platform/kernel/linux-rpi.git
6 years agoscsi: lpfc: Correct target queue depth application changes
James Smart [Mon, 9 Apr 2018 21:24:22 +0000 (14:24 -0700)]
scsi: lpfc: Correct target queue depth application changes

The max_scsicmpl_time parameter can be used to perform scsi cmd queue
depth mgmt based on io completion time: the queue depth is reduced to
make completion time shorter. However, as soon as an io completes and
the completion time is within limits, the code immediately bumps the
queue depth limit back up to the target queue depth. Thus the procedure
restarts, effectively limiting the usefulness of adjusting queue depth
to help completion time.

This patch makes the following changes:

 - Removes the code at io completion that resets the queue depth as soon
   as within limits.

 - As the code removed was where the target queue depth was first
   applied, change target queue depth application so that it occurs when
   the parameter is changed.

 - Makes target queue depth a standard parameter: both a module
   parameter and a sysfs parameter.

 - Optimizes the command pending count by using atomics rather than
   locks.

 - Updates the debugfs nodelist stats to allow better debugging of
   pending command counts.

Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: lpfc: Fix multiple PRLI completion error path
James Smart [Mon, 9 Apr 2018 21:24:21 +0000 (14:24 -0700)]
scsi: lpfc: Fix multiple PRLI completion error path

Nodelist entry for SCSI array ends up in UNMAPPED state. This is due to
illegal discovery State machine transition because of two PRLIs and the
first one failing with LS_RJT. Also, the error path was designed
assuming the PRLIs complete in the order they were sent, FCP first, then
NVME. In a failing case, the array thinks about the first PRLI (FCP),
but issues LS_RJT for the 2nd PRLI immediately.

Fix PRLI completion error path for the ordering expectation.  Ensure the
discovery state machine update is not set until all outstanding PRLIs
are complete.

Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: megaraid_sas: driver version upgrade
Shivasharan S [Fri, 6 Apr 2018 09:02:12 +0000 (02:02 -0700)]
scsi: megaraid_sas: driver version upgrade

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: megaraid_sas: Increase timeout by 1 sec for non-RAID fastpath IOs
Shivasharan S [Fri, 6 Apr 2018 09:02:11 +0000 (02:02 -0700)]
scsi: megaraid_sas: Increase timeout by 1 sec for non-RAID fastpath IOs

Hardware could time out Fastpath IOs one second earlier than the timeout
provided by the host.

For non-RAID devices, driver provides timeout value based on OS provided
timeout value. Under certain scenarios, if the OS provides a timeout
value of 1 second, due to above behavior hardware will timeout
immediately.

Increase timeout value for non-RAID fastpath IOs by 1 second.

Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: megaraid_sas: Use zeroing memory allocator than allocator/memset
Himanshu Jha [Fri, 6 Apr 2018 09:02:10 +0000 (02:02 -0700)]
scsi: megaraid_sas: Use zeroing memory allocator than allocator/memset

Use pci_zalloc_consistent for allocating zeroed memory and remove
unnecessary memset function.

Done using Coccinelle.
Generated by: scripts/coccinelle/api/alloc/kzalloc-simple.cocci

Suggested-by: Luis R. Rodriguez <mcgrof@kernel.org>
Signed-off-by: Himanshu Jha <himanshujha199640@gmail.com>
Signed-off-by: Shivasharan S <shivasharan.srikanteshwara@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: netvsc: Use the vmbus function to calculate ring buffer percentage
Long Li [Wed, 28 Mar 2018 00:48:39 +0000 (17:48 -0700)]
scsi: netvsc: Use the vmbus function to calculate ring buffer percentage

In Vmbus, we have defined a function to calculate available ring buffer
percentage to write.

Use that function and remove netvsc's private version.

[mkp: typo]

Signed-off-by: Long Li <longli@microsoft.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: vmbus: Add function to report available ring buffer to write in total ring...
Long Li [Wed, 28 Mar 2018 00:48:38 +0000 (17:48 -0700)]
scsi: vmbus: Add function to report available ring buffer to write in total ring size percentage

Netvsc has a function to calculate how much ring buffer in percentage is
available to write. This function is also useful for storvsc and other
vmbus devices.

Define a similar function in vmbus to be used by other vmbus devices.

Signed-off-by: Long Li <longli@microsoft.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: libsas: add transport class for ATA devices
Jason Yan [Mon, 26 Mar 2018 09:27:41 +0000 (17:27 +0800)]
scsi: libsas: add transport class for ATA devices

Now ata devices attached with sas controller do not have transport
class, so that we can not see any information of these ata devices in
/sys/class/ata_port(or ata_link or ata_device).

Add transport class for the ata devices attached with sas controller.
The /sys/class directory will show the infomation of the ata devices
as follows:

localhost:/sys/class # ls ata*
ata_device:
dev1.0  dev2.0

ata_link:
link1  link2

ata_port:
ata1  ata2

No functional change of the device scanning and io path. The ata
transport class was deleted when destroying the sas devices.

