From: Andrew Jones <ajones@ventanamicro.com>
To: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
Cc: qemu-devel@nongnu.org, qemu-riscv@nongnu.org,
alistair.francis@wdc.com, bmeng@tinylab.org,
liweiwei@iscas.ac.cn, zhiwei_liu@linux.alibaba.com,
palmer@rivosinc.com
Subject: Re: [PATCH v8 03/19] target/riscv/cpu.c: set satp_max_supported in cpu_riscv_set_satp()
Date: Thu, 2 Nov 2023 14:11:50 +0100 [thread overview]
Message-ID: <20231102-b340fdccfc0cdf9bf9fb935f@orel> (raw)
In-Reply-To: <c51fd5aa-22a9-40dd-9463-60ef23fb77c5@ventanamicro.com>
On Thu, Nov 02, 2023 at 09:53:50AM -0300, Daniel Henrique Barboza wrote:
>
>
> On 11/2/23 06:24, Andrew Jones wrote:
> > On Wed, Nov 01, 2023 at 05:41:48PM -0300, Daniel Henrique Barboza wrote:
> > > The setter() for the boolean attributes that set satp_mode (sv32, sv39,
> > > sv48, sv57, sv64) considers that the CPU will always do a
> > > set_satp_mode_max_supported() during cpu_init().
> > >
> > > This is not the case for the KVM 'host' CPU, and we'll add another CPU
> > > that won't set satp_mode_max() during cpu_init(). Users should be able
> > > to set a max_support in these circunstances.
> > >
> > > Allow cpu_riscv_set_satp() to set satp_mode_max_supported if the CPU
> > > didn't set one prior.
> > >
> > > Signed-off-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
> > > ---
> > > target/riscv/cpu.c | 11 +++++++++++
> > > 1 file changed, 11 insertions(+)
> > >
> > > diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> > > index 822970345c..9f6837ecb7 100644
> > > --- a/target/riscv/cpu.c
> > > +++ b/target/riscv/cpu.c
> > > @@ -1100,6 +1100,7 @@ static void cpu_riscv_get_satp(Object *obj, Visitor *v, const char *name,
> > > static void cpu_riscv_set_satp(Object *obj, Visitor *v, const char *name,
> > > void *opaque, Error **errp)
> > > {
> > > + RISCVCPU *cpu = RISCV_CPU(obj);
> > > RISCVSATPMap *satp_map = opaque;
> > > uint8_t satp = satp_mode_from_str(name);
> > > bool value;
> > > @@ -1108,6 +1109,16 @@ static void cpu_riscv_set_satp(Object *obj, Visitor *v, const char *name,
> > > return;
> > > }
> > > + /*
> > > + * Allow users to set satp max supported if the CPU didn't
> > > + * set any during cpu_init(). First value set to 'true'
> > > + * in this case is assumed to be the max supported for
> > > + * the CPU.
> >
> > Hmm, doesn't that mean if a user does
> >
> > -cpu rv64,sv39=true,sv48=true
> >
> > then the max is set to sv39 instead of sv48?
> >
> > I made a mistake in my last review by stating we shouldn't set the max
> > supported satp for rv64i to the maximum satp that TCG supports. I forgot
> > how all of it worked. Setting the _supported_ modes to the maximum that
> > TCG supports makes sense as long as we don't default to any of them for
> > rv64i. So, I think we should return the set_satp_mode_max_supported() to
> > rv64i's definition (passing VM_1_10_SV57 or maybe even VM_1_10_SV64) and
> > then change set_satp_mode_default_map() to error out for rv64i (or maybe
> > for all "bare" type cpus).
>
> Ok, then let's set max supported mode to SV64 (the maximum we allow).
>
> Then we don't need to do anything else - setting a max supported value will allow
> TCG to handle it accordingly with OpenSBI and the kernel, and a suitable satp_mode
> will be picked by them. We can leave finalize() untouched.
>
> I'll make a note in cpu_init() to remind ourselves that we're not setting a default
> satp_mode value, but a *limit* for the satp_mode value the CPU can handle.
It's both. It's the limit and, in the absence of user input, its used as
the default. Setting the limit is fine for a no-default cpu type, but
allowing it to become the default is not. We need to also modify
set_satp_mode_default_map() to only do what it does for non no-default
cpus types.
