kernel.git/drivers/uio/uio.c, branch linux-4.5.y

uio: fix false positive __might_sleep warning splat

2015-10-04T11:32:25Z

Andy has reported a __might_sleep warning [ 5174.883617] WARNING: CPU: 0 PID: 1532 at /home/agrover/git/kernel/kernel/sched/core.c:7389 __might_sleep+0x7d/0x90() [ 5174.884407] do not call blocking ops when !TASK_RUNNING; state=1 set at [] uio_read+0x91/0x170 [uio] [ 5174.885198] Modules linked in: tcm_loop target_core_user uio target_core_pscsi target_core_file target_core_iblock iscsi_target_mod target_core_mod uinput fuse nfsv3 nfs_acl nfs lockd grace fscache sunrpc microcode i2c_piix4 virtio_balloon i2c_core xfs libcrc32c crc32c_intel virtio_net virtio_blk [ 5174.887351] CPU: 0 PID: 1532 Comm: tcmu-runner Not tainted 4.2.0-rc7+ [ 5174.887853] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.8.1-20150318_183358- 04/01/2014 [ 5174.888633] ffffffff81a3b870 ffff880045393ca8 ffffffff817afaae 0000000000000000 [ 5174.889224] ffff880045393cf8 ffff880045393ce8 ffffffff8109a846 ffff880045393cd8 [ 5174.889793] ffffffffa02a7150 00000000000002dc 0000000000000000 ffff880045008000 [ 5174.890375] Call Trace: [ 5174.890562] [] dump_stack+0x4c/0x65 [ 5174.890938] [] warn_slowpath_common+0x86/0xc0 [ 5174.891388] [] warn_slowpath_fmt+0x46/0x50 [ 5174.891808] [] ? uio_read+0x91/0x170 [uio] [ 5174.892237] [] ? uio_read+0x91/0x170 [uio] [ 5174.892653] [] __might_sleep+0x7d/0x90 [ 5174.893055] [] __might_fault+0x43/0xa0 [ 5174.893448] [] ? schedule+0x3e/0x90 [ 5174.893820] [] uio_read+0x132/0x170 [uio] [ 5174.894240] [] ? wake_up_q+0x70/0x70 [ 5174.894620] [] __vfs_read+0x28/0xe0 [ 5174.894993] [] ? security_file_permission+0xa3/0xc0 [ 5174.895541] [] ? rw_verify_area+0x4f/0xf0 [ 5174.896006] [] vfs_read+0x8a/0x140 [ 5174.896391] [] ? __schedule+0x425/0xcc0 [ 5174.896788] [] SyS_read+0x49/0xb0 The warning is a false positive because uio_read doesn't depent on TASK_INTERRUPTIBLE after copy_to_user so it is safe to silence the warning by an explicit setting the state to TASK_RUNNING in the path which might call into TASK_RUNNING. Reported-by: Andy Grover Signed-off-by: Michal Hocko Signed-off-by: Greg Kroah-Hartman

uio: Destroy uio_idr on module exit

2015-08-05T20:25:59Z

Destroy uio_idr on module exit, reclaiming the allocated memory. This was detected by the following semantic patch (written by Luis Rodriguez ) @ defines_module_init @ declarer name module_init, module_exit; declarer name DEFINE_IDR; identifier init; @@ module_init(init); @ defines_module_exit @ identifier exit; @@ module_exit(exit); @ declares_idr depends on defines_module_init && defines_module_exit @ identifier idr; @@ DEFINE_IDR(idr); @ on_exit_calls_destroy depends on declares_idr && defines_module_exit @ identifier declares_idr.idr, defines_module_exit.exit; @@ exit(void) { ... idr_destroy(&idr); ... } @ missing_module_idr_destroy depends on declares_idr && defines_module_exit && !on_exit_calls_destroy @ identifier declares_idr.idr, defines_module_exit.exit; @@ exit(void) { ... +idr_destroy(&idr); } Signed-off-by: Johannes Thumshirn Signed-off-by: Greg Kroah-Hartman

