From: Igor Mammedov <imammedo@redhat.com>
To: "Philippe Mathieu-Daudé" <philmd@redhat.com>
Cc: "Peter Maydell" <peter.maydell@linaro.org>,
"Joaquin de Andres" <me@xcancerberox.com.ar>,
"Alistair Francis" <alistair@alistair23.me>,
"Philippe Mathieu-Daudé" <f4bug@amsat.org>,
"Andrew Baumann" <Andrew.Baumann@microsoft.com>,
qemu-devel@nongnu.org, "Esteban Bosse" <estebanbosse@gmail.com>,
"Niek Linnenbank" <nieklinnenbank@gmail.com>,
qemu-arm@nongnu.org, "BALATON Zoltan" <balaton@eik.bme.hu>
Subject: Re: [PATCH v3 08/13] hw/arm/raspi: Make board_rev a field of RaspiMachineClass
Date: Mon, 10 Feb 2020 14:09:18 +0100 [thread overview]
Message-ID: <20200210140918.270c8a5d@redhat.com> (raw)
In-Reply-To: <2adc51a2-b526-7f6a-4c01-775f3ac06444@redhat.com>
On Mon, 10 Feb 2020 11:03:40 +0100
Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
> On 2/10/20 10:50 AM, Igor Mammedov wrote:
> > On Sat, 8 Feb 2020 17:56:40 +0100
> > Philippe Mathieu-Daudé <f4bug@amsat.org> wrote:
> >
> >> We want to have a common class_init(). The only value that
> >> matters (and changes) is the board revision.
> >> Pass the board_rev as class_data to class_init().
> >>
> >> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> >> ---
> >> hw/arm/raspi.c | 17 ++++++++++++++---
> >> 1 file changed, 14 insertions(+), 3 deletions(-)
> >>
> >> diff --git a/hw/arm/raspi.c b/hw/arm/raspi.c
> >> index 62b8df3c2e..fbfcd29732 100644
> >> --- a/hw/arm/raspi.c
> >> +++ b/hw/arm/raspi.c
> >> @@ -46,6 +46,7 @@ typedef struct RaspiMachineClass {
> >> /*< private >*/
> >> MachineClass parent_obj;
> >> /*< public >*/
> >> + uint32_t board_rev;
> >> } RaspiMachineClass;
> >>
> >> #define TYPE_RASPI_MACHINE MACHINE_TYPE_NAME("raspi-common")
> >> @@ -227,9 +228,11 @@ static void setup_boot(MachineState *machine, int version, size_t ram_size)
> >> arm_load_kernel(ARM_CPU(first_cpu), machine, &binfo);
> >> }
> >>
> >> -static void raspi_init(MachineState *machine, uint32_t board_rev)
> >> +static void raspi_init(MachineState *machine)
> >> {
> >> + RaspiMachineClass *mc = RASPI_MACHINE_GET_CLASS(machine);
> >> RaspiMachineState *s = RASPI_MACHINE(machine);
> >> + uint32_t board_rev = mc->board_rev;
> >> int version = board_version(board_rev);
> >> uint64_t ram_size = board_ram_size(board_rev);
> >> uint32_t vcram_size;
> >> @@ -279,13 +282,16 @@ static void raspi_init(MachineState *machine, uint32_t board_rev)
> >>
> >> static void raspi2_init(MachineState *machine)
> >> {
> >> - raspi_init(machine, 0xa21041);
> >> + raspi_init(machine);
> >> }
> >>
> >> static void raspi2_machine_class_init(ObjectClass *oc, void *data)
> >> {
> >> MachineClass *mc = MACHINE_CLASS(oc);
> >> + RaspiMachineClass *rmc = RASPI_MACHINE_CLASS(oc);
> >> + uint32_t board_rev = (uint32_t)(uintptr_t)data;
> >>
> >> + rmc->board_rev = board_rev;
> >
> > instead of doing a bit obscure ".class_data = (void *)0xa21041," and
> > using it here, I'd just do
> >
> > rmc->board_rev = 0xa21041;
> >
> > using value specific for each leaf class
>
> Leaf classes are removed in patch #12 "Use a unique
> raspi_machine_class_init() method", see more uses of .class_data from v2:
> https://www.mail-archive.com/qemu-devel@nongnu.org/msg677164.html
> https://www.mail-archive.com/qemu-devel@nongnu.org/msg677166.html
>
> Are you disagreeing with them? Then we should document .class_data as
> deprecated and show example of good code.
we sometimes use .class_data when creating many derived types
(typical example CPU types (x86) ) where it's impractical to code
leaf class_init functions. I'd use .class_data in cases where I
can't get away with explicit .class_init
However in case of machines (even various versioned ones) practice
was to use parent class_init to set common data and leaf class_inits
to set board (version) specific features.
