Skip to contentSkip to conversation

cosmetics for the i6300esb watchdog

PATCH In review Cornelia Huck 4 participants 17 messages Jun 10, 2025 – Jul 30, 2025 Arm/Amlogic Meson SoC support

lore
Share & tools
lore
#0 lore raw

cover cosmetics for the i6300esb watchdog

The i6300esb watchdog device is generic enough to be used on any PCI platform; however, having an "Intel" device on a non-x86 platform looks a bit odd. Just call it a "virtual" device instead, and also fix an old typo for the config option while at it.

Cornelia Huck (2):
watchdog: CONFIG_WDT_IB6300ESB -> CONFIG_WDT_I6300ESB
watchdog: generic name for i6300esb

 hw/watchdog/Kconfig        | 2 +-
 hw/watchdog/meson.build    | 2 +-
 hw/watchdog/wdt_i6300esb.c | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)
Signature · 1 line
2.49.0
#1 lore raw

PATCH 1/2 watchdog: CONFIG_WDT_IB6300ESB -> CONFIG_WDT_I6300ESB

There's no ib6300esb device; let's correct the typo in the config option.

---
 hw/watchdog/Kconfig     | 2 +-
 hw/watchdog/meson.build | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
hw/watchdog/Kconfig +1 -1
index 861fd003341d..1003939e1f33 100644
@@ -2,7 +2,7 @@ config CMSDK_APB_WATCHDOG
     bool
     select PTIMER
 
-config WDT_IB6300ESB
+config WDT_I6300ESB
     bool
     default y if PCI_DEVICES
     depends on PCI
hw/watchdog/meson.build +1 -1
index 15370565bd4d..dafb85dfffc3 100644
@@ -1,7 +1,7 @@
 system_ss.add(files('watchdog.c'))
 system_ss.add(when: 'CONFIG_ALLWINNER_WDT', if_true: files('allwinner-wdt.c'))
 system_ss.add(when: 'CONFIG_CMSDK_APB_WATCHDOG', if_true: files('cmsdk-apb-watchdog.c'))
-system_ss.add(when: 'CONFIG_WDT_IB6300ESB', if_true: files('wdt_i6300esb.c'))
+system_ss.add(when: 'CONFIG_WDT_I6300ESB', if_true: files('wdt_i6300esb.c'))
 system_ss.add(when: 'CONFIG_WDT_IB700', if_true: files('wdt_ib700.c'))
 system_ss.add(when: 'CONFIG_WDT_DIAG288', if_true: files('wdt_diag288.c'))
 system_ss.add(when: 'CONFIG_ASPEED_SOC', if_true: files('wdt_aspeed.c'))
Signature · 1 line
2.49.0
Reviewed-by on v1 1/2
#2 lore raw

Re: [PATCH 1/2] watchdog: CONFIG_WDT_IB6300ESB -> CONFIG_WDT_I6300ESB

Quoted 7 lines · Cornelia Huck

There's no ib6300esb device; let's correct the typo in the config option.

---
 hw/watchdog/Kconfig     | 2 +-
 hw/watchdog/meson.build | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#3 lore raw

PATCH 2/2 watchdog: generic name for i6300esb

In reply to Cornelia Huck

The Intel 6300 Enterprise SouthBridge is a south bridge for a more or less obscure embedded Intel system; however, the i6300esb watchdog device we implement in QEMU is a virtual watchdog device that should work well on any PCI-based machine, is well supported by Linux guests, and used in many examples on how to set up a virtual watchdog.

Let's use "virtual i6300ESB" in the description to make clear that this device will work just fine on non-Intel platforms.

---
 hw/watchdog/wdt_i6300esb.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
hw/watchdog/wdt_i6300esb.c +1 -1
index bb8a2766b69b..5e5d8c48cfd7 100644
@@ -472,7 +472,7 @@ static void i6300esb_class_init(ObjectClass *klass, const void *data)
     device_class_set_legacy_reset(dc, i6300esb_reset);
     dc->vmsd = &vmstate_i6300esb;
     set_bit(DEVICE_CATEGORY_WATCHDOG, dc->categories);
-    dc->desc = "Intel 6300ESB";
+    dc->desc = "virtual i6300ESB";
 }
 
 static const TypeInfo i6300esb_info = {
Signature · 1 line
2.49.0
Reviewed-by on v1 2/2
#4 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 6 lines · Cornelia Huck

The Intel 6300 Enterprise SouthBridge is a south bridge for a more or less obscure embedded Intel system; however, the i6300esb watchdog device we implement in QEMU is a virtual watchdog device that should work well on any PCI-based machine, is well supported by Linux guests, and used in many examples on how to set up a virtual watchdog.

