Commit Graph

20 Commits

Author SHA1 Message Date
Johannes Thumshirn
4d2ec85753 mcb: Acquire reference to carrier module in core
Acquire a reference to the carrier's kernel module in bus code, so
it can't be removed from the kernel while it still has a bus and thus
possibly devices attached to it.

Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Reported-by: Andreas Werner <andreas.werner@men.de>
Tested-by: Andreas Werner <andreas.werner@men.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-06-13 18:49:30 -07:00
Johannes Thumshirn
7bc364097a mcb: Acquire reference to device in probe
mcb_probe() does not aqcuire a reference to the probed device but drops one
when removing the device. As it is actually using the device, it should grab
a reference via get_device().

This could lead to a panic found with a rmmod/modprobe stress test

Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Reported-by: Andreas Werner <andreas.werner@men.de>
Tested-by: Andreas Werner <andreas.werner@men.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-06-13 18:49:30 -07:00
Andreas Werner
8cd1f9d01c mcb: Delete num_cells variable which is not required
The num_cells variable is only used in the dev_dbg print,
but we can directly use the ret variable which also includes the same
value.

Signed-off-by: Andreas Werner <andreas.werner@men.de>
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-05-03 15:52:28 -07:00
Andreas Werner
f75564d343 mcb: Fixed bar number assignment for the gdd
The bar number is found in reg2 within the gdd. Therefore
we need to change the assigment from reg1 to reg2 which
is the correct location.

Signed-off-by: Andreas Werner <andreas.werner@men.de>
Fixes: '3764e82e5' drivers: Introduce MEN Chameleon Bus
Cc: stable@vger.kernel.org # v3.15+
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-05-03 15:52:28 -07:00
Andreas Werner
d19366f1d6 mcb: Replace ioremap and request_region with the devm version
Replaced ioremap with devm_ioremap and request_mem_region with
devm_request_mem_region. This makes the code much more cleaner.

Signed-off-by: Andreas Werner <andreas.werner@men.de>
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-05-03 15:52:28 -07:00
Johannes Thumshirn
5d9e2ab9fe mcb: Implement bus->dev.release callback
The mcb_bus structure previously was released in mcb_release_bus. This lead to
the following warning on module unload:

------------[ cut here ]------------
WARNING: CPU: 1 PID: 2032 at drivers/base/core.c:251 device_release+0x73/0x90
Device 'mcb:0' does not have a release() function, it is broken and must be fixed.
Modules linked in: men_z135_uart mcb_pci(-) mcb
CPU: 1 PID: 2032 Comm: rmmod Not tainted 4.6.0-rc4+ #3
Hardware name: N/A N/A/COMe-mBTi10, BIOS MVV1R921 X64 10/14/2015
 00000286 00000286 c0117de4 c12d6f16 c0117e2c c18be0d3 c0117dfc c104f6e1
 000000fb f5ccbe08 f5ccbe00 f5c64600 c0117e18 c104f728 00000009 00000000
 c0117e10 c18db674 c0117e2c c0117e3c c13ce5c3 c18be0d3 000000fb c18db674
Call Trace:
 [<c12d6f16>] dump_stack+0x47/0x61
 [<c104f6e1>] __warn+0xc1/0xe0
 [<c104f728>] warn_slowpath_fmt+0x28/0x30
 [<c13ce5c3>] device_release+0x73/0x90
 [<c12d92e4>] kobject_release+0x34/0x80
 [<c12d929d>] ? kobject_del+0x2d/0x40
 [<c12d9205>] kobject_put+0x25/0x50
 [<c13ce77f>] put_device+0xf/0x20
 [<c13d114b>] klist_devices_put+0xb/0x10
 [<c1752673>] klist_next+0x73/0xf0
 [<c13d1140>] ? unbind_store+0x100/0x100
 [<f8a23370>] ? mcb_bus_add_devices+0x30/0x30 [mcb]
 [<c13d0a81>] bus_for_each_dev+0x51/0x80
 [<f8a23319>] mcb_release_bus+0x19/0x40 [mcb]
 [<f8a23370>] ? mcb_bus_add_devices+0x30/0x30 [mcb]
 [<f8a2b033>] mcb_pci_remove+0x13/0x20 [mcb_pci]
 [<c130d358>] pci_device_remove+0x28/0xb0
 [<c13d201b>] __device_release_driver+0x7b/0x110
 [<c13d2847>] driver_detach+0x87/0x90
 [<c13d1b9b>] bus_remove_driver+0x3b/0x80
 [<c13d2ed0>] driver_unregister+0x20/0x50
 [<c130be53>] pci_unregister_driver+0x13/0x60
 [<f8a2b1f4>] mcb_pci_driver_exit+0xd/0xf [mcb_pci]
 [<c10be588>] SyS_delete_module+0x138/0x200
 [<c1159208>] ? ____fput+0x8/0x10
 [<c1068054>] ? task_work_run+0x74/0x90
 [<c1001879>] do_fast_syscall_32+0x69/0x120
 [<c1757597>] sysenter_past_esp+0x40/0x6a
---[ end trace 1ed34c2aa3019875 ]---

Release a mcb_bus' memory on the device's release callback, to avoid above
warning.

Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Reported-by: Andreas Werner <andreas.werner@men.de>
Tested-by: Andreas Werner <andreas.werner@men.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-05-03 15:52:28 -07:00
Johannes Thumshirn
803f1ca60d mcb: export bus information via sysfs
Export information about the bus stored in the FPGA's header to userspace via
sysfs, instead of hiding it in pr_debug()s from everyone.

Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Andreas Werner <andreas.werner@men.de>
Tested-by: Andreas Werner <andreas.werner@men.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-05-03 15:07:35 -07:00
Johannes Thumshirn
18d2881980 mcb: Correctly initialize the bus's device
The mcb bus' device member wasn't correctly initialized and thus wasn't placed
correctly into the driver model.

Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Andreas Werner <andreas.werner@men.de>
Tested-by: Andreas Werner <andreas.werner@men.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2016-05-03 15:07:35 -07:00
Johannes Thumshirn
169883a63e mcb: Destroy IDA on module unload
Destroy mcb_ida on module_unload

Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-10-29 09:02:16 +09:00
Alexey Khoroshilov
bf25c19979 mcb: Do not return zero on error path in mcb_pci_probe()
There is an error path in mcb_pci_probe() where
it returns zero instead of error code.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-10-29 09:02:16 +09:00
Alexey Khoroshilov
41ada9df7f mcb: Fix error handling in mcb_pci_probe()
If a MCB PCI Carrier device is IO mapped insted of memory-mapped,
the memory of the PCI device is still not unmapped.

Also the patch adds deallocation of the bus
if chameleon_parse_cells() fails.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-10-05 05:10:01 +01:00
Dan Carpenter
d86fb45b5c mcb: request_mem_region() returns NULL on error
The code here is checking for IS_ERR() when request_mem_region() only
returns NULL on error and never an ERR_PTR.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-04-03 16:15:30 +02:00
Johannes Thumshirn
a48742bce1 mcb: Fix error path of mcb_pci_probe
If a MCB PCI Carrier device is IO mapped insted of memory-mapped (which is
currently unsupported by the upstream driver) the probe function bails out
with -ENOTSUPP.

In this case the memory of the PCI device was not unmapped.
Also rename error label to reflect what will happen at the destination (suggested
by Julia Lawall <julia.lawall@lip6.fr>.

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@men.de>
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-02-03 15:48:51 -08:00
Johannes Thumshirn
7b7c54914f mcb: mcb-pci: Only remap the 1st 0x200 bytes of BAR 0
Currently it is not possible to have a kernel with built-in MCB attached
devices. This results out of the fact that mcb-pci requests PCI BAR 0, then
parses the chameleon table and calls the driver's probe function before
releasing BAR 0 again. When building the kernel with modules this is not a
problem (and therefore it wasn't detected by my tests yet).

A solution is to only remap the 1st 0x200 bytes of a Chameleon PCI device.
0x200 bytes is the maximum size of a Chameleon v2 Table.

Also this patch stops disabling the PCI device on successful registration of MCB
devices.

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@men.de>
Suggested-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2015-01-09 15:46:37 -08:00
Johannes Thumshirn
4ec65b77c6 mcb: Add support for shared PCI IRQs
Add support for shared PCI IRQs to mcb and mcb-pci.

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@men.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-05-27 17:38:11 -07:00
Christoph Jaeger
7c73528273 drivers: mcb: fix memory leak in chameleon_parse_cells() error path
chameleon_parse_cells() bails out if chameleon descriptor type is
invalid but does not free the storage 'header' points to.

Signed-off-by: Christoph Jaeger <christophjaeger@linux.com>
Signed-off-by: Johannes Thumshirn <johannes.thumshirn@men.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-04-16 12:28:47 -07:00
Johannes Thumshirn
07792c7e10 drivers: mcb: Fix build error discovered by 0-day bot
Make mcb depend on HAS_IOMEM and mcb-pci depend on PCI. This fixes build errors
discovered by the 0-day kernel build testing system.

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@men.de>
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-03-09 10:22:46 -07:00
Paul Gortmaker
0db03f3f0c drivers/mcb: do not set default Kconfig variable to modular
Coverage builds found this build fail in ARM ebsa110_defconfig:

drivers/mcb/mcb-parse.c: In function 'chameleon_parse_cells':
drivers/mcb/mcb-parse.c:105:2: error: implicit declaration of function 'memcpy_fromio' [-Werror=implicit-function-declaration]
cc1: some warnings being treated as errors
make[2]: *** [drivers/mcb/mcb-parse.o] Error 1

A simple bisect will output this:

3764e82e51 is the first bad commit
commit 3764e82e51
Author: Johannes Thumshirn <johannes.thumshirn@men.de>
Date:   Wed Feb 26 17:29:05 2014 +0100

    drivers: Introduce MEN Chameleon Bus

The above commit used "default m" which is wrong.  New drivers
should never be globally enabled with "default y/m".  Whether
this driver makes sense to build on ARM is an independent issue.

Here we delete the "default m" line, which is the equivalent of
"default n".

Cc: Johannes Thumshirn <johannes.thumshirn@men.de>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-03-08 21:55:39 -08:00
Johannes Thumshirn
b71bb86398 mcb: Add PCI carrier for MEN Chameleon Bus
Add support for MCB over PCI devices. Both PCI attached on-board Chameleon FPGAs
as well as CompactPCI based MCB carrier cards are supported with this driver.

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@men.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-02-28 15:47:12 -08:00
Johannes Thumshirn
3764e82e51 drivers: Introduce MEN Chameleon Bus
The MCB (MEN Chameleon Bus) is a Bus specific to MEN Mikroelektronik
FPGA based devices. It is used to identify MCB based IP-Cores within
an FPGA and provide the necessary framework for instantiating drivers
for these devices.

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@men.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2014-02-28 15:36:37 -08:00