> > with this change
> > Reviewed-by: Igor Mammedov <imammedo@redhat.com>
> >
> >> mc->desc = "Raspberry Pi 2B";
> >> mc->init = raspi2_init;
> >> mc->block_default_type = IF_SD;
> >> @@ -302,13 +308,16 @@ static void raspi2_machine_class_init(ObjectClass *oc, void *data)
> >> #ifdef TARGET_AARCH64
> >> static void raspi3_init(MachineState *machine)
> >> {
> >> - raspi_init(machine, 0xa02082);
> >> + raspi_init(machine);
> >> }
> >>
> >> static void raspi3_machine_class_init(ObjectClass *oc, void *data)
> >> {
> >> MachineClass *mc = MACHINE_CLASS(oc);
> >> + RaspiMachineClass *rmc = RASPI_MACHINE_CLASS(oc);
> >> + uint32_t board_rev = (uint32_t)(uintptr_t)data;
> >>
> >> + rmc->board_rev = board_rev;
> >> mc->desc = "Raspberry Pi 3B";
> >> mc->init = raspi3_init;
> >> mc->block_default_type = IF_SD;
> >> @@ -327,11 +336,13 @@ static const TypeInfo raspi_machine_types[] = {
> >> .name = MACHINE_TYPE_NAME("raspi2"),
> >> .parent = TYPE_RASPI_MACHINE,
> >> .class_init = raspi2_machine_class_init,
> >> + .class_data = (void *)0xa21041,
> >> #ifdef TARGET_AARCH64
> >> }, {
> >> .name = MACHINE_TYPE_NAME("raspi3"),
> >> .parent = TYPE_RASPI_MACHINE,
> >> .class_init = raspi3_machine_class_init,
> >> + .class_data = (void *)0xa02082,
> >> #endif
> >> }, {
> >> .name = TYPE_RASPI_MACHINE,
> >
>
WARNING: multiple messages have this Message-ID (diff)
From: Igor Mammedov <imammedo@redhat.com>
To: "Philippe Mathieu-Daudé" <philmd@redhat.com>
Cc: "Peter Maydell" <peter.maydell@linaro.org>,
"Joaquin de Andres" <me@xcancerberox.com.ar>,
"Alistair Francis" <alistair@alistair23.me>,
"Philippe Mathieu-Daudé" <f4bug@amsat.org>,
"Andrew Baumann" <Andrew.Baumann@microsoft.com>,
qemu-devel@nongnu.org, "Esteban Bosse" <estebanbosse@gmail.com>,
"Niek Linnenbank" <nieklinnenbank@gmail.com>,
qemu-arm@nongnu.org
Subject: Re: [PATCH v3 08/13] hw/arm/raspi: Make board_rev a field of RaspiMachineClass
Date: Mon, 10 Feb 2020 14:09:18 +0100 [thread overview]
Message-ID: <20200210140918.270c8a5d@redhat.com> (raw)
In-Reply-To: <2adc51a2-b526-7f6a-4c01-775f3ac06444@redhat.com>
On Mon, 10 Feb 2020 11:03:40 +0100
Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
> On 2/10/20 10:50 AM, Igor Mammedov wrote:
> > On Sat, 8 Feb 2020 17:56:40 +0100
> > Philippe Mathieu-Daudé <f4bug@amsat.org> wrote:
> >
> >> We want to have a common class_init(). The only value that
> >> matters (and changes) is the board revision.
> >> Pass the board_rev as class_data to class_init().