Quoted 7 lines

Let's use "virtual i6300ESB" in the description to make clear that this device will work just fine on non-Intel platforms.

---
 hw/watchdog/wdt_i6300esb.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

I'm not entirely sold on the idea that this is needed, but at the same time I won't object so

Quoted 18 lines
hw/watchdog/wdt_i6300esb.c +1 -1
index bb8a2766b69b..5e5d8c48cfd7 100644
@@ -472,7 +472,7 @@ static void i6300esb_class_init(ObjectClass *klass, const void *data)
     device_class_set_legacy_reset(dc, i6300esb_reset);
     dc->vmsd = &vmstate_i6300esb;
     set_bit(DEVICE_CATEGORY_WATCHDOG, dc->categories);
-    dc->desc = "Intel 6300ESB";
+    dc->desc = "virtual i6300ESB";
 }
 
 static const TypeInfo i6300esb_info = {
Signature · 1 line
2.49.0

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#5 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 15 lines · Daniel P. Berrangé
Quoted 6 lines · Cornelia Huck

The Intel 6300 Enterprise SouthBridge is a south bridge for a more or less obscure embedded Intel system; however, the i6300esb watchdog device we implement in QEMU is a virtual watchdog device that should work well on any PCI-based machine, is well supported by Linux guests, and used in many examples on how to set up a virtual watchdog.

Quoted 7 lines

Let's use "virtual i6300ESB" in the description to make clear that this device will work just fine on non-Intel platforms.

---
 hw/watchdog/wdt_i6300esb.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 4 lines

I'm not entirely sold on the idea that this is needed, but at the same time I won't object so

I would argue that this change is incorrect.

While the QEMU device can be used for non-x86 VMs, it *is* faithfully modelled after an Intel part, and the guest OS will recognize it as such:

  # lspci | grep 6300
  07:01.0 System peripheral: Intel Corporation 6300ESB Watchdog Timer

What we actually need to do is create a new QEMU device with distinct PCI IDs, same as we've done in the past for qemu-xhci, pcie-root-port and pcie-to-pci-bridge.

That will take a lot longer to integrate throughout the stack and be supported across the various guest OS, but it's the only approach that eventually leads to truly Intel-free non-x86 VMs.

Signature · 1 line
Andrea Bolognani / Red Hat / Virtualization
#6 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 30 lines · Andrea Bolognani
Quoted 15 lines · Daniel P. Berrangé
Quoted 6 lines · Cornelia Huck

The Intel 6300 Enterprise SouthBridge is a south bridge for a more or less obscure embedded Intel system; however, the i6300esb watchdog device we implement in QEMU is a virtual watchdog device that should work well on any PCI-based machine, is well supported by Linux guests, and used in many examples on how to set up a virtual watchdog.

Quoted 7 lines

Let's use "virtual i6300ESB" in the description to make clear that this device will work just fine on non-Intel platforms.

---
 hw/watchdog/wdt_i6300esb.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 4 lines

I'm not entirely sold on the idea that this is needed, but at the same time I won't object so

I would argue that this change is incorrect.

While the QEMU device can be used for non-x86 VMs, it *is* faithfully modelled after an Intel part, and the guest OS will recognize it as such:

  # lspci | grep 6300
  07:01.0 System peripheral: Intel Corporation 6300ESB Watchdog Timer
Quoted 7 lines

What we actually need to do is create a new QEMU device with distinct PCI IDs, same as we've done in the past for qemu-xhci, pcie-root-port and pcie-to-pci-bridge.

That will take a lot longer to integrate throughout the stack and be supported across the various guest OS, but it's the only approach that eventually leads to truly Intel-free non-x86 VMs.

Hmm. So
- request a new PCI id (probably in the PCI_DEVICE_ID_REDHAT_* space)
- restructure to have two devices base off the same core functionality
- teach guest operating systems about the new device
- teach management software like libvirt about the new device

Not sure how fast we can get an ID (or even how to go about it.) The second step should be reasonably easy. The third step is the most complex one, but at least teaching Linux should hopefully be easy enough, and existing guest operating systems could continue to use the existing device. The last step is probably not that bad.

I can start down that path, if we have some consensus that this is the right way to handle this.

