Re: [PATCH 6/9] hw/arm/aspeed_ast10x0: Map HACE peripheral

[Philippe Mathieu-Daudé]

On 29/12/22 21:52, Peter Delevoryas wrote:
> On Thu, Dec 29, 2022 at 04:23:22PM +0100, Philippe Mathieu-Daudé wrote:
> > Since I don't have access to the datasheet, the relevant
> > values were found in:
> > https://github.com/AspeedTech-BMC/zephyr/blob/v00.01.08/dts/arm/aspeed/ast10x0.dtsi
> >
> > Before on Zephyr:
> >
> >    uart:~$ crypto aes256_cbc_vault
> >    aes256_cbc vault key 1
> >    [00:00:06.699,000] <inf> hace_global: aspeed_crypto_session_setup
> >    [00:00:06.699,000] <inf> hace_global: data->cmd: 1c2098
> >    [00:00:06.699,000] <inf> hace_global: crypto_data_src: 93340
> >    [00:00:06.699,000] <inf> hace_global: crypto_data_dst: 93348
> >    [00:00:06.699,000] <inf> hace_global: crypto_ctx_base: 93300
> >    [00:00:06.699,000] <inf> hace_global: crypto_data_len: 80000040
> >    [00:00:06.699,000] <inf> hace_global: crypto_cmd_reg:  11c2098
> >    [00:00:09.743,000] <inf> hace_global: HACE_STS: 0
> >    [00:00:09.743,000] <err> hace_global: HACE poll timeout
> >    [00:00:09.743,000] <err> crypto: CBC mode ENCRYPT - Failed
> >    [00:00:09.743,000] <inf> hace_global: aspeed_crypto_session_free
> >    uart:~$
> >
> > After:
> >
> >    uart:~$ crypto aes256_cbc_vault
> >    aes256_cbc vault key 1
> >    Was waiting for:
> >    6b c1 be e2 2e 40 9f 96 e9 3d 7e 11 73 93 17 2a
> >    ae 2d 8a 57 1e 03 ac 9c 9e b7 6f ac 45 af 8e 51
> >    30 c8 1c 46 a3 5c e4 11 e5 fb c1 19 1a 0a 52 ef
> >    f6 9f 24 45 df 4f 9b 17 ad 2b 41 7b e6 6c 37 10
> >
> >     But got:
> >    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> >    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> >    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> >    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> >
> >    [00:00:05.771,000] <inf> hace_global: aspeed_crypto_session_setup
> >    [00:00:05.772,000] <inf> hace_global: data->cmd: 1c2098
> >    [00:00:05.772,000] <inf> hace_global: crypto_data_src: 93340
> >    [00:00:05.772,000] <inf> hace_global: crypto_data_dst: 93348
> >    [00:00:05.772,000] <inf> hace_global: crypto_ctx_base: 93300
> >    [00:00:05.772,000] <inf> hace_global: crypto_data_len: 80000040
> >    [00:00:05.772,000] <inf> hace_global: crypto_cmd_reg:  11c2098
> >    [00:00:05.772,000] <inf> hace_global: HACE_STS: 1000
> >    [00:00:05.772,000] <inf> crypto: Output length (encryption): 80
> >    [00:00:05.772,000] <inf> hace_global: aspeed_crypto_session_free
> >    [00:00:05.772,000] <inf> hace_global: aspeed_crypto_session_setup
> >    [00:00:05.772,000] <inf> hace_global: data->cmd: 1c2018
> >    [00:00:05.772,000] <inf> hace_global: crypto_data_src: 93340
> >    [00:00:05.772,000] <inf> hace_global: crypto_data_dst: 93348
> >    [00:00:05.772,000] <inf> hace_global: crypto_ctx_base: 93300
> >    [00:00:05.772,000] <inf> hace_global: crypto_data_len: 80000040
> >    [00:00:05.772,000] <inf> hace_global: crypto_cmd_reg:  11c2018
> >    [00:00:05.772,000] <inf> hace_global: HACE_STS: 1000
> >    [00:00:05.772,000] <inf> crypto: Output length (decryption): 64
> >    [00:00:05.772,000] <err> crypto: CBC mode DECRYPT - Mismatch between 
> > plaintext and decrypted cipher text
> >    [00:00:05.774,000] <inf> hace_global: aspeed_crypto_session_free
> >    uart:~$
> >
> > Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
>
> Awesome!
>
> Reviewed-by: Peter Delevoryas <peter@pjd.dev>
>
> > ---
> > Should we rename HACE 'dram' as 'secram' / 'secure-ram'?
>
> Sure, sounds good to me.
>
> > ---
> >   hw/arm/aspeed_ast10x0.c | 15 +++++++++++++++
> >   1 file changed, 15 insertions(+)
> >
> > diff --git a/hw/arm/aspeed_ast10x0.c b/hw/arm/aspeed_ast10x0.c
> > index 21a2e62345..02636705b6 100644
> > --- a/hw/arm/aspeed_ast10x0.c
> > +++ b/hw/arm/aspeed_ast10x0.c
> > @@ -29,6 +29,7 @@ static const hwaddr aspeed_soc_ast1030_memmap[] = {
> >       [ASPEED_DEV_SPI1]      = 0x7E630000,
> >       [ASPEED_DEV_SPI2]      = 0x7E640000,
> >       [ASPEED_DEV_UDC]       = 0x7E6A2000,
> > +    [ASPEED_DEV_HACE]      = 0x7E6D0000,
> >       [ASPEED_DEV_SCU]       = 0x7E6E2000,
> >       [ASPEED_DEV_JTAG0]     = 0x7E6E4000,
> >       [ASPEED_DEV_JTAG1]     = 0x7E6E4100,
> > @@ -166,6 +167,9 @@ static void aspeed_soc_ast1030_init(Object *obj)
> >       snprintf(typename, sizeof(typename), "aspeed.gpio-%s", socname);
> >       object_initialize_child(obj, "gpio", &s->gpio, typename);
> >   
> > +    snprintf(typename, sizeof(typename), "aspeed.hace-%s", socname);
> > +    object_initialize_child(obj, "hace", &s->hace, typename);
> > +
> >       object_initialize_child(obj, "iomem", &s->iomem, 
> > TYPE_UNIMPLEMENTED_DEVICE);
> >       object_initialize_child(obj, "sbc-unimplemented", &s->sbc_unimplemented,
> >                               TYPE_UNIMPLEMENTED_DEVICE);
> > @@ -359,6 +363,17 @@ static void aspeed_soc_ast1030_realize(DeviceState 
> > *dev_soc, Error **errp)
> >       }
> >       aspeed_mmio_map(s, SYS_BUS_DEVICE(&s->sbc), 0, 
> > sc->memmap[ASPEED_DEV_SBC]);
> >   
> > +    /* HACE */
> > +    object_property_set_link(OBJECT(&s->hace), "dram", OBJECT(&s->secsram),

We need to link the SRAM here, not the sec-SRAM.

Doing so the hash test works:

uart:~$ hash test
sha256_test
tv[0]:PASS
tv[1]:PASS
tv[2]:PASS
tv[3]:PASS
tv[4]:PASS
sha384_test
tv[0]:PASS
tv[1]:PASS
tv[2]:PASS
tv[3]:PASS
tv[4]:PASS
tv[5]:PASS
sha512_test
tv[0]:PASS
tv[1]:PASS
tv[2]:PASS
tv[3]:PASS
tv[4]:PASS
tv[5]:PASS
uart:~$

rsa / aes256_cbc tests still fail.