Skip to contentSkip to conversation

ppc: fix boot with sam460ex

PATCH In review Michael S. Tsirkin 4 participants 9 messages May 26, 2022 – Jun 8, 2022

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ppc: fix boot with sam460ex

Recent changes to pcie_host corrected size of its internal region to match what it expects - only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but it so happened that ppc actually allows guest to configure as large a size as it wants to, and current firmware set it to 1 << 29.

With just qemu-system-ppc -M sam460ex this triggers an assert which seems to happen when the guest (board firmware?) writes a value to CFGMSK reg:

(gdb) bt

This is done in the board firmware here:

https://git.qemu.org/?p=u-boot-sam460ex.git;a=blob;f=arch/powerpc/cpu/ppc4xx/4xx_pcie.c;h=13348be93dccc74c13ea043d6635a7f8ece4b5f0;hb=HEAD

when trying to map config space.

Note that what firmware does matches
https://www.hardware.com.br/comunidade/switch-cisco/1128380/

So it's not clear what the proper fix should be.

However, allowing guest to trigger an assert in qemu is not good practice anyway.

For now let's just force the mask to 256MB on guest write, this way anything outside the expected address range is ignored.

Fixes commit 1f1a7b2269 ("include/hw/pci/pcie_host: Correct PCIE_MMCFG_SIZE_MAX")

---

Affected system is orphan so I guess I will merge the patch unless someone objects.

 hw/ppc/ppc440_uc.c | 8 ++++++++
 1 file changed, 8 insertions(+)
hw/ppc/ppc440_uc.c +8 -0
index 993e3ba955..a1ecf6dd1c 100644
@@ -1180,6 +1180,14 @@ static void dcr_write_pcie(void *opaque, int dcrn, uint32_t val)
     case PEGPL_CFGMSK:
         s->cfg_mask = val;
         size = ~(val & 0xfffffffe) + 1;
+        /*
+         * Firmware sets this register to E0000001. Why we are not sure,
+         * but the current guess is anything above PCIE_MMCFG_SIZE_MAX is
+         * ignored.
+         */
+        if (size > PCIE_MMCFG_SIZE_MAX) {
+            size = PCIE_MMCFG_SIZE_MAX;
+        }
         pcie_host_mmcfg_update(PCIE_HOST_BRIDGE(s), val & 1, s->cfg_base, size);
         break;
     case PEGPL_MSGBAH:
Signature · 1 line
MST
#1 lore raw

Hello,

Some changes to commit message (patch is OK).

Quoted 7 lines · Michael S. Tsirkin

Recent changes to pcie_host corrected size of its internal region to match what it expects - only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but it so happened that ppc actually allows guest to configure as large a size as it wants to, and current firmware set it to 1 << 29.

Quoted 5 lines

With just qemu-system-ppc -M sam460ex this triggers an assert which seems to happen when the guest (board firmware?) writes a value to CFGMSK reg:

(gdb) bt

Backtrace is missing but you could just drop this line and replace : with . at end of previous line as we probably don't need the full backtrace, the commit message is already too long in my opinion.

Quoted 8 lines

This is done in the board firmware here:

https://git.qemu.org/?p=u-boot-sam460ex.git;a=blob;f=arch/powerpc/cpu/ppc4xx/4xx_pcie.c;h=13348be93dccc74c13ea043d6635a7f8ece4b5f0;hb=HEAD

when trying to map config space.

Note that what firmware does matches
https://www.hardware.com.br/comunidade/switch-cisco/1128380/

That's not it. It's a different hardware and firmware, just quoted it as an example that this value seems to be common to that SoC even on different hardware/OS/firmware (probably comes from reference hardware/devel board?). The sam460ex is here

https://www.acube-systems.biz/index.php?page=hardware&pid=5

the U-Boot in above repo is matching the firmware from the acube page but I had to fix some bugs in it to make it compile and work.

Otherwise this should be OK.

Regards,
BALATON Zoltan

Quoted 39 lines

So it's not clear what the proper fix should be.

However, allowing guest to trigger an assert in qemu is not good practice anyway.

For now let's just force the mask to 256MB on guest write, this way anything outside the expected address range is ignored.