> >>
> >> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> >> ---
> >> hw/arm/raspi.c | 17 ++++++++++++++---
> >> 1 file changed, 14 insertions(+), 3 deletions(-)
> >>
> >> diff --git a/hw/arm/raspi.c b/hw/arm/raspi.c
> >> index 62b8df3c2e..fbfcd29732 100644
> >> --- a/hw/arm/raspi.c
> >> +++ b/hw/arm/raspi.c
> >> @@ -46,6 +46,7 @@ typedef struct RaspiMachineClass {
> >> /*< private >*/
> >> MachineClass parent_obj;
> >> /*< public >*/
> >> + uint32_t board_rev;
> >> } RaspiMachineClass;
> >>
> >> #define TYPE_RASPI_MACHINE MACHINE_TYPE_NAME("raspi-common")
> >> @@ -227,9 +228,11 @@ static void setup_boot(MachineState *machine, int version, size_t ram_size)
> >> arm_load_kernel(ARM_CPU(first_cpu), machine, &binfo);
> >> }
> >>
> >> -static void raspi_init(MachineState *machine, uint32_t board_rev)
> >> +static void raspi_init(MachineState *machine)
> >> {
> >> + RaspiMachineClass *mc = RASPI_MACHINE_GET_CLASS(machine);
> >> RaspiMachineState *s = RASPI_MACHINE(machine);
> >> + uint32_t board_rev = mc->board_rev;
> >> int version = board_version(board_rev);
> >> uint64_t ram_size = board_ram_size(board_rev);
> >> uint32_t vcram_size;
> >> @@ -279,13 +282,16 @@ static void raspi_init(MachineState *machine, uint32_t board_rev)
> >>
> >> static void raspi2_init(MachineState *machine)
> >> {
> >> - raspi_init(machine, 0xa21041);
> >> + raspi_init(machine);
> >> }
> >>
> >> static void raspi2_machine_class_init(ObjectClass *oc, void *data)
> >> {
> >> MachineClass *mc = MACHINE_CLASS(oc);
> >> + RaspiMachineClass *rmc = RASPI_MACHINE_CLASS(oc);
> >> + uint32_t board_rev = (uint32_t)(uintptr_t)data;
> >>
> >> + rmc->board_rev = board_rev;
> >
> > instead of doing a bit obscure ".class_data = (void *)0xa21041," and
> > using it here, I'd just do
> >
> > rmc->board_rev = 0xa21041;
> >
> > using value specific for each leaf class
>
> Leaf classes are removed in patch #12 "Use a unique
> raspi_machine_class_init() method", see more uses of .class_data from v2:
> https://www.mail-archive.com/qemu-devel@nongnu.org/msg677164.html
> https://www.mail-archive.com/qemu-devel@nongnu.org/msg677166.html
>
> Are you disagreeing with them? Then we should document .class_data as
> deprecated and show example of good code.
we sometimes use .class_data when creating many derived types
(typical example CPU types (x86) ) where it's impractical to code
leaf class_init functions. I'd use .class_data in cases where I
can't get away with explicit .class_init
However in case of machines (even various versioned ones) practice
was to use parent class_init to set common data and leaf class_inits
to set board (version) specific features.
> > with this change
> > Reviewed-by: Igor Mammedov <imammedo@redhat.com>
> >
> >> mc->desc = "Raspberry Pi 2B";
> >> mc->init = raspi2_init;
> >> mc->block_default_type = IF_SD;
> >> @@ -302,13 +308,16 @@ static void raspi2_machine_class_init(ObjectClass *oc, void *data)
> >> #ifdef TARGET_AARCH64
> >> static void raspi3_init(MachineState *machine)
> >> {
> >> - raspi_init(machine, 0xa02082);
> >> + raspi_init(machine);
> >> }
> >>
> >> static void raspi3_machine_class_init(ObjectClass *oc, void *data)
> >> {
> >> MachineClass *mc = MACHINE_CLASS(oc);
> >> + RaspiMachineClass *rmc = RASPI_MACHINE_CLASS(oc);
> >> + uint32_t board_rev = (uint32_t)(uintptr_t)data;
> >>
> >> + rmc->board_rev = board_rev;
> >> mc->desc = "Raspberry Pi 3B";
> >> mc->init = raspi3_init;
> >> mc->block_default_type = IF_SD;
> >> @@ -327,11 +336,13 @@ static const TypeInfo raspi_machine_types[] = {
> >> .name = MACHINE_TYPE_NAME("raspi2"),
> >> .parent = TYPE_RASPI_MACHINE,
> >> .class_init = raspi2_machine_class_init,
> >> + .class_data = (void *)0xa21041,
> >> #ifdef TARGET_AARCH64
> >> }, {
> >> .name = MACHINE_TYPE_NAME("raspi3"),
> >> .parent = TYPE_RASPI_MACHINE,
> >> .class_init = raspi3_machine_class_init,
> >> + .class_data = (void *)0xa02082,