Signed-off-by: Jason Yan <yanaijie@huawei.com>
CC: Dan Williams <dan.j.williams@intel.com>
CC: Tejun Heo <tj@kernel.org>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: hisi_sas: remove some unneeded structure members
John Garry [Fri, 23 Mar 2018 16:05:15 +0000 (00:05 +0800)]
scsi: hisi_sas: remove some unneeded structure members

This patch removes unneeded structure elements:

- hisi_sas_phy.dev_sas_addr: only ever written
- Also remove associated function which writes it,
  hisi_sas_init_add().

- hisi_sas_device.attached_phy: only ever written
- Also remove code to set it in hisi_sas_dev_found()

Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: hisi_sas: print device id for errors
John Garry [Fri, 23 Mar 2018 16:05:14 +0000 (00:05 +0800)]
scsi: hisi_sas: print device id for errors

When we find an erroneous slot completion, to help aid debugging add the
device index to the current debug log.

Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: hisi_sas: check IPTT is valid before using it for v3 hw
Xiaofei Tan [Fri, 23 Mar 2018 16:05:13 +0000 (00:05 +0800)]
scsi: hisi_sas: check IPTT is valid before using it for v3 hw

There is a bug of v3 hw development version. When AXI error happen, hw
may return an abnormal CQ that IPTT value is 0xffff.  This will cause
IPTT out-of-bounds reference.

This patch adds a check of IPTT in cq_tasklet_v3_hw() and discards
invalid slot. This workaround scheme is just to enhance fault-tolerance
of the driver. So, we will apply this scheme for all version of v3 hw,
although release version has fixed this SoC bug.

Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: hisi_sas: consolidate command check in hisi_sas_get_ata_protocol()
Xiaofei Tan [Fri, 23 Mar 2018 16:05:12 +0000 (00:05 +0800)]
scsi: hisi_sas: consolidate command check in hisi_sas_get_ata_protocol()

Currently we check the fis->command value in 2 locations in
hisi_sas_get_ata_protocol() switch statement. Fix this by consolidating
the check for fis->command value to 1 location only.

Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: hisi_sas: use dma_zalloc_coherent()
Xiang Chen [Fri, 23 Mar 2018 16:05:11 +0000 (00:05 +0800)]
scsi: hisi_sas: use dma_zalloc_coherent()

This is a warning coming from Coccinelle, and need to use new interface
dma_zalloc_coherent() instead of dma_alloc_coherent()/memset().

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: hisi_sas: delete timer when removing hisi_sas driver
Xiang Chen [Fri, 23 Mar 2018 16:05:10 +0000 (00:05 +0800)]
scsi: hisi_sas: delete timer when removing hisi_sas driver

Delete timer for v1 and v3 hw when removing hisi_sas driver.

Signed-off-by: Xiang chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: hisi_sas: update RAS feature for later revision of v3 HW
Xiaofei Tan [Fri, 23 Mar 2018 16:05:09 +0000 (00:05 +0800)]
scsi: hisi_sas: update RAS feature for later revision of v3 HW

There is an modification for later revision of v3 hw. More HW errors are
reported through RAS interrupt. These errors were originally reported
only through MSI.

When report to RAS, some combinations are done to port AXI errors and
FIFO OMIT errors. For example, each port has 4 AXI errors, and they are
combined to one when report to RAS.

This patch does two things:

1. Enable RAS interrupt of these errors and handle them in PCI
   error handlers.

2. Disable MSI interrupts of these errors for this later revision hw.

Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: hisi_sas: make SAS address of SATA disks unique
Xiang Chen [Fri, 23 Mar 2018 16:05:08 +0000 (00:05 +0800)]
scsi: hisi_sas: make SAS address of SATA disks unique

When directly connected with SATA disks in different SAS cores, fill SAS
address with scsi_host's id to make it's fake SAS address unique.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Handle spurious interrupts
Uma Krishnan [Mon, 26 Mar 2018 16:35:42 +0000 (11:35 -0500)]
scsi: cxlflash: Handle spurious interrupts

The following Oops can occur when there is heavy I/O traffic and the host is
reset by a tool such as sg_reset.

[c000200fff3fbc90c00800001690117c process_cmd_doneq+0x104/0x500
                                       [cxlflash] (unreliable)
[c000200fff3fbd80c008000016901648 cxlflash_rrq_irq+0xd0/0x150 [cxlflash]
[c000200fff3fbde0c000000000193130 __handle_irq_event_percpu+0xa0/0x310
[c000200fff3fbea0c0000000001933d8 handle_irq_event_percpu+0x38/0x90
[c000200fff3fbee0c000000000193494 handle_irq_event+0x64/0xb0
[c000200fff3fbf10c000000000198ea0 handle_fasteoi_irq+0xc0/0x230
[c000200fff3fbf40c00000000019182c generic_handle_irq+0x4c/0x70
[c000200fff3fbf60c00000000001794c __do_irq+0x7c/0x1c0
[c000200fff3fbf90c00000000002a390 call_do_irq+0x14/0x24
[c000200e5828fab0c000000000017b2c do_IRQ+0x9c/0x130
[c000200e5828fb00c000000000009b04 h_virt_irq_common+0x114/0x120

When a context is reset, the pending commands are flushed and the AFU is
notified. Before the AFU handles this request there could be command
completion interrupts queued to PHB which are yet to be delivered to the
context. In this scenario, a context could receive an interrupt for a command
that has been flushed, leading to a possible crash when the memory for the
flushed command is accessed.