I'd still argue that patch 1 should be applied regardless :)

#7 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 52 lines · Cornelia Huck
Quoted 30 lines · Andrea Bolognani
Quoted 15 lines · Daniel P. Berrangé
Quoted 6 lines · Cornelia Huck

The Intel 6300 Enterprise SouthBridge is a south bridge for a more or less obscure embedded Intel system; however, the i6300esb watchdog device we implement in QEMU is a virtual watchdog device that should work well on any PCI-based machine, is well supported by Linux guests, and used in many examples on how to set up a virtual watchdog.

Quoted 7 lines

Let's use "virtual i6300ESB" in the description to make clear that this device will work just fine on non-Intel platforms.

---
 hw/watchdog/wdt_i6300esb.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 4 lines

I'm not entirely sold on the idea that this is needed, but at the same time I won't object so

I would argue that this change is incorrect.

While the QEMU device can be used for non-x86 VMs, it *is* faithfully modelled after an Intel part, and the guest OS will recognize it as such:

  # lspci | grep 6300
  07:01.0 System peripheral: Intel Corporation 6300ESB Watchdog Timer
Quoted 7 lines

What we actually need to do is create a new QEMU device with distinct PCI IDs, same as we've done in the past for qemu-xhci, pcie-root-port and pcie-to-pci-bridge.

That will take a lot longer to integrate throughout the stack and be supported across the various guest OS, but it's the only approach that eventually leads to truly Intel-free non-x86 VMs.

Hmm. So
- request a new PCI id (probably in the PCI_DEVICE_ID_REDHAT_* space)
- restructure to have two devices base off the same core functionality
- teach guest operating systems about the new device
- teach management software like libvirt about the new device

Not sure how fast we can get an ID (or even how to go about it.) The second step should be reasonably easy. The third step is the most complex one, but at least teaching Linux should hopefully be easy enough, and existing guest operating systems could continue to use the existing device. The last step is probably not that bad.

Quoted 4 lines

I can start down that path, if we have some consensus that this is the right way to handle this.

I'd still argue that patch 1 should be applied regardless :)

This sounds like a hell of alot of busy work to fix a problem that, IIUC, does not actually exist from a functional POV - it is merely a perception issue that people might be put off by the "Intel 6300ESB" names.

IMHO a better use of time is to expand documentation to clarify this is just fine for all PCI architectures, and change nothing in either QEMU or guest kernels.

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#8 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 18 lines · Daniel P. Berrangé
Quoted 12 lines · Cornelia Huck

Hmm. So
- request a new PCI id (probably in the PCI_DEVICE_ID_REDHAT_* space)
- restructure to have two devices base off the same core functionality
- teach guest operating systems about the new device
- teach management software like libvirt about the new device

Not sure how fast we can get an ID (or even how to go about it.) The second step should be reasonably easy. The third step is the most complex one, but at least teaching Linux should hopefully be easy enough, and existing guest operating systems could continue to use the existing device. The last step is probably not that bad.

Quoted 4 lines

I can start down that path, if we have some consensus that this is the right way to handle this.

I'd still argue that patch 1 should be applied regardless :)

Quoted 7 lines

This sounds like a hell of alot of busy work to fix a problem that, IIUC, does not actually exist from a functional POV - it is merely a perception issue that people might be put off by the "Intel 6300ESB" names.

IMHO a better use of time is to expand documentation to clarify this is just fine for all PCI architectures, and change nothing in either QEMU or guest kernels.

Agreed that it's not the most high-reward endeavor, but IIRC users were getting genuinely confused and annoyed by the string "Intel" showing up in their aarch64 guests.

You can point them to documentation over and over again, or you can work to prevent the confusion/annoyance from showing up in the first place. Which of the two approaches is a better use of anyone's time is up for debate.

I for one am grateful that someone put the time in all those years ago and, as a result, PCI and USB controllers don't suffer from the problem today. Ultimately it's up to Connie though.

Signature · 1 line
Andrea Bolognani / Red Hat / Virtualization
#9 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 23 lines · Andrea Bolognani
Quoted 22 lines · Daniel P. Berrangé
Quoted 12 lines · Cornelia Huck

Hmm. So
- request a new PCI id (probably in the PCI_DEVICE_ID_REDHAT_* space)
- restructure to have two devices base off the same core functionality
- teach guest operating systems about the new device
- teach management software like libvirt about the new device

Not sure how fast we can get an ID (or even how to go about it.) The second step should be reasonably easy. The third step is the most complex one, but at least teaching Linux should hopefully be easy enough, and existing guest operating systems could continue to use the existing device. The last step is probably not that bad.

