user/sven/linux.git/drivers/vme, branch v4.11.3

VME: restore bus_remove function causing incomplete module unload

2017-01-25T10:41:44Z

Commit 050c3d52cc7810d9d17b8cd231708609af6876ae ("vme: make core vme support explicitly non-modular") dropped the remove function because it appeared as if it was for removal of the bus, which is not supported. However, vme_bus_remove() is called when a VME device is removed from the bus and not when the bus is removed; as it calls the VME device driver's cleanup function. Without this function, the remove() in the VME device driver is never called and VME device drivers cannot be reloaded again. Here we restore the remove function that was deleted in that commit, and the reference to the function in the bus structure. Fixes: 050c3d52cc78 ("vme: make core vme support explicitly non-modular") Cc: Manohar Vanga Acked-by: Martyn Welch Cc: devel@driverdev.osuosl.org Signed-off-by: Stefano Babic Signed-off-by: Paul Gortmaker Cc: stable # 4.9 Signed-off-by: Greg Kroah-Hartman

vme: Fix wrong pointer utilization in ca91cx42_slave_get

2017-01-11T09:42:16Z

In ca91cx42_slave_get function, the value pointed by vme_base pointer is set through: *vme_base = ioread32(bridge->base + CA91CX42_VSI_BS[i]); So it must be dereferenced to be used in calculation of pci_base: *pci_base = (dma_addr_t)*vme_base + pci_offset; This bug was caught thanks to the following gcc warning: drivers/vme/bridges/vme_ca91cx42.c: In function ‘ca91cx42_slave_get’: drivers/vme/bridges/vme_ca91cx42.c:467:14: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast] *pci_base = (dma_addr_t)vme_base + pci_offset; Signed-off-by: Augusto Mecking Caringi Acked-By: Martyn Welch Cc: stable Signed-off-by: Greg Kroah-Hartman

vme: vme_get_size potentially returning incorrect value on failure

2016-10-28T12:25:18Z

The function vme_get_size returns the size of the window to the caller, however it doesn't check the return value of the call to vme_master_get. Return 0 on failure rather than anything else. Suggested-by: Dan Carpenter Signed-off-by: Martyn Welch Signed-off-by: Greg Kroah-Hartman

vme: fake: remove unexpected unlock in fake_master_set()

2016-09-27T10:43:35Z

image->lock is unlocked in some error handling path without take the lock, so remove those unexpected unlock. Fixes: 658bcdae9c67 ("vme: Adding Fake VME driver") Signed-off-by: Wei Yongjun Signed-off-by: Greg Kroah-Hartman

vme: fake: mark symbols static where possible

2016-09-27T10:43:34Z

We get 4 warnings when building kernel with W=1: drivers/vme/bridges/vme_fake.c:384:6: warning: no previous prototype for 'fake_lm_check' [-Wmissing-prototypes] drivers/vme/bridges/vme_fake.c:619:6: warning: no previous prototype for 'fake_vmewrite8' [-Wmissing-prototypes] drivers/vme/bridges/vme_fake.c:649:6: warning: no previous prototype for 'fake_vmewrite16' [-Wmissing-prototypes] drivers/vme/bridges/vme_fake.c:679:6: warning: no previous prototype for 'fake_vmewrite32' [-Wmissing-prototypes] In fact, these functions are only used in the file in which they are declared and don't need a declaration, but can be made static. so this patch marks these functions with 'static'. Signed-off-by: Baoyou Xie Acked-by: Martyn Welch Signed-off-by: Greg Kroah-Hartman

vme: fake: fix build for 64-bit dma_addr_t

2016-09-13T15:30:17Z

casting between dma_addr_t and a pointer is generally tricky, as they might not be the same size and almost never point into the same address space. With 32-bit ARM systems and LPAE, we get this warning for the vme_fake driver that stores a pointer in a dma_addr_t variable: drivers/vme/bridges/vme_fake.c: In function 'fake_slave_set': drivers/vme/bridges/vme_fake.c:204:29: error: assignment makes pointer from integer without a cast [-Werror=int-conversion] To make this clearer while fixing the warning, I'm adding a set of helper functions for the type conversion. Signed-off-by: Arnd Bergmann Acked-by: Martyn Welch Signed-off-by: Greg Kroah-Hartman

vme: Adding Fake VME driver

2016-08-31T11:21:12Z

This patch introduces a fake VME bridge driver. This driver currently emulates a subset of the VME bridge functionality. This allows some VME subsystem development and even some VME device driver development to be carried out in the absence of a proper VME bus. Signed-off-by: Martyn Welch Signed-off-by: Greg Kroah-Hartman

vme: make core vme support explicitly non-modular

2016-08-31T11:20:16Z

The Kconfig currently controlling compilation of this code is: drivers/vme/Kconfig:menuconfig VME_BUS drivers/vme/Kconfig: bool "VME bridge support" ...meaning that it currently is not being built as a module by anyone. Lets remove the modular code that is essentially orphaned, so that when reading the driver there is no doubt it is builtin-only. We replace module.h and moduleparam.h (unused) with init.h and also export.h ; the latter since this file does export some syms. Since this is a struct bus_type and not a platform_driver, we don't have any ".suppress_bind_attrs" to be concerned about when we drop the ".remove" code from this file. Since module_init was not in use by this code, the init ordering remains unchanged with this commit. Cc: Manohar Vanga Signed-off-by: Paul Gortmaker Acked-by: Martyn Welch Signed-off-by: Greg Kroah-Hartman

vme: Fix module autoload

2016-08-31T11:20:16Z

These drivers have a PCI device ID table but the PCI module alias information is not created so module autoloading won't work. Signed-off-by: Alessio Igor Bogani Acked-by: Martyn Welch Signed-off-by: Greg Kroah-Hartman

vme: change LM callback argument to void pointer

2016-08-31T11:20:15Z

Make the location monitor callback function prototype more useful by changing the argument from an integer to a void pointer. All VME bridge drivers were simply passing the location monitor index (e.g. 0-3) as the argument to these callbacks. It is much more useful to pass back a pointer to data that the callback-registering driver cares about. There appear to be no in-kernel callers of vme_lm_attach (or vme_lme_request for that matter), so this change only affects the VME subsystem and bridge drivers. This has been tested with Tsi148 hardware, but the CA91Cx42 changes have only been compiled. Signed-off-by: Aaron Sierra Acked-by: Martyn Welch Signed-off-by: Greg Kroah-Hartman