To resolve this problem, a boolean will indicate if the hardware queue is
ready to process interrupts or not. This can be evaluated in the interrupt
handler before proessing an interrupt.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Remove commmands from pending list on timeout
Uma Krishnan [Mon, 26 Mar 2018 16:35:34 +0000 (11:35 -0500)]
scsi: cxlflash: Remove commmands from pending list on timeout

The following Oops can occur if an internal command sent to the AFU does not
complete within the timeout:

[c000000ff101b810c008000016020d94 term_mc+0xfc/0x1b0 [cxlflash]
[c000000ff101b8a0c008000016020fb0 term_afu+0x168/0x280 [cxlflash]
[c000000ff101b930c0080000160232ec cxlflash_pci_error_detected+0x184/0x230
                                       [cxlflash]
[c000000ff101b9e0c00800000d95d468 cxl_vphb_error_detected+0x90/0x150[cxl]
[c000000ff101ba20c00800000d95f27c cxl_pci_error_detected+0xa4/0x240 [cxl]
[c000000ff101bac0c00000000003eaf8 eeh_report_error+0xd8/0x1b0
[c000000ff101bb20c00000000003d0b8 eeh_pe_dev_traverse+0x98/0x170
[c000000ff101bbb0c00000000003f438 eeh_handle_normal_event+0x198/0x580
[c000000ff101bc60c00000000003fba4 eeh_handle_event+0x2a4/0x338
[c000000ff101bd10c0000000000400b8 eeh_event_handler+0x1f8/0x200
[c000000ff101bdc0c00000000013da48 kthread+0x1a8/0x1b0
[c000000ff101be30c00000000000b528 ret_from_kernel_thread+0x5c/0xb4

When an internal command times out, the command buffer is freed while it is
still in the pending commands list of the context. This corrupts the list and
when the context is cleaned up, a crash is encountered.

To resolve this issue, when an AFU command or TMF command times out, the
command should be deleted from the hardware queue pending command list before
freeing the buffer.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Synchronize reset and remove ops
Uma Krishnan [Mon, 26 Mar 2018 16:35:27 +0000 (11:35 -0500)]
scsi: cxlflash: Synchronize reset and remove ops

The following Oops can be encountered if a device removal or system shutdown
is initiated while an EEH recovery is in process:

[c000000ff2f479c0c008000015256f18 cxlflash_pci_slot_reset+0xa0/0x100
                                      [cxlflash]
[c000000ff2f47a30c00800000dae22e0 cxl_pci_slot_reset+0x168/0x290 [cxl]
[c000000ff2f47ae0c00000000003ef1c eeh_report_reset+0xec/0x170
[c000000ff2f47b20c00000000003d0b8 eeh_pe_dev_traverse+0x98/0x170
[c000000ff2f47bb0c00000000003f80c eeh_handle_normal_event+0x56c/0x580
[c000000ff2f47c60c00000000003fba4 eeh_handle_event+0x2a4/0x338
[c000000ff2f47d10c0000000000400b8 eeh_event_handler+0x1f8/0x200
[c000000ff2f47dc0c00000000013da48 kthread+0x1a8/0x1b0
[c000000ff2f47e30c00000000000b528 ret_from_kernel_thread+0x5c/0xb4

The remove handler frees AFU memory while the EEH recovery is in progress,
leading to a race condition. This can result in a crash if the recovery thread
tries to access this memory.

To resolve this issue, the cxlflash remove handler will evaluate the device
state and yield to any active reset or probing threads.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Enable OCXL operations
Uma Krishnan [Mon, 26 Mar 2018 16:35:21 +0000 (11:35 -0500)]
scsi: cxlflash: Enable OCXL operations

This commit enables the OCXL operations for the OCXL devices.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support AFU reset
Uma Krishnan [Mon, 26 Mar 2018 16:35:15 +0000 (11:35 -0500)]
scsi: cxlflash: Support AFU reset

The cxlflash core driver resets the AFU when the master contexts are created
in the initialization or recovery paths. Today, the OCXL provider service to
perform this operation is pending implementation.  To avoid a crash due to a
missing fop, log an error once and return success to continue with execution.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Register for translation errors
Uma Krishnan [Mon, 26 Mar 2018 16:35:07 +0000 (11:35 -0500)]
scsi: cxlflash: Register for translation errors

While enabling a context on the link, a predefined callback can be registered
with the OCXL provider services to be notified on translation errors. These
errors can in turn be passed back to the user on a read operation.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Introduce OCXL context state machine
Uma Krishnan [Mon, 26 Mar 2018 16:35:00 +0000 (11:35 -0500)]
scsi: cxlflash: Introduce OCXL context state machine

In order to protect the OCXL hardware contexts from getting clobbered, a
simple state machine is added to indicate when a context is in open, close or
start state. The expected states are validated throughout the code to prevent
illegal operations on a context. A mutex is added to protect writes to the
context state field.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Update synchronous interrupt status bits
Uma Krishnan [Mon, 26 Mar 2018 16:34:55 +0000 (11:34 -0500)]
scsi: cxlflash: Update synchronous interrupt status bits

The SISLite specification has been updated to define new synchronous interrupt
status bits. These bits are set by the AFU when a given PASID or EA is bad and
a synchronous interrupt is triggered.