Quoted 4 lines

I can start down that path, if we have some consensus that this is the right way to handle this.

I'd still argue that patch 1 should be applied regardless :)

This sounds like a hell of alot of busy work to fix a problem that, IIUC, does not actually exist from a functional POV - it is merely a perception issue that people might be put off by the "Intel 6300ESB" names.

Quoted 7 lines
Quoted 3 lines

IMHO a better use of time is to expand documentation to clarify this is just fine for all PCI architectures, and change nothing in either QEMU or guest kernels.

Agreed that it's not the most high-reward endeavor, but IIRC users were getting genuinely confused and annoyed by the string "Intel" showing up in their aarch64 guests.

So be it, that's far from the only wierd thing in virt.

Quoted 8 lines

You can point them to documentation over and over again, or you can work to prevent the confusion/annoyance from showing up in the first place. Which of the two approaches is a better use of anyone's time is up for debate.

I for one am grateful that someone put the time in all those years ago and, as a result, PCI and USB controllers don't suffer from the problem today. Ultimately it's up to Connie though.

The PCI/USB controller situation is not the same tradeoff though. Those guest kernel drivers will identify and attach to these two controllers regardless of their PCI vendor/product, via the PCI class property. In that case changing the PCI ID and other device metadata in QEMU is cheap as it has no negative impact on guest OS driver compibility.

In the case of 6300ESB though the guest driver is tied directly to the currently used PCI device product/vendor ID.

If we change this then we have actually created new functional problems with guest/QEMU compatibility, in order to placate a non-functional problem. That is not a good thing.

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#10 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 28 lines · Daniel P. Berrangé
Quoted 27 lines · Andrea Bolognani
Quoted 25 lines · Daniel P. Berrangé
Quoted 12 lines · Cornelia Huck

Hmm. So
- request a new PCI id (probably in the PCI_DEVICE_ID_REDHAT_* space)
- restructure to have two devices base off the same core functionality
- teach guest operating systems about the new device
- teach management software like libvirt about the new device

Not sure how fast we can get an ID (or even how to go about it.) The second step should be reasonably easy. The third step is the most complex one, but at least teaching Linux should hopefully be easy enough, and existing guest operating systems could continue to use the existing device. The last step is probably not that bad.

Quoted 4 lines

I can start down that path, if we have some consensus that this is the right way to handle this.

I'd still argue that patch 1 should be applied regardless :)

This sounds like a hell of alot of busy work to fix a problem that, IIUC, does not actually exist from a functional POV - it is merely a perception issue that people might be put off by the "Intel 6300ESB" names.

IMHO a better use of time is to expand documentation to clarify this is just fine for all PCI architectures, and change nothing in either QEMU or guest kernels.

Quoted 5 lines
Quoted 3 lines

Agreed that it's not the most high-reward endeavor, but IIRC users were getting genuinely confused and annoyed by the string "Intel" showing up in their aarch64 guests.

So be it, that's far from the only wierd thing in virt.

So I'm wondering what the general sentiment is towards having an "Intel" device show up in your non-Intel guest. This device is generic enough to be included in PCI_DEVICES, and should be usable on any of those platforms.

Linux treats this as a generic PCI driver as well. Do we know what Windows on non-x86 expects? The *BSDs? Others?

Quoted 17 lines
Quoted 9 lines

You can point them to documentation over and over again, or you can work to prevent the confusion/annoyance from showing up in the first place. Which of the two approaches is a better use of anyone's time is up for debate.

I for one am grateful that someone put the time in all those years ago and, as a result, PCI and USB controllers don't suffer from the problem today. Ultimately it's up to Connie though.

The PCI/USB controller situation is not the same tradeoff though. Those guest kernel drivers will identify and attach to these two controllers regardless of their PCI vendor/product, via the PCI class property. In that case changing the PCI ID and other device metadata in QEMU is cheap as it has no negative impact on guest OS driver compibility.

Quoted 6 lines

In the case of 6300ESB though the guest driver is tied directly to the currently used PCI device product/vendor ID.

If we change this then we have actually created new functional problems with guest/QEMU compatibility, in order to placate a non-functional problem. That is not a good thing.

