Re: [PATCH] aspeed: Enable backend file for eeprom

[Cédric Le Goater]

On 7/28/22 09:20, John Wang wrote:

Cédric Le Goater <c...@kaod.org> 于2022年7月28日周四 14:28写道：

Hello John,

On 7/28/22 08:12, John Wang wrote:

tested on a fp5280g2:

$QEMU_BIN -machine fp5280g2-bmc \
         -nographic \
         -drive file="${IMAGE_PATH}",format=raw,if=mtd \
         -drive file="eeprom.bin",format=raw,if=pflash,index=1 \
         ${NIC}

root@fp5280g2:/sys/bus/i2c/devices/1-0050# hexdump eeprom -C
00000000  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
....
....
00000240  2c 87 a3 a4 1d d3 11 b2  02 d2 c2 9d 44 60 cf 3e  |,...........D`.>|
00000250  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|

It's same as the "eeprom.bin"

Signed-off-by: John Wang <wangzhiqian...@inspur.com>
Change-Id: I5c44785a028144b24aa0b22643266d83addc5eab
---
   hw/arm/aspeed.c | 16 ++++++++++++----
   1 file changed, 12 insertions(+), 4 deletions(-)

diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c
index 4193a3d23d..80aa687372 100644
--- a/hw/arm/aspeed.c
+++ b/hw/arm/aspeed.c
@@ -431,12 +431,20 @@ static void aspeed_machine_init(MachineState *machine)
       arm_load_kernel(ARM_CPU(first_cpu), machine, &aspeed_board_binfo);
   }

-static void at24c_eeprom_init(I2CBus *bus, uint8_t addr, uint32_t rsize)
+static void at24c_eeprom_init(I2CBus *bus, uint8_t addr, uint32_t rsize,
+                              int index)
   {
       I2CSlave *i2c_dev = i2c_slave_new("at24c-eeprom", addr);
       DeviceState *dev = DEVICE(i2c_dev);

+    DriveInfo *dinfo = drive_get_by_index(IF_PFLASH, index);

I don't think IF_PFLASH is the appropriate type.

thanks

Jae proposed a similar patch with IF_NONE which should fit your need :

    
https://lore.kernel.org/all/20220718175214.2087644-1-quic_jaeh...@quicinc.com/

Could you please give it a try ?

I tested on a fp5280g2-bmc, It's ok.  I would abandon my patch :）

It's available on my branch :

    https://github.com/legoater/qemu/commits/aspeed-7.1

I checked it, and will use this tree to module a new machine. :)

Or simply grab the patch if you only need one.

I don't think this is the correct approach for mainline. See this thread :

  
https://lore.kernel.org/all/CAFEAcA8sNjLsknea5Nt-tANEniFF2FYmjiV0xz=pr+vfwkx...@mail.gmail.com/t/

C.

Thanks,

C.

+    BlockBackend *blk = dinfo ? blk_by_legacy_dinfo(dinfo) : NULL;
+
+    if (blk) {
+        qdev_prop_set_drive(DEVICE(dev), "drive", blk);
+    }
       qdev_prop_set_uint32(dev, "rom-size", rsize);
+
       i2c_slave_realize_and_unref(i2c_dev, bus, &error_abort);
   }

@@ -685,7 +693,7 @@ static void fp5280g2_bmc_i2c_init(AspeedMachineState *bmc)
       I2CSlave *i2c_mux;

       /* The at24c256 */
-    at24c_eeprom_init(aspeed_i2c_get_bus(&soc->i2c, 1), 0x50, 32768);
+    at24c_eeprom_init(aspeed_i2c_get_bus(&soc->i2c, 1), 0x50, 32768, 1);

       /* The fp5280g2 expects a TMP112 but a TMP105 is compatible */
       i2c_slave_create_simple(aspeed_i2c_get_bus(&soc->i2c, 2), TYPE_TMP105,
@@ -918,13 +926,13 @@ static void bletchley_bmc_i2c_init(AspeedMachineState 
*bmc)
       }

       /* Bus 6 */
-    at24c_eeprom_init(i2c[6], 0x56, 65536);
+    at24c_eeprom_init(i2c[6], 0x56, 65536, 1);
       /* Missing model: nxp,pcf85263 @ 0x51 , but ds1338 works enough */
       i2c_slave_create_simple(i2c[6], "ds1338", 0x51);


       /* Bus 7 */
-    at24c_eeprom_init(i2c[7], 0x54, 65536);
+    at24c_eeprom_init(i2c[7], 0x54, 65536, 2);

       /* Bus 9 */
       i2c_slave_create_simple(i2c[9], TYPE_TMP421, 0x4f);