The SISLite header file is updated to support these new bits. Note that there
are also some formatting updates to some of the existing bits to allow all of
the definitions to line up uniformly.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Setup LISNs for master contexts
Uma Krishnan [Mon, 26 Mar 2018 16:34:48 +0000 (11:34 -0500)]
scsi: cxlflash: Setup LISNs for master contexts

Similar to user contexts, master contexts also require that the per-context
LISN registers be programmed for certain AFUs. The mapped trigger page is
obtained from underlying transport and registered with AFU for each master
context.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Setup LISNs for user contexts
Uma Krishnan [Mon, 26 Mar 2018 16:34:42 +0000 (11:34 -0500)]
scsi: cxlflash: Setup LISNs for user contexts

The SISLite specification has been updated for OCXL to support communicating
data to generate AFU interrupts to the AFU. This includes a new capability bit
that is advertised for OCXL AFUs and new registers to hold the object handle
and translation PASID of each interrupt. For Power, the object handle is the
mapped trigger page. Note that because these mappings are kernel only, the
PASID of a kernel context must be used to satisfy the translation.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Introduce object handle fop
Uma Krishnan [Mon, 26 Mar 2018 16:34:35 +0000 (11:34 -0500)]
scsi: cxlflash: Introduce object handle fop

OCXL requires that AFUs use an opaque object handle to represent an AFU
interrupt. The specification does not provide a common means to communicate
the object handle to the AFU - each AFU must define this within the AFU
specification. To support this model, the object handle must be passed back to
the core driver as it manages the AFU specification (SISLite) for cxlflash.
Note that for Power systems, the object handle is the effective address of the
trigger page.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support file descriptor mapping
Uma Krishnan [Mon, 26 Mar 2018 16:34:27 +0000 (11:34 -0500)]
scsi: cxlflash: Support file descriptor mapping

The cxlflash core fop API requires a way to invoke the fault and release
handlers of underlying transports using their native file-based APIs. This
provides the core with the ability to insert selectively itself into the
processing stream of these operations for cleanup. Implement these two fops to
map and release when requested.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support adapter context mmap and release
Uma Krishnan [Mon, 26 Mar 2018 16:34:20 +0000 (11:34 -0500)]
scsi: cxlflash: Support adapter context mmap and release

The cxlflash userspace API requires that users be able to mmap and release the
adapter context. Support mapping by implementing the AFU mmap fop to map the
context MMIO space and install the corresponding page table entry upon page
fault. Similarly, implement the AFU release fop to terminate and clean up the
context when invoked.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support adapter context reading
Uma Krishnan [Mon, 26 Mar 2018 16:34:12 +0000 (11:34 -0500)]
scsi: cxlflash: Support adapter context reading

The cxlflash userspace API requires that users be able to read the adapter
context for any pending events or interrupts from the AFU. Support reading
various events by implementing the AFU read fop to copy out event data.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support adapter context polling
Uma Krishnan [Mon, 26 Mar 2018 16:34:03 +0000 (11:34 -0500)]
scsi: cxlflash: Support adapter context polling

The cxlflash userspace API requires that users be able to poll the adapter
context for any pending events or interrupts from the AFU. Support polling on
various events by implementing the AFU poll fop using a waitqueue.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support starting user contexts
Uma Krishnan [Mon, 26 Mar 2018 16:33:55 +0000 (11:33 -0500)]
scsi: cxlflash: Support starting user contexts

User contexts request interrupts and are started using the "start work"
interface. Populate the start_work() fop to allocate and map interrupts before
starting the user context. As part of starting the context, update the user
process identification logic to properly derive the data required by the
SPA. Also, introduce a skeleton interrupt handler using a bitmap, flag, and
spinlock to track interrupts. This handler will be expanded in future commits.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support AFU interrupt mapping and registration
Uma Krishnan [Mon, 26 Mar 2018 16:33:48 +0000 (11:33 -0500)]
scsi: cxlflash: Support AFU interrupt mapping and registration

Add support to map and unmap the irq space and manage irq registrations with
the kernel for each allocated AFU interrupt. Also support mapping the physical
trigger page to obtain an effective address that will be provided to the
cxlflash core in a future commit.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support AFU interrupt management
Uma Krishnan [Mon, 26 Mar 2018 16:33:41 +0000 (11:33 -0500)]
scsi: cxlflash: Support AFU interrupt management

Add support to allocate and free AFU interrupts using the OCXL provider
services. The trigger page returned upon successful allocation will be mapped
and exposed to the cxlflash core in a future commit.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support process element lifecycle
Uma Krishnan [Mon, 26 Mar 2018 16:33:35 +0000 (11:33 -0500)]
scsi: cxlflash: Support process element lifecycle

As part of the context lifecycle, the associated process element within the
Shared Process Area (SPA) of the link must be updated. Each process is defined
by various parameters (pid, tid, PASID mm) that are stored in the SPA upon
starting a context and invalidated when a context is stopped.

