Re: MSI messages in 2.6.19-rc2

[bibo,mao]

msi_init function is called by multithreads at the same time,
there should be mutex lock against status variable access, or
adding this function with __initcall prefix?

thanks
bibo,mao

Stephen Hemminger wrote:
> Got this on boot up, doesn't seem to be reliably reproducible.
> System ran okay. My guess is it is a mulithread probe issue
>
> [    2.773198] assign_interrupt_mode Found MSI capability
> [    2.773227] assign_interrupt_mode Found MSI capability
> [    2.773291] kmem_cache_create: duplicate cache msi_cache
> [    2.773408] assign_interrupt_mode Found MSI capability
> [    2.773485]  [dump_trace+100/461] dump_trace+0x64/0x1cd
> [    2.773573] assign_interrupt_mode Found MSI capability
> [    2.773644]  [show_trace_log_lvl+18/37] <7>PCI: Setting latency timer of device 0000:00:1c.3 to 64
> [    2.773707] assign_interrupt_mode Found MSI capability
> [    2.774538] show_trace_log_lvl+0x12/0x25
> [    2.774606]  [show_trace+13/16] show_trace+0xd/0x10
> [    2.774698]  [dump_stack+25/27] dump_stack+0x19/0x1b
> [    2.774794]  [kmem_cache_create+1065/1116] kmem_cache_create+0x429/0x45c
> [    2.775078]  [msi_init+90/154] msi_init+0x5a/0x9a
> [    2.775554]  [pci_enable_msi+35/546] pci_enable_msi+0x23/0x222
> [    2.776041]  [pcie_port_device_register+522/847] pcie_port_device_register+0x20a/0x34f
> [    2.776519]  [pcie_portdrv_probe+87/128] pcie_portdrv_probe+0x57/0x80
> [    2.777001]  [pci_device_probe+57/91] pci_device_probe+0x39/0x5b
> [    2.777467]  [really_probe+60/181] really_probe+0x3c/0xb5
> [    2.778074]  [kthread+197/243] kthread+0xc5/0xf3
> [    2.778266]  [kernel_thread_helper+7/16] kernel_thread_helper+0x7/0x10
> [    2.778360] DWARF2 unwinder stuck at kernel_thread_helper+0x7/0x10
> [    2.778403]
> [    2.778441] Leftover inexact backtrace:
> [    2.778444]
> [    2.778517]  [show_trace_log_lvl+18/37] show_trace_log_lvl+0x12/0x25
> [    2.778592]  [show_trace+13/16] show_trace+0xd/0x10
> [    2.778665]  [dump_stack+25/27] dump_stack+0x19/0x1b
> [    2.778738]  [kmem_cache_create+1065/1116] kmem_cache_create+0x429/0x45c
> [    2.778817]  [msi_init+90/154] msi_init+0x5a/0x9a
> [    2.778896]  [pci_enable_msi+35/546] pci_enable_msi+0x23/0x222
> [    2.778970]  [pcie_port_device_register+522/847] pcie_port_device_register+0x20a/0x34f
> [    2.779045]  [pcie_portdrv_probe+87/128] pcie_portdrv_probe+0x57/0x80
> [    2.779120]  [pci_device_probe+57/91] pci_device_probe+0x39/0x5b
> [    2.779193]  [really_probe+60/181] really_probe+0x3c/0xb5
> [    2.779267]  [kthread+197/243] kthread+0xc5/0xf3
> [    2.779341]  [kernel_thread_helper+7/16] kernel_thread_helper+0x7/0x10
> [    2.779414]  =======================
> [    2.779455] PCI: MSI cache init failed
> [    2.802070] Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing enabled
> [    2.802247] serial8250: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/