Hi all,
I tried to use the Wifi driver for Intel wifi chips. When running the driver I get the following messages:
[platform_drv] Error: wifi_drv -> : assignment of PCI device 4:0.0
failed phys=0xf8400000 virt=0x10 and
[wifi_drv] [iwl_trans_pcie_grap_nic_access] *WARN* Timeout waiting for
hardware access (CSR_GP_CNTRL 0xffffffff)
The corresponding run file [1] is explicitly written without any default routes. I can only provide a photo of the log output [2] (at least the visible part) as I don't have any other debug method yet on this hardware.
I thought that this was a missing route but I couldn't find a related message and assigning default routes didn't fix it. The error seems to appear in repos/base-nova/src/core/pd_session_support.cc:assign_pci returning false but I couldn't find out why this happens. Does anyone have an idea?
Regards, Johannes
[1]: https://github.com/jklmnn/base-linux-hw/blob/5824238397139b012785f056c126701... [2]: https://imgur.com/a/DkWTq
Hello Johannes,
just an idea...
On Mon, Dec 04, 2017 at 10:04:36AM +0100, Johannes Kliemann wrote:
[platform_drv] Error: wifi_drv -> : assignment of PCI device 4:0.0
failed phys=0xf8400000 virt=0x10
This physical address seems fairly high. Are you sure you enabled the platform_drv to apply constrains to dataspace allocations? Your start node should contain something like
<resource name="RAM" quantum="4M" constrain_phys="yes"/>
Greets
Hello Christian,
thanks for your help. I have taken the platform_drv configuration from repos/base/run/platform_drv.inc and only changed the routes to be explicit. I rechecked and the runfile does indeed contain the line with constrain_phys [1]. I have also tested several other drivers (ahci_drv, audio_drv, nic_drv) with run files based on the same code and only the wifi driver showed this behaviour.
Regards,
Johannes
[1]: https://github.com/jklmnn/base-linux-hw/blob/5824238397139b012785f056c126701...
Am 04.12.2017 um 17:35 schrieb Christian Helmuth:
Hello Johannes,
just an idea...
On Mon, Dec 04, 2017 at 10:04:36AM +0100, Johannes Kliemann wrote:
[platform_drv] Error: wifi_drv -> : assignment of PCI device 4:0.0
failed phys=0xf8400000 virt=0x10
This physical address seems fairly high. Are you sure you enabled the platform_drv to apply constrains to dataspace allocations? Your start node should contain something like
<resource name="RAM" quantum="4M" constrain_phys="yes"/>
Greets
Hello,
On 04.12.2017 10:04, Johannes Kliemann wrote:
Hi all,
I tried to use the Wifi driver for Intel wifi chips. When running the driver I get the following messages:
[platform_drv] Error: wifi_drv -> : assignment of PCI device 4:0.0
failed phys=0xf8400000 virt=0x10 and
[wifi_drv] [iwl_trans_pcie_grap_nic_access] *WARN* Timeout waiting for
hardware access (CSR_GP_CNTRL 0xffffffff)
The corresponding run file [1] is explicitly written without any default routes. I can only provide a photo of the log output [2] (at least the visible part) as I don't have any other debug method yet on this hardware.
I thought that this was a missing route but I couldn't find a related message and assigning default routes didn't fix it. The error seems to appear in repos/base-nova/src/core/pd_session_support.cc:assign_pci returning false but I couldn't find out why this happens. Does anyone have an idea?
What is or maybe are the value(s) for the variable 'res' in repos/base-nova/src/core/pd_session_support.cc:assign_pci ?
On 04.12.2017 22:14, Alexander Boettcher wrote:
I thought that this was a missing route but I couldn't find a related message and assigning default routes didn't fix it. The error seems to appear in repos/base-nova/src/core/pd_session_support.cc:assign_pci returning false but I couldn't find out why this happens. Does anyone have an idea?
What is or maybe are the value(s) for the variable 'res' in repos/base-nova/src/core/pd_session_support.cc:assign_pci ?
Is my assumption right, that you have no kernel log messages in your setup ? In that case issue #2207 may be possibly of help to you if you get it integrated into your setup.
Alex
Hi Alex,
thanks to your hint with Issue #2207 I was able to get the value of res which is 9. What I find interesting is the fact that it now seems to be able to assign the PCI device as you can see in [1].
[platform_drv] wifi_drv -> : assignment of PCI device 4:0.0 succeeded
Yet everything else seems to behave as before. The timeout errors still appear and there doesn't seem to be any actual connection attempt.
I have built this from [2] which currently incorporates log_core_kernel, NOVA VGA support and benders fix free memory calculation (if this information is useful to you).
Regards, Johannes
[1]: https://imgur.com/a/osaVJ [2]: https://github.com/jklmnn/genode/tree/8a64415f52df2dee30e8fcd279e16d1b18eb2f...