Use the OCXL provider services to configure the SPA with the appropriate data
that is unique to the process when starting a context. Initially only kernel
contexts are supported and therefore these process values are not applicable.
Note that the OCXL service used has an optional callback for translation fault
error notification. While not used here, it will be expanded in a future
commit.

Also add a service to stop a context by terminating the corresponding PASID
and remove the process element from the SPA.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Setup OCXL transaction layer
Uma Krishnan [Mon, 26 Mar 2018 16:33:28 +0000 (11:33 -0500)]
scsi: cxlflash: Setup OCXL transaction layer

The first function of the link needs to configure the transaction layer
between the host and device. This is accomplished by a call to the OCXL
provider services.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Setup function OCXL link
Uma Krishnan [Mon, 26 Mar 2018 16:33:21 +0000 (11:33 -0500)]
scsi: cxlflash: Setup function OCXL link

After reading and modifying the function configuration, setup the OCXL link
using the OCXL provider services. The link is released when the adapter is
unconfigured.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support reading adapter VPD data
Uma Krishnan [Mon, 26 Mar 2018 16:33:14 +0000 (11:33 -0500)]
scsi: cxlflash: Support reading adapter VPD data

Use the PCI VPD services to support reading the VPD data of the underlying
adapter.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support AFU state toggling
Uma Krishnan [Mon, 26 Mar 2018 16:33:05 +0000 (11:33 -0500)]
scsi: cxlflash: Support AFU state toggling

The AFU should be enabled following a successful configuration and disabled
near the end of the cleanup path.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support process specific mappings
Uma Krishnan [Mon, 26 Mar 2018 16:32:56 +0000 (11:32 -0500)]
scsi: cxlflash: Support process specific mappings

Once the context is started, the assigned MMIO space can be mapped and
unmapped. Provide means to map and unmap the context MMIO space.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support starting an adapter context
Uma Krishnan [Mon, 26 Mar 2018 16:32:48 +0000 (11:32 -0500)]
scsi: cxlflash: Support starting an adapter context

Once the adapter context is created, it needs to be started by assigning the
MMIO space for the context and by enabling the process element in the
link. This commit adds the skeleton for starting the context and assigns the
context specific MMIO space. Master contexts have access to the global MMIO
space while the rest have access to the context specific space.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: MMIO map the AFU
Uma Krishnan [Mon, 26 Mar 2018 16:32:37 +0000 (11:32 -0500)]
scsi: cxlflash: MMIO map the AFU

When the AFU is configured, the global and per process MMIO regions are
presented by the configuration space. Save these regions and map the global
MMIO region that is used to access all of the control and provisioning data in
the AFU.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support image reload policy modification
Uma Krishnan [Mon, 26 Mar 2018 16:32:29 +0000 (11:32 -0500)]
scsi: cxlflash: Support image reload policy modification

On a PERST, the AFU image can be reloaded or left intact. Provide means to set
this image reload policy.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support adapter context discovery
Uma Krishnan [Mon, 26 Mar 2018 16:32:20 +0000 (11:32 -0500)]
scsi: cxlflash: Support adapter context discovery

Provide means to obtain the process element of an adapter context as well as
locate an adapter context by file.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Support adapter file descriptors for OCXL
Uma Krishnan [Mon, 26 Mar 2018 16:32:09 +0000 (11:32 -0500)]
scsi: cxlflash: Support adapter file descriptors for OCXL

Allocate a file descriptor for an adapter context when requested. In order to
allocate inodes for the file descriptors, a pseudo filesystem is created and
used.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Use IDR to manage adapter contexts
Uma Krishnan [Mon, 26 Mar 2018 16:32:00 +0000 (11:32 -0500)]
scsi: cxlflash: Use IDR to manage adapter contexts

A range of PASIDs are used as identifiers for the adapter contexts. These
contexts may be destroyed and created randomly. Use an IDR to keep track of
contexts that are in use and assign a unique identifier to new ones.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Adapter context support for OCXL
Uma Krishnan [Mon, 26 Mar 2018 16:31:53 +0000 (11:31 -0500)]
scsi: cxlflash: Adapter context support for OCXL

Add support to create and release the adapter contexts for OCXL and provide
means to specify certain contexts as a master.

