Panic with debug kernel

[netfenceventu]

For reasons beyond the scope of this driver, I compiled a 12.2/amd64 kernel with the following:

options INVARIANTS
options INVARIANT_SUPPORT
options WITNESS
options WITNESS_SKIPSPIN
options DEBUG_LOCKS
options DIAGNOSTIC

Booting with it makes this driver panic with:
_mtx_lock_sleep: recursed on non-recursive mutex re0 0 if_re.c:7085

[kostikbel]

There is no useful information in your report. Provide all the data that was printed by kernel, on panic.

[netfenceventu]

On 6/8/21 7:43 PM, Konstantin Belousov wrote: There is no useful information in your report. Provide all the data that was printed by kernel, on panic.

I understand. Unfortunately the panic happened on boot, before the filesystem was even mounted, so nothing was saved. I have a (not so good) picture of the screen: I can try to transcribe portions of it, but the numbers might not be that accurate (i.e. I cannot read them sharply). Would the stack trace suffice? If someone has a better suggestion, I'm open.

[kostikbel]

Load the module after the machine in multiuser.

If screen shots are readable, try posting that.

[netfenceventu]

On 6/9/21 2:05 PM, Konstantin Belousov wrote: Load the module after the machine in multiuser.

Curiously, if I boot single user, then load this module, it does not panic. In multi user mode, it does, but still saves no core (and this machine normally does).

If screen shots are readable, try posting that.

Here it is. HTH. Also, I found this in the logs:

Jun 12 08:39:58 alamar kernel: re0: <Realtek PCIe GbE Family Controller> port 0xe000-0xe0ff mem 0xf7c04000-0xf7c04fff,0xf7c00000-0xf7c03fff irq 17 at device 0.0 on pci2 Jun 12 08:39:58 alamar kernel: re0: Using Memory Mapping! Jun 12 08:39:58 alamar kernel: re0: Using 1 MSI-X message Jun 12 08:39:58 alamar kernel: re0: version:1.96.04 Jun 12 08:39:58 alamar kernel: re0: Ethernet address: d4:3d:7e:b4:74:15 Jun 12 08:39:58 alamar kernel: Jun 12 08:39:58 alamar kernel: This product is covered by one or more of the following patents: Jun 12 08:39:58 alamar kernel: US6,570,884, US6,115,776, and US6,327,625. Jun 12 08:39:58 alamar kernel: bus_dma_tag_create with the following non-sleepable locks held: Jun 12 08:39:58 alamar kernel: exclusive sleep mutex re0 (network driver) r = 0 (0xfffffe0004f29150) locked @ if_re.c:5118 Jun 12 08:39:58 alamar kernel: stack backtrace: Jun 12 08:39:58 alamar kernel: #0 0xffffffff80636c11 at witness_debugger+0x71 Jun 12 08:39:58 alamar kernel: #1 0xffffffff80637b8d at witness_warn+0x40d Jun 12 08:39:58 alamar kernel: #2 0xffffffff808da4ce at bus_dma_tag_create+0x7e Jun 12 08:39:58 alamar kernel: #3 0xffffffff813f9f17 at re_alloc_buf+0x57 Jun 12 08:39:58 alamar kernel: #4 0xffffffff813d4ad1 at re_attach+0x3061 Jun 12 08:39:58 alamar kernel: #5 0xffffffff806043fa at device_attach+0x3ca Jun 12 08:39:58 alamar kernel: #6 0xffffffff8060562a at bus_generic_attach+0x4a Jun 12 08:39:58 alamar kernel: #7 0xffffffff804185d0 at pci_attach+0xe0 Jun 12 08:39:58 alamar kernel: #8 0xffffffff80888245 at acpi_pci_attach+0x15 Jun 12 08:39:58 alamar kernel: #9 0xffffffff806043fa at device_attach+0x3ca Jun 12 08:39:58 alamar kernel: #10 0xffffffff8060562a at bus_generic_attach+0x4a Jun 12 08:39:58 alamar kernel: #11 0xffffffff8088bed0 at acpi_pcib_pci_attach+0xa0 Jun 12 08:39:58 alamar kernel: #12 0xffffffff806043fa at device_attach+0x3ca Jun 12 08:39:58 alamar kernel: #13 0xffffffff8060562a at bus_generic_attach+0x4a Jun 12 08:39:58 alamar kernel: #14 0xffffffff804185d0 at pci_attach+0xe0 Jun 12 08:39:58 alamar kernel: #15 0xffffffff80888245 at acpi_pci_attach+0x15 Jun 12 08:39:58 alamar kernel: #16 0xffffffff806043fa at device_attach+0x3ca Jun 12 08:39:58 alamar kernel: #17 0xffffffff8060562a at bus_generic_attach+0x4a Jun 12 08:39:58 alamar kernel: bus_dma_tag_create with the following non-sleepable locks held: Jun 12 08:39:58 alamar kernel: exclusive sleep mutex re0 (network driver) r = 0 (0xfffffe0004f29150) locked @ if_re.c:5118 Jun 12 08:39:58 alamar kernel: stack backtrace: Jun 12 08:39:58 alamar kernel: #0 0xffffffff80636c11 at witness_debugger+0x71 Jun 12 08:39:58 alamar kernel: #1 0xffffffff80637b8d at witness_warn+0x40d Jun 12 08:39:58 alamar kernel: #2 0xffffffff808da4ce at bus_dma_tag_create+0x7e Jun 12 08:39:58 alamar kernel: #3 0xffffffff813f9f65 at re_alloc_buf+0xa5 Jun 12 08:39:58 alamar kernel: #4 0xffffffff813d4ad1 at re_attach+0x3061 Jun 12 08:39:58 alamar kernel: #5 0xffffffff806043fa at device_attach+0x3ca Jun 12 08:39:58 alamar kernel: #6 0xffffffff8060562a at bus_generic_attach+0x4a Jun 12 08:39:58 alamar kernel: #7 0xffffffff804185d0 at pci_attach+0xe0 Jun 12 08:39:58 alamar kernel: #8 0xffffffff80888245 at acpi_pci_attach+0x15 Jun 12 08:39:58 alamar kernel: #9 0xffffffff806043fa at device_attach+0x3ca Jun 12 08:39:58 alamar kernel: #10 0xffffffff8060562a at bus_generic_attach+0x4a Jun 12 08:39:58 alamar kernel: #11 0xffffffff8088bed0 at acpi_pcib_pci_attach+0xa0 Jun 12 08:39:58 alamar kernel: #12 0xffffffff806043fa at device_attach+0x3ca Jun 12 08:39:58 alamar kernel: #13 0xffffffff8060562a at bus_generic_attach+0x4a Jun 12 08:39:58 alamar kernel: #14 0xffffffff804185d0 at pci_attach+0xe0 Jun 12 08:39:58 alamar kernel: #15 0xffffffff80888245 at acpi_pci_attach+0x15 Jun 12 08:39:58 alamar kernel: #16 0xffffffff806043fa at device_attach+0x3ca Jun 12 08:39:58 alamar kernel: #17 0xffffffff8060562a at bus_generic_attach+0x4a

It goes on for pages, at boot, then ends with:

Jun 12 08:39:58 alamar kernel: re0: Ethernet address: d4:3d:7e:b4:74:15

bye & Thanks

[kostikbel]

On 6/9/21 2:05 PM, Konstantin Belousov wrote: Load the module after the machine in multiuser.
Curiously, if I boot single user, then load this module, it does not panic. In multi user mode, it does, but still saves no core (and this machine normally does).
If screen shots are readable, try posting that.
Here it is. HTH.

Where are the screenshots?

Also, I found this in the logs:
a> lamar kernel: bus_dma_tag_create with the following non-sleepable locks held: Jun 12 08:39:58 alamar kernel: exclusive sleep mutex re0 (network driver) r = 0 (0xfffffe0004f29150) locked @ if_re.c:5118

This might be related, but probably not. I will fix these warnings meantime.

[netfenceventu]

Sorry, I had sent it via mail as an attachment, but it must have been stripped.
I'm uploading it now.

[kostikbel]

I pushed the supposed fix as 53bb979

[netfenceventu]

On 6/13/21 4:48 PM, Konstantin Belousov wrote: I pushed the supposed fix as 53bb979 <53bb979>

Thanks. So far, so good: now I can boot with a debug-enabled kernel.

[nbakker]

This has now been committed to the ports tree: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=256757

Thanks again!