I don't think the suggestion was to disable the existing driver on non-Intel setups, but to add a more generic one. Still, more work to get this actually propagated into guests than doing the change in QEMU. Before I start down that route, I'd like to know whether the issue is actually big enough to make investing time there worth it.

#11 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 34 lines · Cornelia Huck
Quoted 32 lines · Daniel P. Berrangé
Quoted 30 lines · Andrea Bolognani
Quoted 18 lines · Daniel P. Berrangé
Quoted 12 lines · Cornelia Huck

Hmm. So
- request a new PCI id (probably in the PCI_DEVICE_ID_REDHAT_* space)
- restructure to have two devices base off the same core functionality
- teach guest operating systems about the new device
- teach management software like libvirt about the new device

Not sure how fast we can get an ID (or even how to go about it.) The second step should be reasonably easy. The third step is the most complex one, but at least teaching Linux should hopefully be easy enough, and existing guest operating systems could continue to use the existing device. The last step is probably not that bad.

Quoted 4 lines

I can start down that path, if we have some consensus that this is the right way to handle this.

I'd still argue that patch 1 should be applied regardless :)

Quoted 7 lines

This sounds like a hell of alot of busy work to fix a problem that, IIUC, does not actually exist from a functional POV - it is merely a perception issue that people might be put off by the "Intel 6300ESB" names.

IMHO a better use of time is to expand documentation to clarify this is just fine for all PCI architectures, and change nothing in either QEMU or guest kernels.

Agreed that it's not the most high-reward endeavor, but IIRC users were getting genuinely confused and annoyed by the string "Intel" showing up in their aarch64 guests.

Quoted 6 lines
Quoted 1 line

So be it, that's far from the only wierd thing in virt.

So I'm wondering what the general sentiment is towards having an "Intel" device show up in your non-Intel guest. This device is generic enough to be included in PCI_DEVICES, and should be usable on any of those platforms.

That is not all that different from an "Intel" device showing up in my "AMD" guest, or that my "modern" virt guest is exposing an Intel machine board from 2005, or that what we advertize as a "q35" machine is actually a "p35" machine, or many other wierd virt things.

Quoted 2 lines

Linux treats this as a generic PCI driver as well. Do we know what Windows on non-x86 expects? The *BSDs? Others?

I'm not sure offhand.

Quoted 23 lines
Quoted 16 lines
Quoted 8 lines

You can point them to documentation over and over again, or you can work to prevent the confusion/annoyance from showing up in the first place. Which of the two approaches is a better use of anyone's time is up for debate.

I for one am grateful that someone put the time in all those years ago and, as a result, PCI and USB controllers don't suffer from the problem today. Ultimately it's up to Connie though.

The PCI/USB controller situation is not the same tradeoff though. Those guest kernel drivers will identify and attach to these two controllers regardless of their PCI vendor/product, via the PCI class property. In that case changing the PCI ID and other device metadata in QEMU is cheap as it has no negative impact on guest OS driver compibility.

Quoted 6 lines

In the case of 6300ESB though the guest driver is tied directly to the currently used PCI device product/vendor ID.

If we change this then we have actually created new functional problems with guest/QEMU compatibility, in order to placate a non-functional problem. That is not a good thing.

Quoted 5 lines

I don't think the suggestion was to disable the existing driver on non-Intel setups, but to add a more generic one. Still, more work to get this actually propagated into guests than doing the change in QEMU. Before I start down that route, I'd like to know whether the issue is actually big enough to make investing time there worth it.

If we're a mmgmt app provisioning a guest, we have to choose what watchdog to create - either the old one which works everywhere that currently has a driver, or the new one will will work in far fewer places. We'll have to wire up guest OS info about watchdogs into osinfo, and then wire up all the mgmt apps to query this and take action based off it. All possible, but it still feels like a huge waste of time to me.

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#12 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 33 lines · Daniel P. Berrangé
Quoted 26 lines · Cornelia Huck
Quoted 17 lines · Daniel P. Berrangé
Quoted 8 lines · Andrea Bolognani

You can point them to documentation over and over again, or you can work to prevent the confusion/annoyance from showing up in the first place. Which of the two approaches is a better use of anyone's time is up for debate.

I for one am grateful that someone put the time in all those years ago and, as a result, PCI and USB controllers don't suffer from the problem today. Ultimately it's up to Connie though.

The PCI/USB controller situation is not the same tradeoff though. Those guest kernel drivers will identify and attach to these two controllers regardless of their PCI vendor/product, via the PCI class property. In that case changing the PCI ID and other device metadata in QEMU is cheap as it has no negative impact on guest OS driver compibility.