The existing cxlflash core has a design requirement that each host will have a
single host context available by default. To satisfy this requirement, one
host adapter context is created when the hardware AFU is initialized. This is
returned by the get_context() fop.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Setup AFU PASID
Uma Krishnan [Mon, 26 Mar 2018 16:31:44 +0000 (11:31 -0500)]
scsi: cxlflash: Setup AFU PASID

Per the OCXL specification, the maximum PASID supported by the AFU is
indicated by a field within the configuration space. Similar to acTags,
implementations can choose to use any sub-range of PASID within their assigned
range. For cxlflash, the entire range is used.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Setup AFU acTag range
Uma Krishnan [Mon, 26 Mar 2018 16:31:36 +0000 (11:31 -0500)]
scsi: cxlflash: Setup AFU acTag range

The OCXL specification supports distributing acTags amongst different AFUs and
functions on the link. As cxlflash devices are expected to only support a
single AFU per function, the entire range that was assigned to the function is
also assigned to the AFU.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Read host AFU configuration
Uma Krishnan [Mon, 26 Mar 2018 16:31:29 +0000 (11:31 -0500)]
scsi: cxlflash: Read host AFU configuration

The host AFU configuration is read on the initialization path to identify the
features and configuration of the AFU. This data is cached for use in later
configuration steps.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Setup function acTag range
Uma Krishnan [Mon, 26 Mar 2018 16:31:21 +0000 (11:31 -0500)]
scsi: cxlflash: Setup function acTag range

The OCXL specification supports distributing acTags amongst different AFUs and
functions on the link. The platform-specific acTag range for the link is
obtained using the OCXL provider services and then assigned to the host
function based on implementation.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Read host function configuration
Uma Krishnan [Mon, 26 Mar 2018 16:31:09 +0000 (11:31 -0500)]
scsi: cxlflash: Read host function configuration

Per the OCXL specification, the underlying host can have multiple AFUs per
function with each function supporting its own configuration. The host
function configuration is read on the initialization path to evaluate the
number of functions present and identify the features and configuration of the
functions present. This data is cached for use in later configuration
steps. Note that for the OCXL hardware supported by the cxlflash driver, only
one AFU per function is expected.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Hardware AFU for OCXL
Uma Krishnan [Mon, 26 Mar 2018 16:31:01 +0000 (11:31 -0500)]
scsi: cxlflash: Hardware AFU for OCXL

When an adapter is initialized, transport specific configuration and MMIO
mapping details need to be saved. For CXL, this data is managed by the
underlying kernel module. To maintain a separation between the cxlflash core
and underlying transports, introduce a new structure to store data specific to
the OCXL AFU.

Initially only the pointers to underlying PCI and generic devices are added to
this new structure - it will be expanded further in future commits. Services
to create and destroy this hardware AFU are added and integrated in the probe
and exit paths of the driver.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Introduce OCXL backend
Uma Krishnan [Mon, 26 Mar 2018 16:30:51 +0000 (11:30 -0500)]
scsi: cxlflash: Introduce OCXL backend

Add initial infrastructure to support a new cxlflash transport, OCXL.

Claim a dependency on OCXL and add a new file, ocxl_hw.c, which will host the
backend routines that are specific to OCXL.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Add argument identifier names
Uma Krishnan [Mon, 26 Mar 2018 16:30:37 +0000 (11:30 -0500)]
scsi: cxlflash: Add argument identifier names

Checkpatch throws a warning when the argument identifier names are not
included in the function definitions.

To avoid these warnings, argument identifiers are added in the existing
function definitions.

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Avoid clobbering context control register value
Matthew R. Ochs [Mon, 26 Mar 2018 16:30:22 +0000 (11:30 -0500)]
scsi: cxlflash: Avoid clobbering context control register value

The SISLite specification originally defined the context control register with
a single field of bits to represent the LISN and also stipulated that the
register reset value be 0. The cxlflash driver took advantage of this when
programming the LISN for the master contexts via an unconditional write - no
other bits were preserved.

When unmap support was added, SISLite was updated to define bit 0 of the
context control register as a way for the AFU to notify the context owner that
unmap operations were supported. Thus the assumptions under which the register
is setup changed and the existing unconditional write is clobbering the unmap
state for master contexts. This is presently not an issue due to the order in
which the context control register is programmed in relation to the unmap bit
being queried but should be addressed to avoid a future regression in the
event this code is moved elsewhere.

To remedy this issue, preserve the bits when programming the LISN field in the
context control register. Since the LISN will now be programmed using a read
value, assert that the initial state of the LISN field is as described in
SISLite (0).

Signed-off-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoscsi: cxlflash: Preserve number of interrupts for master contexts
Uma Krishnan [Mon, 26 Mar 2018 16:29:56 +0000 (11:29 -0500)]
scsi: cxlflash: Preserve number of interrupts for master contexts

The number of interrupts requested for user contexts are stored in the context
specific structures and utilized to manage the interrupts. For the master
contexts, this number is only used once and therefore not saved.

To prepare for future commits where the number of interrupts will be required
in more than one place, preserve the value in the master context structure.