Fixes commit 1f1a7b2269 ("include/hw/pci/pcie_host: Correct PCIE_MMCFG_SIZE_MAX")

---

Affected system is orphan so I guess I will merge the patch unless someone objects.

hw/ppc/ppc440_uc.c | 8 ++++++++

1 file changed, 8 insertions(+)
hw/ppc/ppc440_uc.c +8 -0
index 993e3ba955..a1ecf6dd1c 100644
@@ -1180,6 +1180,14 @@ static void dcr_write_pcie(void *opaque, int dcrn, uint32_t val)
    case PEGPL_CFGMSK:
        s->cfg_mask = val;
        size = ~(val & 0xfffffffe) + 1;
+        /*
+         * Firmware sets this register to E0000001. Why we are not sure,
+         * but the current guess is anything above PCIE_MMCFG_SIZE_MAX is
+         * ignored.
+         */
+        if (size > PCIE_MMCFG_SIZE_MAX) {
+            size = PCIE_MMCFG_SIZE_MAX;
+        }
        pcie_host_mmcfg_update(PCIE_HOST_BRIDGE(s), val & 1, s->cfg_base, size);
        break;
    case PEGPL_MSGBAH:
#2 lore raw
Quoted 3 lines · BALATON Zoltan

Hello,

Some changes to commit message (patch is OK).

Want to write the commit message for me then?

Quoted 81 lines · Michael S. Tsirkin
Quoted 7 lines

Recent changes to pcie_host corrected size of its internal region to match what it expects - only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but it so happened that ppc actually allows guest to configure as large a size as it wants to, and current firmware set it to 1 << 29.

Quoted 5 lines

With just qemu-system-ppc -M sam460ex this triggers an assert which seems to happen when the guest (board firmware?) writes a value to CFGMSK reg:

(gdb) bt

Backtrace is missing but you could just drop this line and replace : with . at end of previous line as we probably don't need the full backtrace, the commit message is already too long in my opinion.

Quoted 8 lines

This is done in the board firmware here:

https://git.qemu.org/?p=u-boot-sam460ex.git;a=blob;f=arch/powerpc/cpu/ppc4xx/4xx_pcie.c;h=13348be93dccc74c13ea043d6635a7f8ece4b5f0;hb=HEAD

when trying to map config space.

Note that what firmware does matches
https://www.hardware.com.br/comunidade/switch-cisco/1128380/

That's not it. It's a different hardware and firmware, just quoted it as an example that this value seems to be common to that SoC even on different hardware/OS/firmware (probably comes from reference hardware/devel board?). The sam460ex is here

https://www.acube-systems.biz/index.php?page=hardware&pid=5

the U-Boot in above repo is matching the firmware from the acube page but I had to fix some bugs in it to make it compile and work.

Otherwise this should be OK.

Regards,
BALATON Zoltan

Quoted 39 lines

So it's not clear what the proper fix should be.

However, allowing guest to trigger an assert in qemu is not good practice anyway.

For now let's just force the mask to 256MB on guest write, this way anything outside the expected address range is ignored.

Fixes commit 1f1a7b2269 ("include/hw/pci/pcie_host: Correct PCIE_MMCFG_SIZE_MAX")

---

Affected system is orphan so I guess I will merge the patch unless someone objects.

hw/ppc/ppc440_uc.c | 8 ++++++++

1 file changed, 8 insertions(+)
hw/ppc/ppc440_uc.c +8 -0
index 993e3ba955..a1ecf6dd1c 100644
@@ -1180,6 +1180,14 @@ static void dcr_write_pcie(void *opaque, int dcrn, uint32_t val)
    case PEGPL_CFGMSK:
        s->cfg_mask = val;
        size = ~(val & 0xfffffffe) + 1;
+        /*
+         * Firmware sets this register to E0000001. Why we are not sure,
+         * but the current guess is anything above PCIE_MMCFG_SIZE_MAX is
+         * ignored.
+         */
+        if (size > PCIE_MMCFG_SIZE_MAX) {
+            size = PCIE_MMCFG_SIZE_MAX;
+        }
        pcie_host_mmcfg_update(PCIE_HOST_BRIDGE(s), val & 1, s->cfg_base, size);
        break;
    case PEGPL_MSGBAH:
#3 lore raw
Quoted 6 lines · Michael S. Tsirkin
Quoted 3 lines · BALATON Zoltan

Hello,

Some changes to commit message (patch is OK).

Want to write the commit message for me then?

How about:

Recent changes to pcie_host corrected size of its internal region to match what it expects: only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but PPC 4xx actually allows guest to configure different sizes, and some firmwares seem to set it to 1 << 29.

