Re: [PATCH] aspeed: Add an AST2600 eval board
On 22/10/2019 19:20, Peter Maydell wrote: > On Tue, 22 Oct 2019 at 18:16, Cédric Le Goater wrote: >> Is it time for me to install an OSX system for dev and tests or is there >> a way to reproduce the issue ? no errors spotted in valgrind. > > I dunno. You could have a look to see if you can repro > it with the travis builds. I would investigate but I > don't have time this week given impending softfreeze > and KVM Forum next week. Sure. I will reinstall a Mac I have under OSX. C.
Re: [PATCH] aspeed: Add an AST2600 eval board
On 22/10/2019 18:43, Peter Maydell wrote: > On Thu, 17 Oct 2019 at 14:54, Peter Maydell wrote: >> >> On Thu, 17 Oct 2019 at 07:33, Joel Stanley wrote: >>> >>> On Wed, 16 Oct 2019 at 09:08, Cédric Le Goater wrote: Define the board with 1 GiB of RAM but some boards can have up to 2 GiB. Signed-off-by: Cédric Le Goater Reviewed-by: Joel Stanley --- Changes since AST2600 patchset: - lowered the RAM size to 1 GiB as it was breaking the tests on some hosts. >>> >>> Peter, >>> >>> After chatting with Cédric I agree we should merge this patch. >>> >>> As it turns out the EVBs have differing amounts of RAM; his has 1GB >>> while mine has 2GB. So we are not being inaccurate by setting 1GB as >>> the default here. >> >> That's convenient, means we don't have to figure out how to >> special-case the test infrastructure for it :-) > > This is now OK on the 32-bit boxes, but still fails 'make check' > on my OSX system: > > manooth$ QTEST_QEMU_BINARY=aarch64-softmmu/qemu-system-aarch64 > tests/qom-test -p /aarch64/qom/ast2600-evb > /aarch64/qom/ast2600-evb: Broken pipe > /Users/pm215/src/qemu-for-merges/tests/libqtest.c:149: kill_qemu() > detected QEMU death from signal 6 (Abort trap: 6) > Abort trap: 6 > > Dropping from the pullreq again :-( It will be back ! Is it time for me to install an OSX system for dev and tests or is there a way to reproduce the issue ? no errors spotted in valgrind. Thanks, C.
Re: [PATCH] aspeed: Add an AST2600 eval board
On Tue, 22 Oct 2019 at 18:16, Cédric Le Goater wrote: > Is it time for me to install an OSX system for dev and tests or is there > a way to reproduce the issue ? no errors spotted in valgrind. I dunno. You could have a look to see if you can repro it with the travis builds. I would investigate but I don't have time this week given impending softfreeze and KVM Forum next week. thanks -- PMM
Re: [PATCH] aspeed: Add an AST2600 eval board
On Thu, 17 Oct 2019 at 14:54, Peter Maydell wrote: > > On Thu, 17 Oct 2019 at 07:33, Joel Stanley wrote: > > > > On Wed, 16 Oct 2019 at 09:08, Cédric Le Goater wrote: > > > > > > Define the board with 1 GiB of RAM but some boards can have up to 2 > > > GiB. > > > > > > Signed-off-by: Cédric Le Goater > > > Reviewed-by: Joel Stanley > > > --- > > > > > > Changes since AST2600 patchset: > > > > > > - lowered the RAM size to 1 GiB as it was breaking the tests on some > > >hosts. > > > > Peter, > > > > After chatting with Cédric I agree we should merge this patch. > > > > As it turns out the EVBs have differing amounts of RAM; his has 1GB > > while mine has 2GB. So we are not being inaccurate by setting 1GB as > > the default here. > > That's convenient, means we don't have to figure out how to > special-case the test infrastructure for it :-) This is now OK on the 32-bit boxes, but still fails 'make check' on my OSX system: manooth$ QTEST_QEMU_BINARY=aarch64-softmmu/qemu-system-aarch64 tests/qom-test -p /aarch64/qom/ast2600-evb /aarch64/qom/ast2600-evb: Broken pipe /Users/pm215/src/qemu-for-merges/tests/libqtest.c:149: kill_qemu() detected QEMU death from signal 6 (Abort trap: 6) Abort trap: 6 Dropping from the pullreq again :-( thanks -- PMM
Re: [PATCH] aspeed: Add an AST2600 eval board
On Thu, 17 Oct 2019 at 07:33, Joel Stanley wrote: > > On Wed, 16 Oct 2019 at 09:08, Cédric Le Goater wrote: > > > > Define the board with 1 GiB of RAM but some boards can have up to 2 > > GiB. > > > > Signed-off-by: Cédric Le Goater > > Reviewed-by: Joel Stanley > > --- > > > > Changes since AST2600 patchset: > > > > - lowered the RAM size to 1 GiB as it was breaking the tests on some > >hosts. > > Peter, > > After chatting with Cédric I agree we should merge this patch. > > As it turns out the EVBs have differing amounts of RAM; his has 1GB > while mine has 2GB. So we are not being inaccurate by setting 1GB as > the default here. That's convenient, means we don't have to figure out how to special-case the test infrastructure for it :-) Applied to target-arm.next, thanks. -- PMM
Re: [PATCH] aspeed: Add an AST2600 eval board
On Wed, 16 Oct 2019 at 09:08, Cédric Le Goater wrote:
>
> Define the board with 1 GiB of RAM but some boards can have up to 2
> GiB.