[mkp: typo in comment]

Signed-off-by: Uma Krishnan <ukrishn@linux.vnet.ibm.com>
Acked-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
6 years agoLinux 4.17-rc1
Linus Torvalds [Mon, 16 Apr 2018 01:24:20 +0000 (18:24 -0700)]
Linux 4.17-rc1

6 years agoMerge tag 'for-4.17-part2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave...
Linus Torvalds [Mon, 16 Apr 2018 01:08:35 +0000 (18:08 -0700)]
Merge tag 'for-4.17-part2-tag' of git://git./linux/kernel/git/kdave/linux

Pull more btrfs updates from David Sterba:
 "We have queued a few more fixes (error handling, log replay,
  softlockup) and the rest is SPDX updates that touche almost all files
  so the diffstat is long"

* tag 'for-4.17-part2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: Only check first key for committed tree blocks
  btrfs: add SPDX header to Kconfig
  btrfs: replace GPL boilerplate by SPDX -- sources
  btrfs: replace GPL boilerplate by SPDX -- headers
  Btrfs: fix loss of prealloc extents past i_size after fsync log replay
  Btrfs: clean up resources during umount after trans is aborted
  btrfs: Fix possible softlock on single core machines
  Btrfs: bail out on error during replay_dir_deletes
  Btrfs: fix NULL pointer dereference in log_dir_items

6 years agoMerge tag '4.17-rc1SMB3-Fixes' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Mon, 16 Apr 2018 01:06:22 +0000 (18:06 -0700)]
Merge tag '4.17-rc1SMB3-Fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull cifs fixes from Steve French:
 "SMB3 fixes, a few for stable, and some important cleanup work from
  Ronnie of the smb3 transport code"

* tag '4.17-rc1SMB3-Fixes' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: change validate_buf to validate_iov
  cifs: remove rfc1002 hardcoded constants from cifs_discard_remaining_data()
  cifs: Change SMB2_open to return an iov for the error parameter
  cifs: add resp_buf_size to the mid_q_entry structure
  smb3.11: replace a 4 with server->vals->header_preamble_size
  cifs: replace a 4 with server->vals->header_preamble_size
  cifs: add pdu_size to the TCP_Server_Info structure
  SMB311: Improve checking of negotiate security contexts
  SMB3: Fix length checking of SMB3.11 negotiate request
  CIFS: add ONCE flag for cifs_dbg type
  cifs: Use ULL suffix for 64-bit constant
  SMB3: Log at least once if tree connect fails during reconnect
  cifs: smb2pdu: Fix potential NULL pointer dereference

6 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Mon, 16 Apr 2018 00:24:12 +0000 (17:24 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "This is a set of minor (and safe changes) that didn't make the initial
  pull request plus some bug fixes.

  The status handling code is actually a running regression from the
  previous merge window which had an incomplete fix (now reverted) and
  most of the remaining bug fixes are for problems older than the
  current merge window"

[ Side note: this merge also takes the base kernel git repository to 6+
  million objects for the first time. Technically we hit it a couple of
  merges ago already if you count all the tag objects, but now it
  reaches 6M+ objects reachable from HEAD.

  I was joking around that that's when I should switch to 5.0, because
  3.0 happened at the 2M mark, and 4.0 happened at 4M objects. But
  probably not, even if numerology is about as good a reason as any.

                                                              - Linus ]

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: devinfo: Add Microsoft iSCSI target to 1024 sector blacklist
  scsi: cxgb4i: silence overflow warning in t4_uld_rx_handler()
  scsi: dpt_i2o: Use after free in I2ORESETCMD ioctl
  scsi: core: Make scsi_result_to_blk_status() recognize CONDITION MET
  scsi: core: Rename __scsi_error_from_host_byte() into scsi_result_to_blk_status()
  Revert "scsi: core: return BLK_STS_OK for DID_OK in __scsi_error_from_host_byte()"
  scsi: aacraid: Insure command thread is not recursively stopped
  scsi: qla2xxx: Correct setting of SAM_STAT_CHECK_CONDITION
  scsi: qla2xxx: correctly shift host byte
  scsi: qla2xxx: Fix race condition between iocb timeout and initialisation
  scsi: qla2xxx: Avoid double completion of abort command
  scsi: qla2xxx: Fix small memory leak in qla2x00_probe_one on probe failure
  scsi: scsi_dh: Don't look for NULL devices handlers by name
  scsi: core: remove redundant assignment to shost->use_blk_mq

6 years agoMerge tag 'kbuild-v4.17-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy...
Linus Torvalds [Mon, 16 Apr 2018 00:21:30 +0000 (17:21 -0700)]
Merge tag 'kbuild-v4.17-2' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull more Kbuild updates from Masahiro Yamada:

 - pass HOSTLDFLAGS when compiling single .c host programs

 - build genksyms lexer and parser files instead of using shipped
   versions

 - rename *-asn1.[ch] to *.asn1.[ch] for suffix consistency

 - let the top .gitignore globally ignore artifacts generated by flex,
   bison, and asn1_compiler

 - let the top Makefile globally clean artifacts generated by flex,
   bison, and asn1_compiler

 - use safer .SECONDARY marker instead of .PRECIOUS to prevent
   intermediate files from being removed

 - support -fmacro-prefix-map option to make __FILE__ a relative path

 - fix # escaping to prepare for the future GNU Make release

 - clean up deb-pkg by using debian tools instead of handrolled
   source/changes generation

 - improve rpm-pkg portability by supporting kernel-install as a
   fallback of new-kernel-pkg