Quoted 6 lines

In the case of 6300ESB though the guest driver is tied directly to the currently used PCI device product/vendor ID.

If we change this then we have actually created new functional problems with guest/QEMU compatibility, in order to placate a non-functional problem. That is not a good thing.

Quoted 5 lines

I don't think the suggestion was to disable the existing driver on non-Intel setups, but to add a more generic one. Still, more work to get this actually propagated into guests than doing the change in QEMU. Before I start down that route, I'd like to know whether the issue is actually big enough to make investing time there worth it.

Quoted 7 lines

If we're a mmgmt app provisioning a guest, we have to choose what watchdog to create - either the old one which works everywhere that currently has a driver, or the new one will will work in far fewer places. We'll have to wire up guest OS info about watchdogs into osinfo, and then wire up all the mgmt apps to query this and take action based off it. All possible, but it still feels like a huge waste of time to me.

The fact that the device is something emulated and not the Intel hardware device is actually visible to the guest:

00:02.0 System peripheral: Intel Corporation 6300ESB Watchdog Timer
Subsystem: Red Hat, Inc. QEMU Virtual Machine
Flags: fast devsel
Memory at 10804000 (32-bit, non-prefetchable) [size=16] Kernel driver in use: i6300ESB timer
Kernel modules: i6300esb

(lspci -v so unfortunately not immediately obvious, but still)

AFAIK the BSDs do not have a driver for this device at the moment -- and given what turns up when searching for i6300ESB, someone implementing a driver is far more likely to pick the exising PCI ID.

Windows would also need some dance according to Yan's mail, for unclear benefits.

On the whole, I think this is not really worth the hassle, we can simply keep this device as-is.

#13 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 53 lines · Cornelia Huck
Quoted 33 lines · Daniel P. Berrangé
Quoted 26 lines · Cornelia Huck
Quoted 17 lines · Daniel P. Berrangé
Quoted 8 lines · Andrea Bolognani

You can point them to documentation over and over again, or you can work to prevent the confusion/annoyance from showing up in the first place. Which of the two approaches is a better use of anyone's time is up for debate.

I for one am grateful that someone put the time in all those years ago and, as a result, PCI and USB controllers don't suffer from the problem today. Ultimately it's up to Connie though.

The PCI/USB controller situation is not the same tradeoff though. Those guest kernel drivers will identify and attach to these two controllers regardless of their PCI vendor/product, via the PCI class property. In that case changing the PCI ID and other device metadata in QEMU is cheap as it has no negative impact on guest OS driver compibility.

Quoted 6 lines

In the case of 6300ESB though the guest driver is tied directly to the currently used PCI device product/vendor ID.

If we change this then we have actually created new functional problems with guest/QEMU compatibility, in order to placate a non-functional problem. That is not a good thing.

Quoted 5 lines

I don't think the suggestion was to disable the existing driver on non-Intel setups, but to add a more generic one. Still, more work to get this actually propagated into guests than doing the change in QEMU. Before I start down that route, I'd like to know whether the issue is actually big enough to make investing time there worth it.

Quoted 7 lines

If we're a mmgmt app provisioning a guest, we have to choose what watchdog to create - either the old one which works everywhere that currently has a driver, or the new one will will work in far fewer places. We'll have to wire up guest OS info about watchdogs into osinfo, and then wire up all the mgmt apps to query this and take action based off it. All possible, but it still feels like a huge waste of time to me.

The fact that the device is something emulated and not the Intel hardware device is actually visible to the guest:

00:02.0 System peripheral: Intel Corporation 6300ESB Watchdog Timer
Subsystem: Red Hat, Inc. QEMU Virtual Machine
Flags: fast devsel
Memory at 10804000 (32-bit, non-prefetchable) [size=16] Kernel driver in use: i6300ESB timer
Kernel modules: i6300esb

Quoted 5 lines

(lspci -v so unfortunately not immediately obvious, but still)

AFAIK the BSDs do not have a driver for this device at the moment -- and given what turns up when searching for i6300ESB, someone implementing a driver is far more likely to pick the exising PCI ID.

I see vague references (with unfortunately 404 links) to FreeBSD supporting some ICH watchdogs, which might mean it is compatible with the q35 built-in watchdog that all x86 q35 machines get by default. That wouldn't help non-x86 BSD though.

Quoted 2 lines

Windows would also need some dance according to Yan's mail, for unclear benefits.

