Skip to contentSkip to conversation

hw/riscv: virt: Enable booting S-mode firmware from pflash

PATCH Superseded Sunil V L 2 participants 3 messages Sep 5, 2022 RISC-V

v1 of 5 Superseded Latest is v5 →

lore
Share & tools
lore

Sunil V L

author
#0 lore raw

PATCH 1/1 hw/riscv: virt: Enable booting S-mode firmware from pflash

To boot S-mode firmware payload like EDK2 from persistent flash storage, qemu needs to pass the flash address as the next_addr in fw_dynamic_info to the opensbi.

Update the kernel_entry with the flash address when no kernel is specified via QEMU command line so that opensbi
can jump to the flash address. This assumes flash drive unit 1 is used to keep dynamic payload similar to other architectures.

---
 hw/riscv/virt.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
hw/riscv/virt.c +12 -0
index ff8c0df5cd..1ee855fded 100644
@@ -1284,6 +1284,18 @@ static void virt_machine_done(Notifier *notifier, void *data)
         start_addr = virt_memmap[VIRT_FLASH].base;
     }
 
+    if (drive_get(IF_PFLASH, 0, 1)) {
+        /*
+         * Assume second pflash unit (unit=1) to hold the S-mode firmware
+         * like EDK2. Set kernel_entry to flash address if kernel argument
+         * is not set and pflash unit 1 is configured.
+         */
+        if (!kernel_entry) {
+            kernel_entry = virt_memmap[VIRT_FLASH].base +
+                             virt_memmap[VIRT_FLASH].size / 2;
+        }
+    }
+
     /*
      * Init fw_cfg.  Must be done before riscv_load_fdt, otherwise the device
      * tree cannot be altered and we get FDT_ERR_NOSPACE.
Signature · 1 line
2.25.1
#1 lore raw
Quoted 41 lines · Sunil V L

To boot S-mode firmware payload like EDK2 from persistent flash storage, qemu needs to pass the flash address as the next_addr in fw_dynamic_info to the opensbi.

Update the kernel_entry with the flash address when no kernel is specified via QEMU command line so that opensbi
can jump to the flash address. This assumes flash drive unit 1 is used to keep dynamic payload similar to other architectures.

---
 hw/riscv/virt.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
hw/riscv/virt.c +12 -0
index ff8c0df5cd..1ee855fded 100644
@@ -1284,6 +1284,18 @@ static void virt_machine_done(Notifier *notifier, void *data)
         start_addr = virt_memmap[VIRT_FLASH].base;
     }
 
+    if (drive_get(IF_PFLASH, 0, 1)) {
+        /*
+         * Assume second pflash unit (unit=1) to hold the S-mode firmware
+         * like EDK2. Set kernel_entry to flash address if kernel argument
+         * is not set and pflash unit 1 is configured.
+         */
+        if (!kernel_entry) {
+            kernel_entry = virt_memmap[VIRT_FLASH].base +
+                             virt_memmap[VIRT_FLASH].size / 2;
+        }
+    }
+
     /*
      * Init fw_cfg.  Must be done before riscv_load_fdt, otherwise the device
      * tree cannot be altered and we get FDT_ERR_NOSPACE.
Signature · 1 line
2.25.1

Hi Sunil,

We should ensure we can use both edk2 and the '-kernel' option. For example, guest installation will typically do direct kernel boot, and when it's desired / necessary to boot that kernel with ACPI, then we'll also need edk2.

Thanks,
drew

author
#2 lore raw
Quoted 45 lines · Andrew Jones
Quoted 41 lines · Sunil V L

To boot S-mode firmware payload like EDK2 from persistent flash storage, qemu needs to pass the flash address as the next_addr in fw_dynamic_info to the opensbi.

Update the kernel_entry with the flash address when no kernel is specified via QEMU command line so that opensbi
can jump to the flash address. This assumes flash drive unit 1 is used to keep dynamic payload similar to other architectures.

---
 hw/riscv/virt.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
hw/riscv/virt.c +12 -0
index ff8c0df5cd..1ee855fded 100644
@@ -1284,6 +1284,18 @@ static void virt_machine_done(Notifier *notifier, void *data)
         start_addr = virt_memmap[VIRT_FLASH].base;
     }
 
+    if (drive_get(IF_PFLASH, 0, 1)) {
+        /*
+         * Assume second pflash unit (unit=1) to hold the S-mode firmware
+         * like EDK2. Set kernel_entry to flash address if kernel argument
+         * is not set and pflash unit 1 is configured.
+         */
+        if (!kernel_entry) {
+            kernel_entry = virt_memmap[VIRT_FLASH].base +
+                             virt_memmap[VIRT_FLASH].size / 2;
+        }
+    }
+
     /*
      * Init fw_cfg.  Must be done before riscv_load_fdt, otherwise the device
      * tree cannot be altered and we get FDT_ERR_NOSPACE.
Signature · 1 line
2.25.1

Hi Sunil,

Quoted 4 lines

We should ensure we can use both edk2 and the '-kernel' option. For example, guest installation will typically do direct kernel boot, and when it's desired / necessary to boot that kernel with ACPI, then we'll also need edk2.

Thank you very much, Drew. Let me revise the patch and send.

Thanks
Sunil

Quoted 3 lines

Thanks,
drew