This caused e.g. qemu-system-ppc -M sam460ex to exit with an assert when the guest writes a value to CFGMSK register when trying to map config space. This is done in the board firmware in ppc4xx_init_pcie_port() in roms/u-boot-sam460ex/arch/powerpc/cpu/ppc4xx/4xx_pcie.c

It's not clear what the proper fix should be but for now let's force the size to 256MB, so anything outside the expected address range is ignored.

Quoted 34 lines
Quoted 32 lines
Quoted 32 lines
Fixes commit 1f1a7b2269 ("include/hw/pci/pcie_host: Correct PCIE_MMCFG_SIZE_MAX")

---

Affected system is orphan so I guess I will merge the patch unless someone objects.

hw/ppc/ppc440_uc.c | 8 ++++++++

1 file changed, 8 insertions(+)
hw/ppc/ppc440_uc.c +8 -0
index 993e3ba955..a1ecf6dd1c 100644
@@ -1180,6 +1180,14 @@ static void dcr_write_pcie(void *opaque, int dcrn, uint32_t val)
    case PEGPL_CFGMSK:
        s->cfg_mask = val;
        size = ~(val & 0xfffffffe) + 1;
+        /*
+         * Firmware sets this register to E0000001. Why we are not sure,
+         * but the current guess is anything above PCIE_MMCFG_SIZE_MAX is
+         * ignored.
+         */
+        if (size > PCIE_MMCFG_SIZE_MAX) {
+            size = PCIE_MMCFG_SIZE_MAX;
+        }
        pcie_host_mmcfg_update(PCIE_HOST_BRIDGE(s), val & 1, s->cfg_base, size);
        break;
    case PEGPL_MSGBAH:
Acked-by on v1 1/1
#4 lore raw
In reply to Michael S. Tsirkin
Quoted 29 lines · Michael S. Tsirkin

Recent changes to pcie_host corrected size of its internal region to match what it expects - only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but it so happened that ppc actually allows guest to configure as large a size as it wants to, and current firmware set it to 1 << 29.

With just qemu-system-ppc -M sam460ex this triggers an assert which seems to happen when the guest (board firmware?) writes a value to CFGMSK reg:

(gdb) bt

This is done in the board firmware here:

https://git.qemu.org/?p=u-boot-sam460ex.git;a=blob;f=arch/powerpc/cpu/ppc4xx/4xx_pcie.c;h=13348be93dccc74c13ea043d6635a7f8ece4b5f0;hb=HEAD

when trying to map config space.

Note that what firmware does matches
https://www.hardware.com.br/comunidade/switch-cisco/1128380/

So it's not clear what the proper fix should be.

However, allowing guest to trigger an assert in qemu is not good practice anyway.

For now let's just force the mask to 256MB on guest write, this way anything outside the expected address range is ignored.

Quoted 8 lines
Fixes commit 1f1a7b2269 ("include/hw/pci/pcie_host: Correct PCIE_MMCFG_SIZE_MAX")

---

Affected system is orphan so I guess I will merge the patch unless someone objects.

Fine with me.

On the orphan status,

MAINTAINERS file says the sam460ex machine is maintained. I understand that these files :

hw/ppc/ppc440_uc.c
hw/ppc/ppc440.h
hw/ppc/ppc440_pcix.c
hw/ppc/ppc4xx_devs.c
include/hw/ppc/ppc4xx.h

and these

include/hw/i2c/ppc4xx_i2c.h
hw/i2c/ppc4xx_i2c.c
hw/intc/ppc-uic.c
include/hw/intc/ppc-uic.h

should be under the same entry since sam460ex depends on it.

The ppc440 support is a bit of a mess to be honest. We have two 440 machines bamboo and sam460ex which have a lot in common a part from the PCI host bridge.

Thanks,

C.

Quoted 23 lines
  hw/ppc/ppc440_uc.c | 8 ++++++++
  1 file changed, 8 insertions(+)
hw/ppc/ppc440_uc.c +8 -0
index 993e3ba955..a1ecf6dd1c 100644
@@ -1180,6 +1180,14 @@ static void dcr_write_pcie(void *opaque, int dcrn, uint32_t val)
      case PEGPL_CFGMSK:
          s->cfg_mask = val;
          size = ~(val & 0xfffffffe) + 1;
+        /*
+         * Firmware sets this register to E0000001. Why we are not sure,
+         * but the current guess is anything above PCIE_MMCFG_SIZE_MAX is
+         * ignored.
+         */
+        if (size > PCIE_MMCFG_SIZE_MAX) {
+            size = PCIE_MMCFG_SIZE_MAX;
+        }
          pcie_host_mmcfg_update(PCIE_HOST_BRIDGE(s), val & 1, s->cfg_base, size);
          break;
      case PEGPL_MSGBAH:
Acked-by on v1 1/1
#5 lore raw
In reply to Michael S. Tsirkin

[ resend with a reply-all ]

Quoted 29 lines · Michael S. Tsirkin

Recent changes to pcie_host corrected size of its internal region to match what it expects - only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but it so happened that ppc actually allows guest to configure as large a size as it wants to, and current firmware set it to 1 << 29.