>
> Signed-off-by: Cédric Le Goater
> Reviewed-by: Joel Stanley
> ---
>
> Changes since AST2600 patchset:
>
> - lowered the RAM size to 1 GiB as it was breaking the tests on some
>hosts.
Peter,
After chatting with Cédric I agree we should merge this patch.
As it turns out the EVBs have differing amounts of RAM; his has 1GB
while mine has 2GB. So we are not being inaccurate by setting 1GB as
the default here.
Cheers,
Joel
>
> include/hw/arm/aspeed.h | 1 +
> hw/arm/aspeed.c | 23 +++
> 2 files changed, 24 insertions(+)
>
> diff --git a/include/hw/arm/aspeed.h b/include/hw/arm/aspeed.h
> index 02073a6b4d61..f49bc7081e4d 100644
> --- a/include/hw/arm/aspeed.h
> +++ b/include/hw/arm/aspeed.h
> @@ -18,6 +18,7 @@ typedef struct AspeedBoardConfig {
> const char *desc;
> const char *soc_name;
> uint32_t hw_strap1;
> +uint32_t hw_strap2;
> const char *fmc_model;
> const char *spi_model;
> uint32_t num_cs;
> diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c
> index 52993f84b461..028191ff36fc 100644
> --- a/hw/arm/aspeed.c
> +++ b/hw/arm/aspeed.c
> @@ -88,6 +88,10 @@ struct AspeedBoardState {
> /* Witherspoon hardware value: 0xF10AD216 (but use romulus definition) */
> #define WITHERSPOON_BMC_HW_STRAP1 ROMULUS_BMC_HW_STRAP1
>
> +/* AST2600 evb hardware value */
> +#define AST2600_EVB_HW_STRAP1 0x00C0
> +#define AST2600_EVB_HW_STRAP2 0x0003
> +
> /*
> * The max ram region is for firmwares that scan the address space
> * with load/store to guess how much RAM the SoC has.
> @@ -187,6 +191,8 @@ static void aspeed_board_init(MachineState *machine,
> _abort);
> object_property_set_int(OBJECT(>soc), cfg->hw_strap1, "hw-strap1",
> _abort);
> +object_property_set_int(OBJECT(>soc), cfg->hw_strap2, "hw-strap2",
> +_abort);
> object_property_set_int(OBJECT(>soc), cfg->num_cs, "num-cs",
> _abort);
> object_property_set_int(OBJECT(>soc), machine->smp.cpus, "num-cpus",
> @@ -308,6 +314,12 @@ static void ast2500_evb_i2c_init(AspeedBoardState *bmc)
> i2c_create_slave(aspeed_i2c_get_bus(DEVICE(>i2c), 11), "ds1338",
> 0x32);
> }
>
> +static void ast2600_evb_i2c_init(AspeedBoardState *bmc)
> +{
> +/* Start with some devices on our I2C busses */
> +ast2500_evb_i2c_init(bmc);
> +}
> +
> static void romulus_bmc_i2c_init(AspeedBoardState *bmc)
> {
> AspeedSoCState *soc = >soc;
> @@ -455,6 +467,17 @@ static const AspeedBoardConfig aspeed_boards[] = {
> .num_cs= 2,
> .i2c_init = witherspoon_bmc_i2c_init,
> .ram = 512 * MiB,
> +}, {
> +.name = MACHINE_TYPE_NAME("ast2600-evb"),
> +.desc = "Aspeed AST2600 EVB (Cortex A7)",
> +.soc_name = "ast2600-a0",
> +.hw_strap1 = AST2600_EVB_HW_STRAP1,
> +.hw_strap2 = AST2600_EVB_HW_STRAP2,
> +.fmc_model = "w25q512jv",
> +.spi_model = "mx66u51235f",
> +.num_cs= 1,
> +.i2c_init = ast2600_evb_i2c_init,
> +.ram = 1 * GiB,
> },
> };
>
> --
> 2.21.0
>