Thanks,
drew
next prev parent reply other threads:[~2023-11-02 13:12 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-11-01 20:41 [PATCH v8 00/19] rv64i CPU, RVA22U64 profile support Daniel Henrique Barboza
2023-11-01 20:41 ` [PATCH v8 01/19] target/riscv: create TYPE_RISCV_VENDOR_CPU Daniel Henrique Barboza
2023-11-02 2:32 ` Alistair Francis
2023-11-01 20:41 ` [PATCH v8 02/19] target/riscv/tcg: do not use "!generic" CPU checks Daniel Henrique Barboza
2023-11-02 2:38 ` Alistair Francis
2023-11-01 20:41 ` [PATCH v8 03/19] target/riscv/cpu.c: set satp_max_supported in cpu_riscv_set_satp() Daniel Henrique Barboza
2023-11-02 9:24 ` Andrew Jones
2023-11-02 12:53 ` Daniel Henrique Barboza
2023-11-02 13:11 ` Andrew Jones [this message]
2023-11-01 20:41 ` [PATCH v8 04/19] target/riscv/cpu.c: set satp_mode_max MBARE during satp_finalize() Daniel Henrique Barboza
2023-11-02 9:32 ` Andrew Jones
2023-11-01 20:41 ` [PATCH v8 05/19] target/riscv/tcg: update priv_ver on user_set extensions Daniel Henrique Barboza
2023-11-02 9:47 ` Andrew Jones
2023-11-02 13:42 ` Daniel Henrique Barboza
2023-11-02 13:48 ` Andrew Jones
2023-11-01 20:41 ` [PATCH v8 06/19] target/riscv: add rv64i CPU Daniel Henrique Barboza
2023-11-02 9:59 ` Andrew Jones
2023-11-02 14:23 ` Daniel Henrique Barboza
2023-11-01 20:41 ` [PATCH v8 07/19] target/riscv: add zicbop extension flag Daniel Henrique Barboza
2023-11-01 20:41 ` [PATCH v8 08/19] target/riscv/tcg: add 'zic64b' support Daniel Henrique Barboza
2023-11-01 20:41 ` [PATCH v8 09/19] riscv-qmp-cmds.c: expose named features in cpu_model_expansion Daniel Henrique Barboza
2023-11-01 20:41 ` [PATCH v8 10/19] target/riscv: add rva22u64 profile definition Daniel Henrique Barboza
2023-11-01 20:41 ` [PATCH v8 11/19] target/riscv/kvm: add 'rva22u64' flag as unavailable Daniel Henrique Barboza
2023-11-01 20:41 ` [PATCH v8 12/19] target/riscv/tcg: add user flag for profile support Daniel Henrique Barboza
2023-11-01 20:41 ` [PATCH v8 13/19] target/riscv/tcg: add MISA user options hash Daniel Henrique Barboza
2023-11-01 20:41 ` [PATCH v8 14/19] target/riscv/tcg: add riscv_cpu_write_misa_bit() Daniel Henrique Barboza
2023-11-01 20:42 ` [PATCH v8 15/19] target/riscv/tcg: handle profile MISA bits Daniel Henrique Barboza
2023-11-01 20:42 ` [PATCH v8 16/19] target/riscv/tcg: add hash table insert helpers Daniel Henrique Barboza
2023-11-01 20:42 ` [PATCH v8 17/19] target/riscv/tcg: honor user choice for G MISA bits Daniel Henrique Barboza
2023-11-01 20:42 ` [PATCH v8 18/19] target/riscv/tcg: validate profiles during finalize Daniel Henrique Barboza
2023-11-01 20:42 ` [PATCH v8 19/19] riscv-qmp-cmds.c: add profile flags in cpu-model-expansion Daniel Henrique Barboza
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20231102-b340fdccfc0cdf9bf9fb935f@orel \
--to=ajones@ventanamicro.com \
--cc=alistair.francis@wdc.com \
--cc=bmeng@tinylab.org \
--cc=dbarboza@ventanamicro.com \
--cc=liweiwei@iscas.ac.cn \
--cc=palmer@rivosinc.com \
--cc=qemu-devel@nongnu.org \
--cc=qemu-riscv@nongnu.org \
--cc=zhiwei_liu@linux.alibaba.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).