With just qemu-system-ppc -M sam460ex this triggers an assert which seems to happen when the guest (board firmware?) writes a value to CFGMSK reg:

(gdb) bt

This is done in the board firmware here:

https://git.qemu.org/?p=u-boot-sam460ex.git;a=blob;f=arch/powerpc/cpu/ppc4xx/4xx_pcie.c;h=13348be93dccc74c13ea043d6635a7f8ece4b5f0;hb=HEAD

when trying to map config space.

Note that what firmware does matches
https://www.hardware.com.br/comunidade/switch-cisco/1128380/

So it's not clear what the proper fix should be.

However, allowing guest to trigger an assert in qemu is not good practice anyway.

For now let's just force the mask to 256MB on guest write, this way anything outside the expected address range is ignored.

Quoted 8 lines
Fixes commit 1f1a7b2269 ("include/hw/pci/pcie_host: Correct PCIE_MMCFG_SIZE_MAX")

---

Affected system is orphan so I guess I will merge the patch unless someone objects.

Fine with me.

On the orphan status,

MAINTAINERS file says the sam460ex machine is maintained. I understand that these files :

hw/ppc/ppc440_uc.c
hw/ppc/ppc440.h
hw/ppc/ppc440_pcix.c
hw/ppc/ppc4xx_devs.c
include/hw/ppc/ppc4xx.h

and these

include/hw/i2c/ppc4xx_i2c.h
hw/i2c/ppc4xx_i2c.c
hw/intc/ppc-uic.c
include/hw/intc/ppc-uic.h

should be under the same entry since sam460ex depends on it.

The ppc440 support is a bit of a mess to be honest. We have two 440 machines bamboo and sam460ex which have a lot in common a part from the PCI host bridge.

Thanks,

C.

Quoted 23 lines
  hw/ppc/ppc440_uc.c | 8 ++++++++
  1 file changed, 8 insertions(+)
hw/ppc/ppc440_uc.c +8 -0
index 993e3ba955..a1ecf6dd1c 100644
@@ -1180,6 +1180,14 @@ static void dcr_write_pcie(void *opaque, int dcrn, uint32_t val)
      case PEGPL_CFGMSK:
          s->cfg_mask = val;
          size = ~(val & 0xfffffffe) + 1;
+        /*
+         * Firmware sets this register to E0000001. Why we are not sure,
+         * but the current guess is anything above PCIE_MMCFG_SIZE_MAX is
+         * ignored.
+         */
+        if (size > PCIE_MMCFG_SIZE_MAX) {
+            size = PCIE_MMCFG_SIZE_MAX;
+        }
          pcie_host_mmcfg_update(PCIE_HOST_BRIDGE(s), val & 1, s->cfg_base, size);
          break;
      case PEGPL_MSGBAH:
#6 lore raw
Quoted 41 lines · Cédric Le Goater

[ resend with a reply-all ]

Quoted 29 lines · Michael S. Tsirkin

Recent changes to pcie_host corrected size of its internal region to match what it expects - only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but it so happened that ppc actually allows guest to configure as large a size as it wants to, and current firmware set it to 1 << 29.

With just qemu-system-ppc -M sam460ex this triggers an assert which seems to happen when the guest (board firmware?) writes a value to CFGMSK reg:

(gdb) bt

This is done in the board firmware here:

https://git.qemu.org/?p=u-boot-sam460ex.git;a=blob;f=arch/powerpc/cpu/ppc4xx/4xx_pcie.c;h=13348be93dccc74c13ea043d6635a7f8ece4b5f0;hb=HEAD

when trying to map config space.

Note that what firmware does matches
https://www.hardware.com.br/comunidade/switch-cisco/1128380/

So it's not clear what the proper fix should be.

However, allowing guest to trigger an assert in qemu is not good practice anyway.

For now let's just force the mask to 256MB on guest write, this way anything outside the expected address range is ignored.

