linux-stable.git/drivers/uio, branch v5.13.2

uio_hv_generic: Fix another memory leak in error handling paths

2021-05-14T11:26:04+00:00

Memory allocated by 'vmbus_alloc_ring()' at the beginning of the probe
function is never freed in the error handling path.

Add the missing 'vmbus_free_ring()' call.

Note that it is already freed in the .remove function.

Fixes: cdfa835c6e5e ("uio_hv_generic: defer opening vmbus until first use")
Cc: stable 
Signed-off-by: Christophe JAILLET 
Link: https://lore.kernel.org/r/0d86027b8eeed8e6360bc3d52bcdb328ff9bdca1.1620544055.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Greg Kroah-Hartman

uio_hv_generic: Fix a memory leak in error handling paths

2021-05-14T11:26:01+00:00

If 'vmbus_establish_gpadl()' fails, the (recv|send)_gpadl will not be
updated and 'hv_uio_cleanup()' in the error handling path will not be
able to free the corresponding buffer.

In such a case, we need to free the buffer explicitly.

Fixes: cdfa835c6e5e ("uio_hv_generic: defer opening vmbus until first use")
Cc: stable 
Signed-off-by: Christophe JAILLET 
Link: https://lore.kernel.org/r/4fdaff557deef6f0475d02ba7922ddbaa1ab08a6.1620544055.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Greg Kroah-Hartman

uio/uio_pci_generic: fix return value changed in refactoring

2021-05-14T11:25:27+00:00

Commit ef84928cff58 ("uio/uio_pci_generic: use device-managed function
equivalents") was able to simplify various error paths thanks to no
longer having to clean up on the way out. Some error paths were dropped,
others were simplified. In one of those simplifications, the return
value was accidentally changed from -ENODEV to -ENOMEM. Restore the old
return value.

Fixes: ef84928cff58 ("uio/uio_pci_generic: use device-managed function equivalents")
Cc: stable 
Acked-by: Michael S. Tsirkin 
Signed-off-by: Martin Ågren 
Link: https://lore.kernel.org/r/20210422192240.1136373-1-martin.agren@gmail.com
Signed-off-by: Greg Kroah-Hartman

uio: uio_dfl: add userspace i/o driver for DFL bus

2021-03-28T12:58:18+00:00

This patch supports the DFL drivers be written in userspace. This is
realized by exposing the userspace I/O device interfaces.

The driver now only binds the ether group feature, which has no irq. So
the irq support is not implemented yet.

Reviewed-by: Tom Rix 
Signed-off-by: Xu Yilun 
Link: https://lore.kernel.org/r/1615168776-8553-2-git-send-email-yilun.xu@intel.com
Signed-off-by: Greg Kroah-Hartman

uio: uio_pci_generic: don't fail probe if pdev->irq equals to IRQ_NOTCONNECTED

2021-02-09T11:25:32+00:00

Some devices use 255 as default value of Interrupt Line register, and this
maybe causes pdev->irq is set as IRQ_NOTCONNECTED in some scenarios. For
example, NVMe controller connects to Intel Volume Management Device (VMD).
In this situation, IRQ_NOTCONNECTED means INTx line is not connected, not
fault. If bind uio_pci_generic to these devices, uio frame will return
-ENOTCONN through request_irq.

This patch allows binding uio_pci_generic to device with dev->irq of
IRQ_NOTCONNECTED.

Acked-by: Kyungsan Kim 
Signed-off-by: Jie Li 
Link: https://lore.kernel.org/r/1612153559-17028-1-git-send-email-jie6.li@samsung.com
Signed-off-by: Greg Kroah-Hartman

uio: uio_hv_generic: use devm_kzalloc() for private data alloc

2020-12-09T18:59:00+00:00

This is a minor cleanup for the management of the private object of this
driver. The allocation can be tied to the life-time of the hv_device
object.
This cleans up a bit the exit & error paths, since the object doesn't need
to be explicitly free'd anymore.

Signed-off-by: Alexandru Ardelean 
Link: https://lore.kernel.org/r/20201119154903.82099-4-alexandru.ardelean@analog.com
Signed-off-by: Greg Kroah-Hartman

uio: uio_fsl_elbc_gpcm: use device-managed allocators

2020-12-09T18:59:00+00:00

This change moves all the simple allocations to use device-managed
allocator functions. This way their life-time is tied to the
platform_device object, so when this gets free'd these allocations also get
cleaned up.
The final effect is that error & exit paths get cleaned up a bit.

Signed-off-by: Alexandru Ardelean 
Link: https://lore.kernel.org/r/20201119154903.82099-3-alexandru.ardelean@analog.com
Signed-off-by: Greg Kroah-Hartman

uio: uio_aec: use devm_kzalloc() for uio_info object

2020-12-09T18:59:00+00:00

The uio_info object is free'd last, so it's life-time is tied PCI device
object. Using devm_kzalloc() cleans up the error path a bit and the exit
path.

Signed-off-by: Alexandru Ardelean 
Link: https://lore.kernel.org/r/20201119154903.82099-2-alexandru.ardelean@analog.com
Signed-off-by: Greg Kroah-Hartman

uio: uio_cif: use devm_kzalloc() for uio_info object

2020-12-09T18:59:00+00:00

The uio_info object is free'd last, so it's life-time is tied PCI device
object. Using devm_kzalloc() cleans up the error path a bit and the exit
path.

Signed-off-by: Alexandru Ardelean 
Link: https://lore.kernel.org/r/20201119154903.82099-1-alexandru.ardelean@analog.com
Signed-off-by: Greg Kroah-Hartman

uio: uio_netx: use devm_kzalloc() for or uio_info object

2020-12-09T18:58:54+00:00

This change uses the devm_kzalloc() function to tie the life-time of the
uio_info object to PCI device. This cleans up the exit & error path a bit.

Signed-off-by: Alexandru Ardelean 
Link: https://lore.kernel.org/r/20201120084207.50736-3-alexandru.ardelean@analog.com
Signed-off-by: Greg Kroah-Hartman