Off-list, Richard Jones pointed to the ACPI Watchdog WADT specification from Microsoft which appears to the most viable solution for Windows guests - at least from x86 POV, but hopefully any future Wndows aarch64 too:

https://download.microsoft.com/download/a/f/7/af7777e5-7dcd-4800-8a0a-b18336565f5b/HardwareWDTSpec.doc

The ACPI watchdog sounds like potentially the best bet for a working solution across Linux and Windows, on any arch that does ACPI.... if we can just find someone to write a QEMU driver for it....

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#14 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

Quoted 56 lines · Daniel P. Berrangé
Quoted 55 lines · Cornelia Huck
Quoted 33 lines · Daniel P. Berrangé
Quoted 26 lines · Cornelia Huck
Quoted 17 lines · Daniel P. Berrangé
Quoted 8 lines · Andrea Bolognani

You can point them to documentation over and over again, or you can work to prevent the confusion/annoyance from showing up in the first place. Which of the two approaches is a better use of anyone's time is up for debate.

I for one am grateful that someone put the time in all those years ago and, as a result, PCI and USB controllers don't suffer from the problem today. Ultimately it's up to Connie though.

The PCI/USB controller situation is not the same tradeoff though. Those guest kernel drivers will identify and attach to these two controllers regardless of their PCI vendor/product, via the PCI class property. In that case changing the PCI ID and other device metadata in QEMU is cheap as it has no negative impact on guest OS driver compibility.

Quoted 6 lines

In the case of 6300ESB though the guest driver is tied directly to the currently used PCI device product/vendor ID.

If we change this then we have actually created new functional problems with guest/QEMU compatibility, in order to placate a non-functional problem. That is not a good thing.

Quoted 5 lines

I don't think the suggestion was to disable the existing driver on non-Intel setups, but to add a more generic one. Still, more work to get this actually propagated into guests than doing the change in QEMU. Before I start down that route, I'd like to know whether the issue is actually big enough to make investing time there worth it.

Quoted 7 lines

If we're a mmgmt app provisioning a guest, we have to choose what watchdog to create - either the old one which works everywhere that currently has a driver, or the new one will will work in far fewer places. We'll have to wire up guest OS info about watchdogs into osinfo, and then wire up all the mgmt apps to query this and take action based off it. All possible, but it still feels like a huge waste of time to me.

The fact that the device is something emulated and not the Intel hardware device is actually visible to the guest:

00:02.0 System peripheral: Intel Corporation 6300ESB Watchdog Timer
Subsystem: Red Hat, Inc. QEMU Virtual Machine
Flags: fast devsel
Memory at 10804000 (32-bit, non-prefetchable) [size=16] Kernel driver in use: i6300ESB timer
Kernel modules: i6300esb

(lspci -v so unfortunately not immediately obvious, but still)

Quoted 8 lines
Quoted 3 lines

AFAIK the BSDs do not have a driver for this device at the moment -- and given what turns up when searching for i6300ESB, someone implementing a driver is far more likely to pick the exising PCI ID.

I see vague references (with unfortunately 404 links) to FreeBSD supporting some ICH watchdogs, which might mean it is compatible with the q35 built-in watchdog that all x86 q35 machines get by default. That wouldn't help non-x86 BSD though.

I found https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=259673 -- I'm not sure if anyone has tried to implement something cross-arch.

Quoted 9 lines
Quoted 3 lines

Windows would also need some dance according to Yan's mail, for unclear benefits.

Off-list, Richard Jones pointed to the ACPI Watchdog WADT specification from Microsoft which appears to the most viable solution for Windows guests - at least from x86 POV, but hopefully any future Wndows aarch64 too:

Quoted 5 lines

https://download.microsoft.com/download/a/f/7/af7777e5-7dcd-4800-8a0a-b18336565f5b/HardwareWDTSpec.doc

The ACPI watchdog sounds like potentially the best bet for a working solution across Linux and Windows, on any arch that does ACPI.... if we can just find someone to write a QEMU driver for it....

Yes, if someone(tm) wrote a QEMU device for that, we'd have something that should be quite useful across archs (well, excluding s390, which has its own watchdog anyway.)

Given all of that, I still think that leaving i6300ESB as-is is the right thing, with a possible option of implementing the WADT watchdog.

