I have the following BUG-warnings in dmesg. It seems to concern the fglrx driver from the ATI video drivers package ati-drivers-8.501. There are no noticable degradations in video performance and/or stability, but on the other hand I'm not exactly sure where I should look, since I don't quite understand what these BUG-warnings are about. fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel. Symbol init_mm is marked as UNUSED, however this module is using it. This symbol will go away in the future. Please evalute if this is the right api to use, and if it really is, submit a report the linux kernel mailinglist together with submitting your code for inclusion. [fglrx] vendor: 1002 device: 5b60 count: 1 [fglrx] Maximum main memory to use for locked dma buffers: 3268 MBytes. [fglrx] PAT is enabled successfully! [fglrx] module loaded - fglrx 8.50.3 [Jun 2 2008] with 1 minors ACPI: PCI Interrupt 0000:00:1f.3[B] -> GSI 17 (level, low) -> IRQ 17 input: PC Speaker as /devices/platform/pcspkr/input/input4 ACPI: PCI Interrupt 0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16 BUG: using smp_processor_id() in preemptible [00000000] code: X/5219 caller is __ke_spin_lock+0x22/0x2e [fglrx] Pid: 5219, comm: X Tainted: P 2.6.25-gentoo-r7 #3 [<c028cff1>] debug_smp_processor_id+0x91/0xa4 [<f8b20b54>] __ke_spin_lock+0x22/0x2e [fglrx] [<f8b20b6a>] KAS_GetExecutionLevel+0xa/0x9d [fglrx] [<f8b22e76>] KAS_InterlockedList_InsertAtTail+0x27/0x76 [fglrx] [<f8b38b88>] MCIL_ZeroMemory+0x58/0x60 [fglrx] [<f8b37781>] MCIL_InterlockedList+0xf1/0x1a0 [fglrx] [<c014ccfa>] get_page_from_freelist+0x24f/0x394 [<f8bdde68>] RingBuffer_Initialize+0x148/0x160 [fglrx] [<f8b37805>] MCIL_InterlockedList+0x175/0x1a0 [fglrx] [<f8b32422>] PcsCommand+0x262/0x600 [fglrx] [<f8b322ea>] PcsCommand+0x12a/0x600 [fglrx] [<f8be5ed8>] DispatchRing_Initialize+0xf8/0x1a0 [fglrx] [<f8b3677f>] MCIL_GetRegistrykey+0x1cf/0x2c0 [fglrx] [<f8be6c22>] IRQMGR_ZeroMemory+0x32/0x40 [fglrx] [<f8bea1d0>] R1XX_InitializeFlags+0xb0/0x1d0 [fglrx] [<f8bde397>] InitializeTimerInfo+0xb7/0x100 [fglrx] [<f8be883c>] InitializeInterrupts+0x3c/0x60 [fglrx] [<f8bdc786>] IRQMGR_InitializeIRQMGR+0xc6/0xd0 [fglrx] [<f8b3a444>] irqmgr_wrap_initialize+0x84/0x100 [fglrx] [<f8b3a0d1>] firegl_install_irq+0x181/0x2c0 [fglrx] [<f8b30046>] firegl_get_lockid+0x56/0x80 [fglrx] [<f8b39f50>] firegl_install_irq+0x0/0x2c0 [fglrx] [<f8b2d37e>] firegl_ioctl+0x19e/0x1030 [fglrx] [<f8b22a4b>] ip_firegl_ioctl+0xc/0x10 [fglrx] [<c016e3ff>] vfs_ioctl+0x47/0x5d [<c016e64d>] do_vfs_ioctl+0x238/0x24b [<c016e68c>] sys_ioctl+0x2c/0x44 [<c0103b62>] sysenter_past_esp+0x5f/0x85 ======================= (... a couple more like the above, see dmesg.log ...) [fglrx] Reserved FB block: Shared offset:0, size:40000 [fglrx] Reserved FB block: Unshared offset:3fbf000, size:40000 [fglrx] Reserved FB block: Unshared offset:3fff000, size:1000 I'm not sure if Gentoo Bugzilla is the right place for this report. If not, I'd appreciate directions.
