this can be fixed: First patch
fixes riscv64, and the second one fixes riscv32.

Thomas Huth (2):
  hw/char/riscv_htif: Fix printing of console characters on big endian
    hosts
  hw/char/riscv_htif: Fix the console syscall on big endian hosts

 hw/char/riscv_htif.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

-- 
2.39.3

[PATCH 1/2] hw/char/riscv_htif: Fix printing of console characters on big endian hosts

[Thomas Huth]

The character that should be printed is stored in the 64 bit "payload"
variable. The code currently tries to print it by taking the address
of the variable and passing this pointer to qemu_chr_fe_write(). However,
this only works on little endian hosts where the least significant bits
are stored on the lowest address. To do this in a portable way, we have
to store the value in an uint8_t variable instead.

Fixes: 5033606780 ("RISC-V HTIF Console")
Signed-off-by: Thomas Huth <thuth@redhat.com>
---
 hw/char/riscv_htif.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/hw/char/riscv_htif.c b/hw/char/riscv_htif.c
index 37d3ccc76b..f96df40124 100644
--- a/hw/char/riscv_htif.c
+++ b/hw/char/riscv_htif.c
@@ -232,7 +232,8 @@ static void htif_handle_tohost_write(HTIFState *s, uint64_t val_written)
             s->tohost = 0; /* clear to indicate we read */
             return;
         } else if (cmd == HTIF_CONSOLE_CMD_PUTC) {
-            qemu_chr_fe_write(&s->chr, (uint8_t *)&payload, 1);
+            uint8_t ch = (uint8_t)payload;
+            qemu_chr_fe_write(&s->chr, &ch, 1);
             resp = 0x100 | (uint8_t)payload;
         } else {
             qemu_log("HTIF device %d: unknown command\n", device);
-- 
2.39.3

Re: [PATCH 1/2] hw/char/riscv_htif: Fix printing of console characters on big endian hosts

[Bin Meng]

On Fri, Jul 21, 2023 at 5:48 PM Thomas Huth <thuth@redhat.com> wrote:
>
> The character that should be printed is stored in the 64 bit "payload"
> variable. The code currently tries to print it by taking the address
> of the variable and passing this pointer to qemu_chr_fe_write(). However,
> this only works on little endian hosts where the least significant bits
> are stored on the lowest address. To do this in a portable way, we have
> to store the value in an uint8_t variable instead.
>
> Fixes: 5033606780 ("RISC-V HTIF Console")
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---
>  hw/char/riscv_htif.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>

Reviewed-by: Bin Meng <bmeng@tinylab.org>

Re: [PATCH 1/2] hw/char/riscv_htif: Fix printing of console characters on big endian hosts

[Daniel Henrique Barboza]

On 7/21/23 06:47, Thomas Huth wrote:
> The character that should be printed is stored in the 64 bit "payload"
> variable. The code currently tries to print it by taking the address
> of the variable and passing this pointer to qemu_chr_fe_write(). However,
> this only works on little endian hosts where the least significant bits
> are stored on the lowest address. To do this in a portable way, we have
> to store the value in an uint8_t variable instead.
> 
> Fixes: 5033606780 ("RISC-V HTIF Console")
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---

Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>

>   hw/char/riscv_htif.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/hw/char/riscv_htif.c b/hw/char/riscv_htif.c
> index 37d3ccc76b..f96df40124 100644
> --- a/hw/char/riscv_htif.c
> +++ b/hw/char/riscv_htif.c
> @@ -232,7 +232,8 @@ static void htif_handle_tohost_write(HTIFState *s, uint64_t val_written)
>               s->tohost = 0; /* clear to indicate we read */
>               return;
>           } else if (cmd == HTIF_CONSOLE_CMD_PUTC) {
> -            qemu_chr_fe_write(&s->chr, (uint8_t *)&payload, 1);
> +            uint8_t ch = (uint8_t)payload;
> +            qemu_chr_fe_write(&s->chr, &ch, 1);
>               resp = 0x100 | (uint8_t)payload;
>           } else {
>               qemu_log("HTIF device %d: unknown command\n", device);

Re: [PATCH 1/2] hw/char/riscv_htif: Fix printing of console characters on big endian hosts

[Philippe Mathieu-Daudé]

On 21/7/23 11:47, Thomas Huth wrote:
> The character that should be printed is stored in the 64 bit "payload"
> variable. The code currently tries to print it by taking the address
> of the variable and passing this pointer to qemu_chr_fe_write(). However,
> this only works on little endian hosts where the least significant bits
> are stored on the lowest address. To do this in a portable way, we have
> to store the value in an uint8_t variable instead.
> 
> Fixes: 5033606780 ("RISC-V HTIF Console")
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---
>   hw/char/riscv_htif.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>

Re: [PATCH 1/2] hw/char/riscv_htif: Fix printing of console characters on big endian hosts