> >> #endif
> >> }, {
> >> .name = TYPE_RASPI_MACHINE,
> >
>
next prev parent reply other threads:[~2020-02-10 13:09 UTC|newest]
Thread overview: 55+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-02-08 16:56 [PATCH v3 00/13] hw/arm/raspi: Dynamically create machines based on the board revision Philippe Mathieu-Daudé
2020-02-08 16:56 ` Philippe Mathieu-Daudé
2020-02-08 16:56 ` [PATCH v3 01/13] hw/arm/raspi: Use BCM2708 machine type with pre Device Tree kernels Philippe Mathieu-Daudé
2020-02-08 16:56 ` Philippe Mathieu-Daudé
2020-02-09 22:53 ` Niek Linnenbank
2020-02-08 16:56 ` [PATCH v3 02/13] hw/arm/raspi: Correct the board descriptions Philippe Mathieu-Daudé
2020-02-09 22:51 ` Niek Linnenbank
2020-02-09 22:51 ` Niek Linnenbank
2020-02-09 23:02 ` Philippe Mathieu-Daudé
2020-02-09 23:02 ` Philippe Mathieu-Daudé
2020-02-08 16:56 ` [PATCH v3 03/13] hw/arm/raspi: Extract the version from the board revision Philippe Mathieu-Daudé
2020-02-13 13:40 ` Peter Maydell
2020-02-13 13:40 ` Peter Maydell
2020-02-13 13:53 ` Philippe Mathieu-Daudé
2020-02-13 13:53 ` Philippe Mathieu-Daudé
2020-02-08 16:56 ` [PATCH v3 04/13] hw/arm/raspi: Extract the RAM size " Philippe Mathieu-Daudé
2020-02-08 16:56 ` Philippe Mathieu-Daudé
2020-02-08 16:56 ` [PATCH v3 05/13] hw/arm/raspi: Extract the processor type " Philippe Mathieu-Daudé
2020-02-08 16:56 ` [PATCH v3 06/13] hw/arm/raspi: Trivial code movement Philippe Mathieu-Daudé
2020-02-08 16:56 ` Philippe Mathieu-Daudé
2020-02-08 16:56 ` Philippe Mathieu-Daudé
2020-02-10 9:58 ` Igor Mammedov
2020-02-10 9:58 ` Igor Mammedov
2020-02-08 16:56 ` [PATCH v3 07/13] hw/arm/raspi: Make machines children of abstract RaspiMachineClass Philippe Mathieu-Daudé
2020-02-08 16:56 ` Philippe Mathieu-Daudé
2020-02-10 9:45 ` Igor Mammedov
2020-02-08 16:56 ` [PATCH v3 08/13] hw/arm/raspi: Make board_rev a field of RaspiMachineClass Philippe Mathieu-Daudé
2020-02-08 16:56 ` Philippe Mathieu-Daudé
2020-02-10 9:50 ` Igor Mammedov
2020-02-10 9:50 ` Igor Mammedov
2020-02-10 10:03 ` Philippe Mathieu-Daudé
2020-02-10 10:03 ` Philippe Mathieu-Daudé
2020-02-10 13:09 ` Igor Mammedov [this message]
2020-02-10 13:09 ` Igor Mammedov
2020-02-08 16:56 ` [PATCH v3 09/13] hw/arm/raspi: Let class_init() directly call raspi_machine_init() Philippe Mathieu-Daudé
2020-02-10 9:55 ` Igor Mammedov
2020-02-10 9:55 ` Igor Mammedov
2020-02-08 16:56 ` [PATCH v3 10/13] hw/arm/raspi: Set default RAM size to size encoded in board revision Philippe Mathieu-Daudé
2020-02-08 16:56 ` [PATCH v3 11/13] hw/arm/raspi: Extract the board model from the " Philippe Mathieu-Daudé
2020-02-08 16:56 ` Philippe Mathieu-Daudé
2020-02-08 16:56 ` [PATCH v3 12/13] hw/arm/raspi: Use a unique raspi_machine_class_init() method Philippe Mathieu-Daudé
2020-02-10 10:01 ` Igor Mammedov
2020-02-10 10:01 ` Igor Mammedov
2020-02-13 13:59 ` Peter Maydell
2020-02-13 14:15 ` Philippe Mathieu-Daudé
2020-02-13 14:15 ` Philippe Mathieu-Daudé
2020-02-13 14:32 ` Peter Maydell
2020-02-13 15:33 ` Philippe Mathieu-Daudé
2020-02-13 15:33 ` Philippe Mathieu-Daudé
2020-02-15 17:45 ` Philippe Mathieu-Daudé
2020-02-15 17:45 ` Philippe Mathieu-Daudé
2020-02-08 16:56 ` [PATCH v3 13/13] hw/arm/raspi: Extract the cores count from the board revision Philippe Mathieu-Daudé
2020-02-10 10:03 ` Igor Mammedov
2020-02-10 10:03 ` Igor Mammedov
2020-02-13 14:00 ` [PATCH v3 00/13] hw/arm/raspi: Dynamically create machines based on " Peter Maydell
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=20200210140918.270c8a5d@redhat.com \
--to=imammedo@redhat.com \
--cc=Andrew.Baumann@microsoft.com \
--cc=alistair@alistair23.me \
--cc=balaton@eik.bme.hu \
--cc=estebanbosse@gmail.com \
--cc=f4bug@amsat.org \
--cc=me@xcancerberox.com.ar \
--cc=nieklinnenbank@gmail.com \
--cc=peter.maydell@linaro.org \
--cc=philmd@redhat.com \
--cc=qemu-arm@nongnu.org \
--cc=qemu-devel@nongnu.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.