qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH v3 16/16] i2c:smbus_eeprom: Add a reset function


From: Corey Minyard
Subject: Re: [Qemu-devel] [PATCH v3 16/16] i2c:smbus_eeprom: Add a reset function to smbus_eeprom
Date: Mon, 26 Nov 2018 16:41:18 -0600
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.2.1

On 11/26/18 2:42 PM, Philippe Mathieu-Daudé wrote:
Hi Corey,

On 26/11/18 21:04, address@hidden wrote:
From: Corey Minyard <address@hidden>

Reset the contents to init data and reset the offset on a machine
reset.

Signed-off-by: Corey Minyard <address@hidden>
---
  hw/i2c/smbus_eeprom.c | 8 +++++++-
  1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/hw/i2c/smbus_eeprom.c b/hw/i2c/smbus_eeprom.c
index a0dcadbd60..de3a492df4 100644
--- a/hw/i2c/smbus_eeprom.c
+++ b/hw/i2c/smbus_eeprom.c
@@ -107,7 +107,7 @@ static const VMStateDescription vmstate_smbus_eeprom = {
      }
  };
-static void smbus_eeprom_realize(DeviceState *dev, Error **errp)
+static void smbus_eeprom_reset(DeviceState *dev)
  {
      SMBusEEPROMDevice *eeprom = SMBUS_EEPROM(dev);
'git diff -U4' also shows this line:

        memcpy(eeprom->data, eeprom->init_data, SMBUS_EEPROM_SIZE);

I don't think this is correct.

One test I'd like to have is a machine booting, updating the EPROM then
rebooting calling hw reset() to use the new values (BIOS use this).

With this patch this won't work, you'll restore the EPROM content on
each machine reset.

I'd move the memcpy() call to the realize() function.

What do you think?

There was some debate on this in the earlier patch set.  The general principle is that a reset is the same as starting up qemu from scratch, so I added this
code based on that principle.  But I'm not really sure.

      eeprom->offset = 0;
This is correct, the offset reset belongs to the reset() function.

Actually, on a real system, a hardware reset will generally not affect the
eeprom current offset register.  So if we don't take the above code, then
IMHO this is wrong, too.

-corey


Regards,

Phil.

  }
+static void smbus_eeprom_realize(DeviceState *dev, Error **errp)
+{
+    smbus_eeprom_reset(dev);
+}
+
  static Property smbus_eeprom_properties[] = {
      DEFINE_PROP_PTR("data", SMBusEEPROMDevice, init_data),
      DEFINE_PROP_END_OF_LIST(),
@@ -126,6 +131,7 @@ static void smbus_eeprom_class_initfn(ObjectClass *klass, 
void *data)
      SMBusDeviceClass *sc = SMBUS_DEVICE_CLASS(klass);
dc->realize = smbus_eeprom_realize;
+    dc->reset = smbus_eeprom_reset;
      sc->receive_byte = eeprom_receive_byte;
      sc->write_data = eeprom_write_data;
      dc->props = smbus_eeprom_properties;





reply via email to

[Prev in Thread] Current Thread [Next in Thread]