February 2014 - kernelnewbies@kernelnewbies.org

16-Byte aligned memory allocated in user space and then
by m silverstri Feb. 5, 2014

Feb. 5, 2014

Hi, If I allocated memory in user space and make it 16 byte aligned memory, and then pass it to kernel and setup DMA for my kernel driver, will my kernel driver still see the memory in 16-Byte aligned memory? The hard that my kernel driver takes to need 16-Byte aligned memory. Thank you.

2 1

How to map more than 1 page of buffer to kernel address?
by m silverstri Feb. 5, 2014

Feb. 5, 2014

Hi, I need to map an buffer object to kernel address. What should I buffer is more than 1 page and I want to map the whole buffer to the kernel address? https://www.kernel.org/doc/htmldocs/device-drivers/API-dma-buf-kmap.html dma_buf_kmap -- Map a page of the buffer object into kernel address space. The same restrictions as for kmap and friends apply. Thanks.

2 1

Where to find USB VID and PID?
by Eric Fowler Feb. 4, 2014

Feb. 4, 2014

I have an Emcraft SmartFusion2 board with a USB interface, running ucLinux. When I plug the board to a Windows machine, said Win machine sees an 'unknown device', which usually (always) means VID_0000 and PID_0000. This is separate and independent of the USB and JTAG interfaces that are used for programming, network access, and debugging, which are all working as expected. How do I: - Find the device on the linux side that maps to that endpoint, if it exists (I have already found the device object for the root hub) - Interrogate for the VID & PID that a given USB device presents to the connecting host - Manipulate the VID & PID on a USB device on ucLinux -- cc:NSA

2 1

GSoc
by David kiarie Feb. 4, 2014

Feb. 4, 2014

Hi everybody. I am College student and I happen to be following a couple of open-source mailing lists.I have noted that threads related to the GSoc projects have come up again. Now could anyone have any ideas on some nice GSoc Projects for Linux?

2 2

Question about using spinlock to synchronize between kernel driver and an interrupt handler
by m silverstri Feb. 4, 2014

Feb. 4, 2014

Hi, I read this article http://www.linuxjournal.com/article/5833 to learn about spinlock. I try this to use it in my kernel driver. Here is what my driver code needs to do: In f1(), it will get the spin lock, and caller can call f2() will wait for the lock since the spin lock is not being unlock. The spin lock will be unlock in my interrupt handler (triggered by the HW). void f1() { spin_lock(&mylock); // write hardware REG_ADDR += FLAG_A; } void f2() { spin_lock(&mylock); //... } The hardware will send the application an interrupt and my interrupt handler will call spin_unlock(&mylock); My question is if I call f1() f2() // i want this to block until the interrupt return saying setting REG_ADDR is done. when I run this, I get an exception in kernel saying a deadlock " INFO: possible recursive locking detected" How can I re-write my code so that kernel does not think I have a deadlock? I want my driver code to wait until HW sends me an interrupt saying setting REG_ADDR is done. Thank you.