[Alistair Francis]

On Fri, Jul 21, 2023 at 7:48 PM Thomas Huth <thuth@redhat.com> wrote:
>
> The character that should be printed is stored in the 64 bit "payload"
> variable. The code currently tries to print it by taking the address
> of the variable and passing this pointer to qemu_chr_fe_write(). However,
> this only works on little endian hosts where the least significant bits
> are stored on the lowest address. To do this in a portable way, we have
> to store the value in an uint8_t variable instead.
>
> Fixes: 5033606780 ("RISC-V HTIF Console")
> Signed-off-by: Thomas Huth <thuth@redhat.com>

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>

Alistair

> ---
>  hw/char/riscv_htif.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/hw/char/riscv_htif.c b/hw/char/riscv_htif.c
> index 37d3ccc76b..f96df40124 100644
> --- a/hw/char/riscv_htif.c
> +++ b/hw/char/riscv_htif.c
> @@ -232,7 +232,8 @@ static void htif_handle_tohost_write(HTIFState *s, uint64_t val_written)
>              s->tohost = 0; /* clear to indicate we read */
>              return;
>          } else if (cmd == HTIF_CONSOLE_CMD_PUTC) {
> -            qemu_chr_fe_write(&s->chr, (uint8_t *)&payload, 1);
> +            uint8_t ch = (uint8_t)payload;
> +            qemu_chr_fe_write(&s->chr, &ch, 1);
>              resp = 0x100 | (uint8_t)payload;
>          } else {
>              qemu_log("HTIF device %d: unknown command\n", device);
> --
> 2.39.3
>
>

[PATCH 2/2] hw/char/riscv_htif: Fix the console syscall on big endian hosts

[Thomas Huth]

Values that have been read via cpu_physical_memory_read() from the
guest's memory have to be swapped in case the host endianess differs
from the guest.

Fixes: a6e13e31d5 ("riscv_htif: Support console output via proxy syscall")
Signed-off-by: Thomas Huth <thuth@redhat.com>
---
 hw/char/riscv_htif.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/hw/char/riscv_htif.c b/hw/char/riscv_htif.c
index f96df40124..40de6b8b77 100644
--- a/hw/char/riscv_htif.c
+++ b/hw/char/riscv_htif.c
@@ -30,6 +30,7 @@
 #include "qemu/timer.h"
 #include "qemu/error-report.h"
 #include "exec/address-spaces.h"
+#include "exec/tswap.h"
 #include "sysemu/dma.h"
 
 #define RISCV_DEBUG_HTIF 0
@@ -209,11 +210,11 @@ static void htif_handle_tohost_write(HTIFState *s, uint64_t val_written)
             } else {
                 uint64_t syscall[8];
                 cpu_physical_memory_read(payload, syscall, sizeof(syscall));
-                if (syscall[0] == PK_SYS_WRITE &&
-                    syscall[1] == HTIF_DEV_CONSOLE &&
-                    syscall[3] == HTIF_CONSOLE_CMD_PUTC) {
+                if (tswap64(syscall[0]) == PK_SYS_WRITE &&
+                    tswap64(syscall[1]) == HTIF_DEV_CONSOLE &&
+                    tswap64(syscall[3]) == HTIF_CONSOLE_CMD_PUTC) {
                     uint8_t ch;
-                    cpu_physical_memory_read(syscall[2], &ch, 1);
+                    cpu_physical_memory_read(tswap64(syscall[2]), &ch, 1);
                     qemu_chr_fe_write(&s->chr, &ch, 1);
                     resp = 0x100 | (uint8_t)payload;
                 } else {
-- 
2.39.3

Re: [PATCH 2/2] hw/char/riscv_htif: Fix the console syscall on big endian hosts

[Bin Meng]

On Fri, Jul 21, 2023 at 5:48 PM Thomas Huth <thuth@redhat.com> wrote:
>
> Values that have been read via cpu_physical_memory_read() from the
> guest's memory have to be swapped in case the host endianess differs

typo: endianness

> from the guest.
>
> Fixes: a6e13e31d5 ("riscv_htif: Support console output via proxy syscall")
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---
>  hw/char/riscv_htif.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
>

Reviewed-by: Bin Meng <bmeng@tinylab.org>

Re: [PATCH 2/2] hw/char/riscv_htif: Fix the console syscall on big endian hosts

[Daniel Henrique Barboza]

On 7/21/23 06:47, Thomas Huth wrote:
> Values that have been read via cpu_physical_memory_read() from the
> guest's memory have to be swapped in case the host endianess differs
> from the guest.
> 
> Fixes: a6e13e31d5 ("riscv_htif: Support console output via proxy syscall")
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---

Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>

>   hw/char/riscv_htif.c | 9 +++++----
>   1 file changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/hw/char/riscv_htif.c b/hw/char/riscv_htif.c
> index f96df40124..40de6b8b77 100644
> --- a/hw/char/riscv_htif.c
> +++ b/hw/char/riscv_htif.c
> @@ -30,6 +30,7 @@
>   #include "qemu/timer.h"
>   #include "qemu/error-report.h"
>   #include "exec/address-spaces.h"
> +#include "exec/tswap.h"
>   #include "sysemu/dma.h"
>   
>   #define RISCV_DEBUG_HTIF 0
> @@ -209,11 +210,11 @@ static void htif_handle_tohost_write(HTIFState *s, uint64_t val_written)
>               } else {
>                   uint64_t syscall[8];
>                   cpu_physical_memory_read(payload, syscall, sizeof(syscall));
> -                if (syscall[0] == PK_SYS_WRITE &&
> -                    syscall[1] == HTIF_DEV_CONSOLE &&
> -                    syscall[3] == HTIF_CONSOLE_CMD_PUTC) {
> +                if (tswap64(syscall[0]) == PK_SYS_WRITE &&
> +                    tswap64(syscall[1]) == HTIF_DEV_CONSOLE &&
> +                    tswap64(syscall[3]) == HTIF_CONSOLE_CMD_PUTC) {
>                       uint8_t ch;
> -                    cpu_physical_memory_read(syscall[2], &ch, 1);
> +                    cpu_physical_memory_read(tswap64(syscall[2]), &ch, 1);
>                       qemu_chr_fe_write(&s->chr, &ch, 1);
>                       resp = 0x100 | (uint8_t)payload;
>                   } else {

Re: [PATCH 2/2] hw/char/riscv_htif: Fix the console syscall on big endian hosts

[Alistair Francis]

On Fri, Jul 21, 2023 at 7:48 PM Thomas Huth <thuth@redhat.com> wrote:
>
> Values that have been read via cpu_physical_memory_read() from the
> guest's memory have to be swapped in case the host endianess differs
> from the guest.
>
> Fixes: a6e13e31d5 ("riscv_htif: Support console output via proxy syscall")
> Signed-off-by: Thomas Huth <thuth@redhat.com>

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>

Alistair

> ---
>  hw/char/riscv_htif.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
>
> diff --git a/hw/char/riscv_htif.c b/hw/char/riscv_htif.c
> index f96df40124..40de6b8b77 100644
> --- a/hw/char/riscv_htif.c
> +++ b/hw/char/riscv_htif.c
> @@ -30,6 +30,7 @@
>  #include "qemu/timer.h"
>  #include "qemu/error-report.h"
>  #include "exec/address-spaces.h"
> +#include "exec/tswap.h"
>  #include "sysemu/dma.h"
>
>  #define RISCV_DEBUG_HTIF 0
> @@ -209,11 +210,11 @@ static void htif_handle_tohost_write(HTIFState *s, uint64_t val_written)
>              } else {
>                  uint64_t syscall[8];
>                  cpu_physical_memory_read(payload, syscall, sizeof(syscall));
> -                if (syscall[0] == PK_SYS_WRITE &&
> -                    syscall[1] == HTIF_DEV_CONSOLE &&
> -                    syscall[3] == HTIF_CONSOLE_CMD_PUTC) {
> +                if (tswap64(syscall[0]) == PK_SYS_WRITE &&
> +                    tswap64(syscall[1]) == HTIF_DEV_CONSOLE &&
> +                    tswap64(syscall[3]) == HTIF_CONSOLE_CMD_PUTC) {
>                      uint8_t ch;
> -                    cpu_physical_memory_read(syscall[2], &ch, 1);
> +                    cpu_physical_memory_read(tswap64(syscall[2]), &ch, 1);
>                      qemu_chr_fe_write(&s->chr, &ch, 1);
>                      resp = 0x100 | (uint8_t)payload;
>                  } else {
> --
> 2.39.3
>
>

Re: [PATCH 0/2] riscv: Fix the console of the Spike machine on big endian hosts

[Alistair Francis]

On Fri, Jul 21, 2023 at 7:48 PM Thomas Huth <thuth@redhat.com> wrote:
>
> The tests/avocado/riscv_opensbi.py avocado test is currently failing
> on big endian hosts since the console of the Spike machine is not
> working there. With two small patches, this can be fixed: First patch
> fixes riscv64, and the second one fixes riscv32.
>
> Thomas Huth (2):
>   hw/char/riscv_htif: Fix printing of console characters on big endian
>     hosts
>   hw/char/riscv_htif: Fix the console syscall on big endian hosts

Thanks!

Applied to riscv-to-apply.next

Alistair

>
>  hw/char/riscv_htif.c | 12 +++++++-----
>  1 file changed, 7 insertions(+), 5 deletions(-)
>
> --
> 2.39.3
>
>