Quoted 8 lines
Fixes commit 1f1a7b2269 ("include/hw/pci/pcie_host: Correct PCIE_MMCFG_SIZE_MAX")

---

Affected system is orphan so I guess I will merge the patch unless someone objects.

Quoted 8 lines

Fine with me.

On the orphan status,

MAINTAINERS file says the sam460ex machine is maintained. I understand that

I'm maintaining sam460ex but I don't know much about other parts so don't want to take up all the PPC4xx emulation so only files I've written or made extensive changes so I know them are listed. Others are part of pre-existing PPC4xx emulation which may be orphan.

Quoted 5 lines

these files :

hw/ppc/ppc440_uc.c
hw/ppc/ppc440.h
hw/ppc/ppc440_pcix.c

The ppc440_pcix is already listed, the rest aren't something I know much about so wouldn't volunteer to maintain those. They are shared with other PPC 4xx machines.

Regards,
BALATON Zoltan

Quoted 47 lines

hw/ppc/ppc4xx_devs.c
include/hw/ppc/ppc4xx.h

and these

include/hw/i2c/ppc4xx_i2c.h
hw/i2c/ppc4xx_i2c.c
hw/intc/ppc-uic.c
include/hw/intc/ppc-uic.h

should be under the same entry since sam460ex depends on it.

The ppc440 support is a bit of a mess to be honest. We have two 440 machines bamboo and sam460ex which have a lot in common a part from the PCI host bridge.

Thanks,

C.

Quoted 25 lines
  hw/ppc/ppc440_uc.c | 8 ++++++++
  1 file changed, 8 insertions(+)
hw/ppc/ppc440_uc.c +8 -0
index 993e3ba955..a1ecf6dd1c 100644
@@ -1180,6 +1180,14 @@ static void dcr_write_pcie(void *opaque, int dcrn, 
 uint32_t val)
      case PEGPL_CFGMSK:
          s->cfg_mask = val;
          size = ~(val & 0xfffffffe) + 1;
+        /*
+         * Firmware sets this register to E0000001. Why we are not sure,
+         * but the current guess is anything above PCIE_MMCFG_SIZE_MAX is
+         * ignored.
+         */
+        if (size > PCIE_MMCFG_SIZE_MAX) {
+            size = PCIE_MMCFG_SIZE_MAX;
+        }
          pcie_host_mmcfg_update(PCIE_HOST_BRIDGE(s), val & 1, s->cfg_base, 
 size);
          break;
      case PEGPL_MSGBAH:
#7 lore raw
In reply to Michael S. Tsirkin

Michael,

I'll queue this patch with the commit msg proposed by Zoltan as follows:

Author: Michael S. Tsirkin <[email protected]>
Date:   Thu May 26 18:43:43 2022 -0400

ppc: fix boot with sam460ex

Recent changes to pcie_host corrected size of its internal region to match what it expects: only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but PPC 4xx actually allows guest to configure different sizes, and some firmwares seem to set it to 1 << 29.

This caused e.g. qemu-system-ppc -M sam460ex to exit with an assert when the guest writes a value to CFGMSK register when trying to map config space. This is done in the board firmware in ppc4xx_init_pcie_port() in roms/u-boot-sam460ex/arch/powerpc/cpu/ppc4xx/4xx_pcie.c

It's not clear what the proper fix should be but for now let's force the size to 256MB, so anything outside the expected address range is ignored.

Is that ok with you?

Thanks,

Daniel

Quoted 60 lines · Michael S. Tsirkin

Recent changes to pcie_host corrected size of its internal region to match what it expects - only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but it so happened that ppc actually allows guest to configure as large a size as it wants to, and current firmware set it to 1 << 29.

With just qemu-system-ppc -M sam460ex this triggers an assert which seems to happen when the guest (board firmware?) writes a value to CFGMSK reg:

(gdb) bt

This is done in the board firmware here:

https://git.qemu.org/?p=u-boot-sam460ex.git;a=blob;f=arch/powerpc/cpu/ppc4xx/4xx_pcie.c;h=13348be93dccc74c13ea043d6635a7f8ece4b5f0;hb=HEAD

when trying to map config space.

Note that what firmware does matches
https://www.hardware.com.br/comunidade/switch-cisco/1128380/

So it's not clear what the proper fix should be.

However, allowing guest to trigger an assert in qemu is not good practice anyway.

For now let's just force the mask to 256MB on guest write, this way anything outside the expected address range is ignored.