Created attachment 163203 [details] dmesg log
I saw the line: Symbol init_mm is marked as UNUSED, however this module is using it. in my dmesg too. The ATI fglrx driver is in use. I made it go away by changing the kernel configuration using make menuconfig to unselect the Direct Rendering Manager. Device Drivers ---> Graphics support ---> < > Direct Rendering Manager (XFree86 4.1.0 and higher DRI support) ---> This may have disabled something important, but it works much better with my ASUS EAH3450 ATI based PCIe graphics card.
Actually the message is still there. This does not fix it. (In reply to comment #2) > I saw the line: > > Symbol init_mm is marked as UNUSED, however this module is using it. > > in my dmesg too. The ATI fglrx driver is in use. >
I am seeing this too, on gentoo-sources-2.6.26-r9. Same stack as is shown. The dump is repeated ten times after ACPI reports that it is loading and hooking interrupts. fglrx continues to load, and direct rendering it reported by glxinfo. I don't see any other evidence of a problem.
The dump is being produced in debug_smp_processor_id. It is not a panic, but a few printk's and a call to dump_stack. The problem is ati's fglrx is calling smp_processor_id in a kernel premptible context, which is bad. ATI must fix this.
If anyone would like to turn off the noise this makes, enable CONFIG_DEBUG_PREEMPT in Kernel Hacking. This doesn't solve the problem, it just silences the alarm.
(In reply to comment #6) > If anyone would like to turn off the noise this makes, enable Sorry, that should have been "disable".
Same here. Linux Kernel v2.6.27-gentoo-r7 x11-drivers/ati-drivers-8.542 BUG: using smp_processor_id() in preemptible [00000000] code: X/4842 caller is __ke_spin_lock+0x27/0x3b [fglrx] Pid: 4842, comm: X Tainted: P 2.6.27-gentoo-r7-phenom #5 Call Trace: [<ffffffff803b76b7>] debug_smp_processor_id+0xc7/0xdc [<ffffffffa00010bd>] __ke_spin_lock+0x27/0x3b [fglrx] [<ffffffffa00010dc>] KAS_GetExecutionLevel+0xb/0xaa [fglrx] [<ffffffffa00024c1>] KAS_SlabCache_AllocEntry+0x50/0x13f [fglrx] [<ffffffff80220f70>] print_trace_address+0x11/0x20 [<ffffffffa00180a5>] MCIL_LookasideList+0x165/0x2d0 [fglrx] [<ffffffffa00bd525>] IRQMGR_LookasideList_AllocateEntry+0x35/0x40 [fglrx] [<ffffffff80221f0d>] dump_stack+0x69/0x6f [<ffffffff803b13bd>] __ratelimit+0xb5/0xc0 [<ffffffffa00b970c>] AllocateCallbackEntry+0x8c/0xb0 [fglrx] [<ffffffffa00b3fb0>] RegisterIRQClient+0x90/0x1a0 [fglrx] [<ffffffffa001979d>] MCIL_CopyMemory+0x5d/0x70 [fglrx] [<ffffffffa00b0df8>] IRQMGR_Access+0xb8/0x140 [fglrx] [<ffffffffa002eee0>] fireglAsyncioUnregisterIntMsgHandlers+0x3a0/0x440 [fglrx] [<ffffffffa002ecba>] fireglAsyncioUnregisterIntMsgHandlers+0x17a/0x440 [fglrx] [<ffffffffa002eee0>] fireglAsyncioUnregisterIntMsgHandlers+0x3a0/0x440 [fglrx] [<ffffffffa006306b>] asyncIONotifyMsg+0x30b/0x3a0 [fglrx] [<ffffffffa0062f64>] asyncIONotifyMsg+0x204/0x3a0 [fglrx] [<ffffffffa002e3df>] firegl_asyncio_write+0x18f/0x250 [fglrx] [<ffffffffa0003b4b>] ip_firegl_write+0x48/0x78 [fglrx] [<ffffffff8029e3c3>] vfs_write+0xad/0x136 [<ffffffff8029e84b>] sys_write+0x45/0x6e [<ffffffff8021fb4b>] system_call_fastpath+0x16/0x1b It began happening after changing from: Processor type and features ---> Preemption Model (Voluntary Kernel Preemption (Desktop)) General setup ---> [*] Group CPU scheduler to: Preemption Model (Preemptible Kernel (Low-Latency Desktop)) General setup ---> [ ] Group CPU scheduler
this ati-drivers version was removed from tree