uio: don't free irq that was not requested

2015-05-24T18:55:15Z

UIO base driver should only free_irq that it has requested. UIO supports drivers without interrupts (irq == 0) or custom handlers. This fixes warnings like: WARNING: CPU: 1 PID: 5478 at kernel/irq/manage.c:1244 __free_irq+0xa9/0x1e0() Trying to free already-free IRQ 0 Signed-off-by: Stephen Hemminger Signed-off-by: Greg Kroah-Hartman

uio: Request/free irq separate from dev lifecycle

2015-03-20T12:18:16Z

Separate irq request/free from the device lifecycle. After device unregister the parent module can call pci_disable_msi. >From the PCI MSI how to: "Before calling this function, a device driver must always call free_irq() on any interrupt for which it previously called request_irq(). Failure to do so results in a BUG_ON(), leaving the device with MSI enabled and thus leaking its vector." So we need to separately free the irq at unregister to allow the device to be kept around in the case of it still having open FDs. Signed-off-by: Brian Russell Signed-off-by: Greg Kroah-Hartman

uio: support memory sizes larger than 32 bits

2014-11-07T18:09:07Z

This is a completion to 27a90700a4275c5178b883b65927affdafa5185c The size field is also increased to allow values larger than 32 bits on platforms that have more than 32 bit physical addresses. Signed-off-by: Cristian Stoica Signed-off-by: Greg Kroah-Hartman

uio: Export definition of struct uio_device

2014-10-03T04:35:54Z

In order to prevent a O(n) search of the filesystem to link up its uio node with its target configuration, TCMU needs to know the minor number that UIO assigned. Expose the definition of this struct so TCMU can access this field. Signed-off-by: Andy Grover Signed-off-by: Nicholas Bellinger

Revert "uio: fix vma io range check in mmap"

2014-06-17T23:07:08Z

This reverts commit ddb09754e6c7239e302c7b675df9bbd415f8de5d. Linus objected to this originally, I can see why it might be needed, but given that no one spoke up defending this patch, I'm going to revert it. If you have hardware that requires this change, please speak up in the future and defend the patch. Reported-by: Linus Torvalds Cc: Bin Wang Cc: Nobuhiro Iwamatsu Cc: Norbert Ciosek Cc: stable Signed-off-by: Greg Kroah-Hartman

uio: fix vma io range check in mmap

2014-05-27T21:11:06Z

the vma range size is always page size aligned in mmap, while the real io space range may not be page aligned, thus leading to range check failure in the uio_mmap_physical(). for example, in a case of io range size "mem->size == 1KB", and we have (vma->vm_end - vma->vm_start) == 4KB, due to "len" is aligned to page size in do_mmap_pgoff(). now fix this issue by align mem->size to page size in the check. Signed-off-by: Bin Wang Signed-off-by: Nobuhiro Iwamatsu Cc: stable Signed-off-by: Greg Kroah-Hartman

uio: fix devm_request_irq usage

2013-12-20T16:49:17Z

Commit e6789cd3dfb553077606ccafeb05e0043f072481 (uio: Simplify uio error path by using devres functions) converted uio to use devm_request_irq(). This introduced a change in behaviour since the IRQ is associated with the parent device instead of the created UIO device. The IRQ will remain active after uio_unregister_device() is called, and some drivers will crash because of this. The patch fixes this. Signed-off-by: Aaro Koskinen Cc: stable # 3.13 Signed-off-by: Greg Kroah-Hartman

uio: we cannot mmap unaligned page contents

2013-12-02T19:50:37Z

In commit 7314e613d5ff ("Fix a few incorrectly checked [io_]remap_pfn_range() calls") the uio driver started more properly checking the passed-in user mapping arguments against the size of the actual uio driver data. That in turn exposed that some driver authors apparently didn't realize that mmap can only work on a page granularity, and had tried to use it with smaller mappings, with the new size check catching that out. So since it's not just the user mmap() arguments that can be confused, make the uio mmap code also verify that the uio driver has the memory allocated at page boundaries in order for mmap to work. If the device memory isn't properly aligned, we return [ENODEV] The fildes argument refers to a file whose type is not supported by mmap(). as per the open group documentation on mmap. Reported-by: Holger Brunck Acked-by: Greg KH Signed-off-by: Linus Torvalds