Fixes commit 1f1a7b2269 ("include/hw/pci/pcie_host: Correct PCIE_MMCFG_SIZE_MAX")

---

Affected system is orphan so I guess I will merge the patch unless someone objects.

  hw/ppc/ppc440_uc.c | 8 ++++++++
  1 file changed, 8 insertions(+)
hw/ppc/ppc440_uc.c +8 -0
index 993e3ba955..a1ecf6dd1c 100644
@@ -1180,6 +1180,14 @@ static void dcr_write_pcie(void *opaque, int dcrn, uint32_t val)
      case PEGPL_CFGMSK:
          s->cfg_mask = val;
          size = ~(val & 0xfffffffe) + 1;
+        /*
+         * Firmware sets this register to E0000001. Why we are not sure,
+         * but the current guess is anything above PCIE_MMCFG_SIZE_MAX is
+         * ignored.
+         */
+        if (size > PCIE_MMCFG_SIZE_MAX) {
+            size = PCIE_MMCFG_SIZE_MAX;
+        }
          pcie_host_mmcfg_update(PCIE_HOST_BRIDGE(s), val & 1, s->cfg_base, size);
          break;
      case PEGPL_MSGBAH:
#8 lore raw
Quoted 25 lines · Daniel Henrique Barboza

Michael,

I'll queue this patch with the commit msg proposed by Zoltan as follows:

Author: Michael S. Tsirkin <[email protected]>
Date:   Thu May 26 18:43:43 2022 -0400

ppc: fix boot with sam460ex
Recent changes to pcie_host corrected size of its internal region to match what it expects: only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but PPC 4xx actually allows guest to configure different sizes, and some firmwares seem to set it to 1 << 29.
This caused e.g. qemu-system-ppc -M sam460ex to exit with an assert when the guest writes a value to CFGMSK register when trying to map config space. This is done in the board firmware in ppc4xx_init_pcie_port() in roms/u-boot-sam460ex/arch/powerpc/cpu/ppc4xx/4xx_pcie.c
It's not clear what the proper fix should be but for now let's force the size to 256MB, so anything outside the expected address range is ignored.

Quoted 7 lines

Is that ok with you?

Thanks,

Daniel

ACK

Quoted 62 lines · Michael S. Tsirkin
Quoted 60 lines

Recent changes to pcie_host corrected size of its internal region to match what it expects - only the low 28 bits are ever decoded. Previous code just ignored bit 29 (if size was 1 << 29) in the address which does not make much sense. We are now asserting on size > 1 << 28 instead, but it so happened that ppc actually allows guest to configure as large a size as it wants to, and current firmware set it to 1 << 29.

With just qemu-system-ppc -M sam460ex this triggers an assert which seems to happen when the guest (board firmware?) writes a value to CFGMSK reg:

(gdb) bt

This is done in the board firmware here:

https://git.qemu.org/?p=u-boot-sam460ex.git;a=blob;f=arch/powerpc/cpu/ppc4xx/4xx_pcie.c;h=13348be93dccc74c13ea043d6635a7f8ece4b5f0;hb=HEAD

when trying to map config space.

Note that what firmware does matches
https://www.hardware.com.br/comunidade/switch-cisco/1128380/

So it's not clear what the proper fix should be.

However, allowing guest to trigger an assert in qemu is not good practice anyway.

For now let's just force the mask to 256MB on guest write, this way anything outside the expected address range is ignored.

Fixes commit 1f1a7b2269 ("include/hw/pci/pcie_host: Correct PCIE_MMCFG_SIZE_MAX")

---

Affected system is orphan so I guess I will merge the patch unless someone objects.

  hw/ppc/ppc440_uc.c | 8 ++++++++
  1 file changed, 8 insertions(+)
hw/ppc/ppc440_uc.c +8 -0
index 993e3ba955..a1ecf6dd1c 100644
@@ -1180,6 +1180,14 @@ static void dcr_write_pcie(void *opaque, int dcrn, uint32_t val)
      case PEGPL_CFGMSK:
          s->cfg_mask = val;
          size = ~(val & 0xfffffffe) + 1;
+        /*
+         * Firmware sets this register to E0000001. Why we are not sure,
+         * but the current guess is anything above PCIE_MMCFG_SIZE_MAX is
+         * ignored.
+         */
+        if (size > PCIE_MMCFG_SIZE_MAX) {
+            size = PCIE_MMCFG_SIZE_MAX;
+        }
          pcie_host_mmcfg_update(PCIE_HOST_BRIDGE(s), val & 1, s->cfg_base, size);
          break;
      case PEGPL_MSGBAH: