Re: [Qemu-devel] [PATCH v3 4/8] hw/i386: Improve some of the warning messages

[Markus Armbruster <armbru@redhat.com>]

Alistair Francis <alistair.francis@xilinx.com> writes:

> Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
> Suggested-by: Eduardo Habkost <ehabkost@redhat.com>

You forgot to cc: Eduardo.  Fixed.

> ---
>
>  hw/i386/acpi-build.c | 7 ++++---
>  hw/i386/pc.c         | 9 ++++-----
>  hw/i386/pc_q35.c     | 4 ++--
>  3 files changed, 10 insertions(+), 10 deletions(-)
>
> diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
> index 6b7bade183..f9efb6be41 100644
> --- a/hw/i386/acpi-build.c
> +++ b/hw/i386/acpi-build.c
> @@ -2766,7 +2766,8 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
>                       ACPI_BUILD_ALIGN_SIZE);
>          if (tables_blob->len > legacy_table_size) {
>              /* Should happen only with PCI bridges and -M pc-i440fx-2.0.  */
> -            warn_report("migration may not work.");
> +            warn_report("ACPI tables are larger than legacy_table_size");
> +            warn_report("migration may not work");

The user has no idea what legacy_table_size means, what its value might
be, or what he can do to reduce it.

Recommend

               warn_report("ACPI tables too large, migration may not work");

If the user can do something to reduce the table size, printing suitable
hints would be nice.  Printing both tables_blob->len and
legacy_table_size might also help then.

>          }
>          g_array_set_size(tables_blob, legacy_table_size);
>      } else {
> @@ -2774,9 +2775,9 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
>          if (tables_blob->len > ACPI_BUILD_TABLE_SIZE / 2) {
>              /* As of QEMU 2.1, this fires with 160 VCPUs and 255 memory slots.  */
>              warn_report("ACPI tables are larger than 64k.");

The warning text hardcodes the value of ACPI_BUILD_TABLE_SIZE / 2.  Not
nice.  Clean up while there?

> -            warn_report("migration may not work.");
> +            warn_report("migration may not work");
>              warn_report("please remove CPUs, NUMA nodes, "
> -                        "memory slots or PCI bridges.");
> +                        "memory slots or PCI bridges");

Aha, here's what the user can do.

What about:

               warn_report("ACPI tables are large, migration may not work");
               error_printf("Try removing CPUs, NUMA nodes, memory slots"
                            " or PCI bridges.");

If we want to show actual size and limit, then this might do instead:

               warn_report("ACPI table size %u exceeds %d bytes,"
                           " migration may not work",
                           tables_blob->len, ACPI_BUILD_TABLE_SIZE / 2);
               error_printf("Try removing CPUs, NUMA nodes, memory slots"
                            " or PCI bridges.");

>          }
>          acpi_align_size(tables_blob, ACPI_BUILD_TABLE_SIZE);
>      }
> diff --git a/hw/i386/pc.c b/hw/i386/pc.c
> index 465e91cc5b..084ca796c2 100644
> --- a/hw/i386/pc.c
> +++ b/hw/i386/pc.c
> @@ -383,8 +383,8 @@ ISADevice *pc_find_fdc0(void)
>      if (state.multiple) {
>          warn_report("multiple floppy disk controllers with "
>                      "iobase=0x3f0 have been found");
> -        error_printf("the one being picked for CMOS setup might not reflect "
> -                     "your intent\n");
> +        warn_report("the one being picked for CMOS setup might not reflect "
> +                    "your intent");

Please keep error_printf() here.

>      }
>  
>      return state.floppy;
> @@ -2087,9 +2087,8 @@ static void pc_machine_set_max_ram_below_4g(Object *obj, Visitor *v,
>      }
>  
>      if (value < (1ULL << 20)) {
> -        warn_report("small max_ram_below_4g(%"PRIu64
> -                    ") less than 1M.  BIOS may not work..",
> -                    value);
> +        warn_report("max_ram_below_4g (%" PRIu64 ") is less than 1M; "
> +                    "BIOS may not work.", value);

The user has no idea what max_ram_below_4g might be.  Suggest:

           warn_report("Only %" PRIu64 " bytes of RAM below the 4GiB boundary,"
                       "BIOS may not work with less than 1MiB");

>      }
>  
>      pcms->max_ram_below_4g = value;
> diff --git a/hw/i386/pc_q35.c b/hw/i386/pc_q35.c
> index 1653a47f0a..682c576cf1 100644
> --- a/hw/i386/pc_q35.c
> +++ b/hw/i386/pc_q35.c
> @@ -101,8 +101,8 @@ static void pc_q35_init(MachineState *machine)
>          lowmem = pcms->max_ram_below_4g;
>          if (machine->ram_size - lowmem > lowmem &&
>              lowmem & ((1ULL << 30) - 1)) {
> -            warn_report("Large machine and max_ram_below_4g(%"PRIu64
> -                        ") not a multiple of 1G; possible bad performance.",
> +            warn_report("Large machine and max_ram_below_4g (%"PRIu64") not a "
> +                        "multiple of 1G; possible bad performance.",

Space between string literal and PRIu64, please.

The user has no idea what max_ram_below_4g might be, or what makes the
machine "large".

>                          pcms->max_ram_below_4g);
>          }
>      }