Am 04.12.2017 um 22:14 schrieb Alexander Boettcher:
Hello,
On 04.12.2017 10:04, Johannes Kliemann wrote:
Hi all,
I tried to use the Wifi driver for Intel wifi chips. When running the driver I get the following messages:
[platform_drv] Error: wifi_drv -> : assignment of PCI device 4:0.0
failed phys=0xf8400000 virt=0x10 and
[wifi_drv] [iwl_trans_pcie_grap_nic_access] *WARN* Timeout waiting for
hardware access (CSR_GP_CNTRL 0xffffffff)
The corresponding run file [1] is explicitly written without any default routes. I can only provide a photo of the log output [2] (at least the visible part) as I don't have any other debug method yet on this hardware.
I thought that this was a missing route but I couldn't find a related message and assigning default routes didn't fix it. The error seems to appear in repos/base-nova/src/core/pd_session_support.cc:assign_pci returning false but I couldn't find out why this happens. Does anyone have an idea?
What is or maybe are the value(s) for the variable 'res' in repos/base-nova/src/core/pd_session_support.cc:assign_pci ?
Hi Alex,
as we have discussed yesterday, I tried enabling and disabling Vt-d. This does deterministically affect the PCI assignment so I think that missing Vt-d was the reason for the PCI error.
Yet this doesn't explain the second error that iwl is timing out waiting for the hardware.
Regards, Johannes
Am 05.12.2017 um 17:06 schrieb Johannes Kliemann:
Hi Alex,
thanks to your hint with Issue #2207 I was able to get the value of res which is 9. What I find interesting is the fact that it now seems to be able to assign the PCI device as you can see in [1].
[platform_drv] wifi_drv -> : assignment of PCI device 4:0.0 succeeded
Yet everything else seems to behave as before. The timeout errors still appear and there doesn't seem to be any actual connection attempt.
I have built this from [2] which currently incorporates log_core_kernel, NOVA VGA support and benders fix free memory calculation (if this information is useful to you).
Regards, Johannes
Am 04.12.2017 um 22:14 schrieb Alexander Boettcher:
Hello,
On 04.12.2017 10:04, Johannes Kliemann wrote:
Hi all,
I tried to use the Wifi driver for Intel wifi chips. When running the driver I get the following messages:
[platform_drv] Error: wifi_drv -> : assignment of PCI device 4:0.0
failed phys=0xf8400000 virt=0x10 and
[wifi_drv] [iwl_trans_pcie_grap_nic_access] *WARN* Timeout waiting for
hardware access (CSR_GP_CNTRL 0xffffffff)
The corresponding run file [1] is explicitly written without any default routes. I can only provide a photo of the log output [2] (at least the visible part) as I don't have any other debug method yet on this hardware.
I thought that this was a missing route but I couldn't find a related message and assigning default routes didn't fix it. The error seems to appear in repos/base-nova/src/core/pd_session_support.cc:assign_pci returning false but I couldn't find out why this happens. Does anyone have an idea?
What is or maybe are the value(s) for the variable 'res' in repos/base-nova/src/core/pd_session_support.cc:assign_pci ?
Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ genode-main mailing list genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main
Hi Johannes,
On 07.12.2017 10:11, Johannes Kliemann wrote:
Yet this doesn't explain the second error that iwl is timing out waiting for the hardware.
as discussed off-line, I could reproduce it on one of our machines. May you please give the commits of [0] a spin, which solves the issue for the one machine from us.
Thanks,
Alex.
[0] https://github.com/genodelabs/genode/issues/2620
Regards, Johannes
Am 05.12.2017 um 17:06 schrieb Johannes Kliemann:
Hi Alex,
thanks to your hint with Issue #2207 I was able to get the value of res which is 9. What I find interesting is the fact that it now seems to be able to assign the PCI device as you can see in [1].
[platform_drv] wifi_drv -> : assignment of PCI device 4:0.0 succeeded
Yet everything else seems to behave as before. The timeout errors still appear and there doesn't seem to be any actual connection attempt.
I have built this from [2] which currently incorporates log_core_kernel, NOVA VGA support and benders fix free memory calculation (if this information is useful to you).
Regards, Johannes
Am 04.12.2017 um 22:14 schrieb Alexander Boettcher:
Hello,
On 04.12.2017 10:04, Johannes Kliemann wrote:
Hi all,
I tried to use the Wifi driver for Intel wifi chips. When running the driver I get the following messages:
[platform_drv] Error: wifi_drv -> : assignment of PCI device 4:0.0
failed phys=0xf8400000 virt=0x10 and
[wifi_drv] [iwl_trans_pcie_grap_nic_access] *WARN* Timeout waiting for
hardware access (CSR_GP_CNTRL 0xffffffff)
The corresponding run file [1] is explicitly written without any default routes. I can only provide a photo of the log output [2] (at least the visible part) as I don't have any other debug method yet on this hardware.
I thought that this was a missing route but I couldn't find a related message and assigning default routes didn't fix it. The error seems to appear in repos/base-nova/src/core/pd_session_support.cc:assign_pci returning false but I couldn't find out why this happens. Does anyone have an idea?