5 15

Decoding a Linux kernel oops panic due to DMAR error
by Ahmed A Feb. 3, 2014

Feb. 3, 2014

Hello, I have a server with onboard Intel 10G ports (82599). When I load the kernel module driver for these ports, everything is fine, I can see the newly created ethX devices using "ip addr show". However, after I assign an IP address, and right after I issue command to bring up the port, I get a kernel panic related to DMAR (DMA remapping) in the VFIO (Virtual Function IO) module. I am not even sure why I am getting this panic since this Intel kernel module does not use VFIO. I know why the panic is happening, NULL being sent as a parameter to function vfio_group_get(), in which it is being de-referenced. I know NULL is passed, because register RDI, which is used to pass the first argument to a function, contains 0. Linux kernel 3.6.11 Following is the stack trace of panic: # [11036.855410] BUG: unable to handle kernel [11036.887249] ixgbe 0000:84:00.0: eth6: detected SFP+: 3 NULL pointer dereference at (null) [11037.010224] IP: [<ffffffffa006615a>] vfio_group_get+0x9/0x27 [vfio] [11037.085047] PGD 1fd6b5b067 PUD 20404b1067 PMD 0 [11037.140181] Oops: 0000 [#1] SMP [11037.178676] Modules linked in: ixgbe(O) nfsv3 autofs4 nfsd nfs_acl nfs lockd sunrpc vfio_pci vfio_iommu_type1 vfio i2c_mux i2c_smbus i2c_dev container ide_pci_generic ide_core uhci_hcd isci ata_generic [11037.393137] CPU 0 [11037.414974] Pid: 14045, comm: kworker/0:0 Tainted: G O 3.6.11 [11037.539628] RIP: 0010:[<ffffffffa006615a>] [<ffffffffa006615a>] vfio_group_get+0x9/0x27 [vfio] [11037.643521] RSP: 0018:ffff881f52453d00 EFLAGS: 00010282 [11037.706886] RAX: ffff881fd6740680 RBX: 0000000000000000 RCX: ffff88204157ec00 [11037.792053] RDX: 0000000000000084 RSI: 0000000001f5327a RDI: 0000000000000000 [11037.877221] RBP: ffff881f52453d10 R08: ffff881f5327abe0 R09: 0000000000000000 [11037.962394] R10: 0000000000000000 R11: 0000000000000000 R12: ffff88204157f800 [11038.024995] ixgbe 0000:84:00.0: eth6: NIC Link is Up 10 Gbps, Flow Control: RX/TX [11038.025144] IPv6: ADDRCONF(NETDEV_CHANGE): eth6: link becomes ready [11038.211671] R13: 0000000000000084 R14: 0000000000000000 R15: 0000000000000000 [11038.296842] FS: 0000000000000000(0000) GS:ffff88204f000000(0000) knlGS:0000000000000000 [11038.393430] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [11038.461988] CR2: 0000000000000000 CR3: 0000001fd686d000 CR4: 00000000001407f0 [11038.547156] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [11038.632326] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 [11038.717496] Process kworker/0:0 (pid: 14045, threadinfo ffff881f52452000, task ffff882034d61950) [11038.822392] Stack: [11038.846298] 0000000000000084 ffff881fd6740680 ffff881f52453d30 ffffffffa006618a [11038.934688] 0000000001f5327a ffff882035e23e00 ffff881f52453d50 ffffffffa0066442 [11039.023078] ffff881f52453d70 ffff881fd6740680 ffff881f52453d70 ffffffffa0072072 [11039.111465] Call Trace: [11039.140571] [<ffffffffa006618a>] vfio_device_get+0x12/0x30 [vfio] [11039.214324] [<ffffffffa0066442>] vfio_device_get_from_dev+0x19/0x1f [vfio] [11039.297425] [<ffffffffa0072072>] vfio_pci_dmar_error_handler+0x13/0x4a [vfio_pci] [11039.387796] [<ffffffff81420cc6>] dmar_fault_do_one+0xd4/0xf1 [11039.456366] [<ffffffff8104175d>] process_one_work+0x1c2/0x311 [11039.525968] [<ffffffff81041568>] ? manage_workers+0x23a/0x24c [11039.595566] [<ffffffff81420bf2>] ? dmar_get_fault_reason+0x52/0x52 [11039.670354] [<ffffffff81041b42>] worker_thread+0x26c/0x34a [11039.736840] [<ffffffff810418d6>] ? process_scheduled_works+0x2a/0x2a [11039.813710] [<ffffffff8104583a>] kthread+0x86/0x8e [11039.871891] [<ffffffff81604bf4>] kernel_thread_helper+0x4/0x10 [11039.942524] [<ffffffff810457b4>] ? kthread_freezable_should_stop+0x4d/0x4d [11040.025618] [<ffffffff81604bf0>] ? gs_change+0xb/0xb [11040.085865] Code: 48 8b 00 48 8b 40 20 48 85 c0 74 0c 55 48 8b 7f 40 48 89 e5 ff d0 eb 08 48 c7 c0 ea ff ff ff c3 5d c3 55 48 89 e5 53 48 89 fb 52 <8b> 07 85 c0 75 11 be 2a 00 00 00 48 c7 c7 38 76 06 a0 e8 32 84 [11040.312869] RIP [<ffffffffa006615a>] vfio_group_get+0x9/0x27 [vfio] [11040.388722] RSP <ffff881f52453d00> [11040.430282] CR2: 0000000000000000 - Can someone please help me understand the damr/vfio related function calls in the back trace, and why they are getting invoked? I know what causes DMAR error, but not sure how this could be happening, since none of the devices is managed by VFIO. - Looking at the source code, it seems dmar_fault_do_one() is called from interrupt handler dmar_fault(). I am just curious, why dmar_fault() is not part of the stack trace? - What is the significance of the "?" in front of some of the functions in the backtrace (e.g. dmar_get_fault_reason()). Thank you, Ahmed.

2 1

Issue Float point in ARM WARNING: "__aeabi_d2iz"
by naveen yadav Feb. 3, 2014

Feb. 3, 2014

Dear All, We are using ARM SOC cortex A15. we are running 3.10.x kernel. I am using gcc 4.6 I need floating point in kernel module. So when I build I got warning WARNING: "__aeabi_d2iz" [/home/test/tt/hello.ko] undefined! WARNING: "__aeabi_dmul" [/home/test/tt/hello.ko] undefined! WARNING: "__aeabi_dsub" [/home/test/tt/hello.ko] undefined! WARNING: "__aeabi_ddiv" [/home/test/tt/hello.ko] undefined! WARNING: "__aeabi_f2d" [/home/test/tt/hello.ko] undefined! static int __init hello_init(void) { float a=3.14; float b=3.1222; b=7.2/a ; printk(KERN_INFO "Hello world! %f\n",a/b); return 0; // Non-zero return means that the module couldn't be loaded. } static void __exit hello_cleanup(void) { printk(KERN_INFO "Cleaning up module.\n"); } module_init(hello_init); module_exit(hello_cleanup); So I change in my Linux kernel make file. arch/arm/Makefile I remove -msoft-float . -KBUILD_CFLAGS +=$(CFLAGS_ABI) $(CFLAGS_ISA) $(arch-y) $(tune-y) $(call cc-option,-mshort-load-bytes,$(call cc-option,-malignment-traps,)) -msoft-float -Uarm -KBUILD_AFLAGS +=$(CFLAGS_ABI) $(AFLAGS_ISA) $(arch-y) $(tune-y) -include asm/unified.h -msoft-float +KBUILD_CFLAGS +=$(CFLAGS_ABI) $(CFLAGS_ISA) $(arch-y) $(tune-y) $(call cc-option,-mshort-load-bytes,$(call cc-option,-malignment-traps,)) -Uarm +KBUILD_AFLAGS +=$(CFLAGS_ABI) $(AFLAGS_ISA) $(arch-y) $(tune-y) -include asm/unified.h I run my module and it work well. So my question is . 1. Is it correct way ? 2. will it effect my performance in normal scenario ? 3. how can I check more stability? Regards.

2 1

Loading / Unloading kernel module from another module
by Chetan Nanda Feb. 3, 2014

Feb. 3, 2014

Hi, Is there a way to load and unload a module from another module? I know 'request_module' can be used to load but what is the API for unloading a module. Thanks for your help. Thanks, Chetan Nanda

3 2