#15 lore raw

Re: [PATCH 2/2] watchdog: generic name for i6300esb

In reply to Cornelia Huck
Quoted 37 lines · Cornelia Huck
Quoted 28 lines · Daniel P. Berrangé
Quoted 27 lines · Andrea Bolognani
Quoted 25 lines · Daniel P. Berrangé
Quoted 12 lines · Cornelia Huck

Hmm. So
- request a new PCI id (probably in the PCI_DEVICE_ID_REDHAT_* space)
- restructure to have two devices base off the same core functionality
- teach guest operating systems about the new device
- teach management software like libvirt about the new device

Not sure how fast we can get an ID (or even how to go about it.) The second step should be reasonably easy. The third step is the most complex one, but at least teaching Linux should hopefully be easy enough, and existing guest operating systems could continue to use the existing device. The last step is probably not that bad.

Quoted 4 lines

I can start down that path, if we have some consensus that this is the right way to handle this.

I'd still argue that patch 1 should be applied regardless :)

This sounds like a hell of alot of busy work to fix a problem that, IIUC, does not actually exist from a functional POV - it is merely a perception issue that people might be put off by the "Intel 6300ESB" names.

IMHO a better use of time is to expand documentation to clarify this is just fine for all PCI architectures, and change nothing in either QEMU or guest kernels.

Quoted 5 lines
Quoted 3 lines

Agreed that it's not the most high-reward endeavor, but IIRC users were getting genuinely confused and annoyed by the string "Intel" showing up in their aarch64 guests.

So be it, that's far from the only wierd thing in virt.

Quoted 7 lines

So I'm wondering what the general sentiment is towards having an "Intel" device show up in your non-Intel guest. This device is generic enough to be included in PCI_DEVICES, and should be usable on any of those platforms.

Linux treats this as a generic PCI driver as well. Do we know what Windows on non-x86 expects? The *BSDs? Others?

Eric just asked me about the i6300ESB on ARM and Windows. Windows has i6300ESB device PNP ID in machine.inf:
%WatchdogTimer_Desc% = NO_DRV, PCI\VEN_8086&DEV_25AB NO_DRV means that there is no device driver installed for the device, but the device will have some "Friendly Name" in the device manager and will be treated as installed device. Users could then install Intel's driver.

Changing the device ID, will make this device as not-installed from Windows perspective. If such device will be included by default for some machine types, those machine types will not be able to pass Microsoft SVVP certification because the OS will have a device without a driver. The mitigation is to keep and certify our own dummy INF, which is a grey zone as Intel should have their own driver certified with MS.
This is very similar to the discussion we had about AMD IOMMU PCIe endpoint PNP Id.

Best regards,
Yan.

Quoted 32 lines
Quoted 18 lines
Quoted 10 lines

You can point them to documentation over and over again, or you can work to prevent the confusion/annoyance from showing up in the first place. Which of the two approaches is a better use of anyone's time is up for debate.

I for one am grateful that someone put the time in all those years ago and, as a result, PCI and USB controllers don't suffer from the problem today. Ultimately it's up to Connie though.

The PCI/USB controller situation is not the same tradeoff though. Those guest kernel drivers will identify and attach to these two controllers regardless of their PCI vendor/product, via the PCI class property. In that case changing the PCI ID and other device metadata in QEMU is cheap as it has no negative impact on guest OS driver compibility.

Quoted 6 lines

In the case of 6300ESB though the guest driver is tied directly to the currently used PCI device product/vendor ID.

If we change this then we have actually created new functional problems with guest/QEMU compatibility, in order to placate a non-functional problem. That is not a good thing.

I don't think the suggestion was to disable the existing driver on non-Intel setups, but to add a more generic one. Still, more work to get this actually propagated into guests than doing the change in QEMU. Before I start down that route, I'd like to know whether the issue is actually big enough to make investing time there worth it.

#16 lore raw
In reply to Cornelia Huck
Quoted 5 lines · Cornelia Huck

The i6300esb watchdog device is generic enough to be used on any PCI platform; however, having an "Intel" device on a non-x86 platform looks a bit odd. Just call it a "virtual" device instead, and also fix an old typo for the config option while at it.

Quoted 8 lines

Cornelia Huck (2):
watchdog: CONFIG_WDT_IB6300ESB -> CONFIG_WDT_I6300ESB
watchdog: generic name for i6300esb

 hw/watchdog/Kconfig        | 2 +-
 hw/watchdog/meson.build    | 2 +-
 hw/watchdog/wdt_i6300esb.c | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)

Friendly ping -- any sentiment regarding those patches?