What is or maybe are the value(s) for the variable 'res' in repos/base-nova/src/core/pd_session_support.cc:assign_pci ?
Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ genode-main mailing list genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main
Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ genode-main mailing list genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main
Hi Alex,
unfortunately I don't have the device with me currently. I will test those commits next time I get it into my hands ;)
Johannes
Am 18.12.2017 um 14:28 schrieb Alexander Boettcher:
Hi Johannes,
On 07.12.2017 10:11, Johannes Kliemann wrote:
Yet this doesn't explain the second error that iwl is timing out waiting for the hardware.
as discussed off-line, I could reproduce it on one of our machines. May you please give the commits of [0] a spin, which solves the issue for the one machine from us.
Thanks,
Alex.
[0] https://github.com/genodelabs/genode/issues/2620
Regards, Johannes
Am 05.12.2017 um 17:06 schrieb Johannes Kliemann:
Hi Alex,
thanks to your hint with Issue #2207 I was able to get the value of res which is 9. What I find interesting is the fact that it now seems to be able to assign the PCI device as you can see in [1].
[platform_drv] wifi_drv -> : assignment of PCI device 4:0.0 succeeded
Yet everything else seems to behave as before. The timeout errors still appear and there doesn't seem to be any actual connection attempt.
I have built this from [2] which currently incorporates log_core_kernel, NOVA VGA support and benders fix free memory calculation (if this information is useful to you).
Regards, Johannes
Am 04.12.2017 um 22:14 schrieb Alexander Boettcher:
Hello,
On 04.12.2017 10:04, Johannes Kliemann wrote:
Hi all,
I tried to use the Wifi driver for Intel wifi chips. When running the driver I get the following messages:
[platform_drv] Error: wifi_drv -> : assignment of PCI device 4:0.0
failed phys=0xf8400000 virt=0x10 and
[wifi_drv] [iwl_trans_pcie_grap_nic_access] *WARN* Timeout waiting for
hardware access (CSR_GP_CNTRL 0xffffffff)
The corresponding run file [1] is explicitly written without any default routes. I can only provide a photo of the log output [2] (at least the visible part) as I don't have any other debug method yet on this hardware.
I thought that this was a missing route but I couldn't find a related message and assigning default routes didn't fix it. The error seems to appear in repos/base-nova/src/core/pd_session_support.cc:assign_pci returning false but I couldn't find out why this happens. Does anyone have an idea?
What is or maybe are the value(s) for the variable 'res' in repos/base-nova/src/core/pd_session_support.cc:assign_pci ?
Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ genode-main mailing list genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main
Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ genode-main mailing list genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main
Hi Alex,
I have tested your fix and can confirm that it solves the issue.
Regards, Johannes
Am 18.12.2017 um 14:28 schrieb Alexander Boettcher:
Hi Johannes,
On 07.12.2017 10:11, Johannes Kliemann wrote:
Yet this doesn't explain the second error that iwl is timing out waiting for the hardware.
as discussed off-line, I could reproduce it on one of our machines. May you please give the commits of [0] a spin, which solves the issue for the one machine from us.
Thanks,
Alex.
[0] https://github.com/genodelabs/genode/issues/2620
Regards, Johannes
Am 05.12.2017 um 17:06 schrieb Johannes Kliemann:
Hi Alex,
thanks to your hint with Issue #2207 I was able to get the value of res which is 9. What I find interesting is the fact that it now seems to be able to assign the PCI device as you can see in [1].
[platform_drv] wifi_drv -> : assignment of PCI device 4:0.0 succeeded
Yet everything else seems to behave as before. The timeout errors still appear and there doesn't seem to be any actual connection attempt.
I have built this from [2] which currently incorporates log_core_kernel, NOVA VGA support and benders fix free memory calculation (if this information is useful to you).
Regards, Johannes
Am 04.12.2017 um 22:14 schrieb Alexander Boettcher:
Hello,
On 04.12.2017 10:04, Johannes Kliemann wrote:
Hi all,
I tried to use the Wifi driver for Intel wifi chips. When running the driver I get the following messages:
[platform_drv] Error: wifi_drv -> : assignment of PCI device 4:0.0
failed phys=0xf8400000 virt=0x10 and
[wifi_drv] [iwl_trans_pcie_grap_nic_access] *WARN* Timeout waiting for
hardware access (CSR_GP_CNTRL 0xffffffff)
The corresponding run file [1] is explicitly written without any default routes. I can only provide a photo of the log output [2] (at least the visible part) as I don't have any other debug method yet on this hardware.
I thought that this was a missing route but I couldn't find a related message and assigning default routes didn't fix it. The error seems to appear in repos/base-nova/src/core/pd_session_support.cc:assign_pci returning false but I couldn't find out why this happens. Does anyone have an idea?
What is or maybe are the value(s) for the variable 'res' in repos/base-nova/src/core/pd_session_support.cc:assign_pci ?
Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ genode-main mailing list genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main
Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ genode-main mailing list genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main
-
Alexander Boettcher -
Christian Helmuth -
Johannes Kliemann