 - extend Kconfig listnewconfig target to provide more information

* tag 'kbuild-v4.17-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  kconfig: extend output of 'listnewconfig'
  kbuild: rpm-pkg: use kernel-install as a fallback for new-kernel-pkg
  Kbuild: fix # escaping in .cmd files for future Make
  kbuild: deb-pkg: split generating packaging and build
  kbuild: use -fmacro-prefix-map to make __FILE__ a relative path
  kbuild: mark $(targets) as .SECONDARY and remove .PRECIOUS markers
  kbuild: rename *-asn1.[ch] to *.asn1.[ch]
  kbuild: clean up *-asn1.[ch] patterns from top-level Makefile
  .gitignore: move *-asn1.[ch] patterns to the top-level .gitignore
  kbuild: add %.dtb.S and %.dtb to 'targets' automatically
  kbuild: add %.lex.c and %.tab.[ch] to 'targets' automatically
  genksyms: generate lexer and parser during build instead of shipping
  kbuild: clean up *.lex.c and *.tab.[ch] patterns from top-level Makefile
  .gitignore: move *.lex.c *.tab.[ch] patterns to the top-level .gitignore
  kbuild: use HOSTLDFLAGS for single .c executables

6 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 15 Apr 2018 23:12:35 +0000 (16:12 -0700)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Thomas Gleixner:
 "A set of fixes and updates for x86:

   - Address a swiotlb regression which was caused by the recent DMA
     rework and made driver fail because dma_direct_supported() returned
     false

   - Fix a signedness bug in the APIC ID validation which caused invalid
     APIC IDs to be detected as valid thereby bloating the CPU possible
     space.

   - Fix inconsisten config dependcy/select magic for the MFD_CS5535
     driver.

   - Fix a corruption of the physical address space bits when encryption
     has reduced the address space and late cpuinfo updates overwrite
     the reduced bit information with the original value.

   - Dominiks syscall rework which consolidates the architecture
     specific syscall functions so all syscalls can be wrapped with the
     same macros. This allows to switch x86/64 to struct pt_regs based
     syscalls. Extend the clearing of user space controlled registers in
     the entry patch to the lower registers"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/apic: Fix signedness bug in APIC ID validity checks
  x86/cpu: Prevent cpuinfo_x86::x86_phys_bits adjustment corruption
  x86/olpc: Fix inconsistent MFD_CS5535 configuration
  swiotlb: Use dma_direct_supported() for swiotlb_ops
  syscalls/x86: Adapt syscall_wrapper.h to the new syscall stub naming convention
  syscalls/core, syscalls/x86: Rename struct pt_regs-based sys_*() to __x64_sys_*()
  syscalls/core, syscalls/x86: Clean up compat syscall stub naming convention
  syscalls/core, syscalls/x86: Clean up syscall stub naming convention
  syscalls/x86: Extend register clearing on syscall entry to lower registers
  syscalls/x86: Unconditionally enable 'struct pt_regs' based syscalls on x86_64
  syscalls/x86: Use 'struct pt_regs' based syscall calling for IA32_EMULATION and x32
  syscalls/core: Prepare CONFIG_ARCH_HAS_SYSCALL_WRAPPER=y for compat syscalls
  syscalls/x86: Use 'struct pt_regs' based syscall calling convention for 64-bit syscalls
  syscalls/core: Introduce CONFIG_ARCH_HAS_SYSCALL_WRAPPER=y
  x86/syscalls: Don't pointlessly reload the system call number
  x86/mm: Fix documentation of module mapping range with 4-level paging
  x86/cpuid: Switch to 'static const' specifier

6 years agoMerge branch 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 15 Apr 2018 20:35:29 +0000 (13:35 -0700)]
Merge branch 'x86-pti-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 pti updates from Thomas Gleixner:
 "Another series of PTI related changes:

   - Remove the manual stack switch for user entries from the idtentry
     code. This debloats entry by 5k+ bytes of text.

   - Use the proper types for the asm/bootparam.h defines to prevent
     user space compile errors.

   - Use PAGE_GLOBAL for !PCID systems to gain back performance

   - Prevent setting of huge PUD/PMD entries when the entries are not
     leaf entries otherwise the entries to which the PUD/PMD points to
     and are populated get lost"

* 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/pgtable: Don't set huge PUD/PMD on non-leaf entries
  x86/pti: Leave kernel text global for !PCID
  x86/pti: Never implicitly clear _PAGE_GLOBAL for kernel image
  x86/pti: Enable global pages for shared areas
  x86/mm: Do not forbid _PAGE_RW before init for __ro_after_init
  x86/mm: Comment _PAGE_GLOBAL mystery
  x86/mm: Remove extra filtering in pageattr code
  x86/mm: Do not auto-massage page protections
  x86/espfix: Document use of _PAGE_GLOBAL
  x86/mm: Introduce "default" kernel PTE mask
  x86/mm: Undo double _PAGE_PSE clearing
  x86/mm: Factor out pageattr _PAGE_GLOBAL setting
  x86/entry/64: Drop idtentry's manual stack switch for user entries
  x86/uapi: Fix asm/bootparam.h userspace compilation errors

6 years agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 15 Apr 2018 19:43:30 +0000 (12:43 -0700)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Thomas Gleixner:
 "A few scheduler fixes:

   - Prevent a bogus warning vs. runqueue clock update flags in
     do_sched_rt_period_timer()

   - Simplify the helper functions which handle requests for skipping
     the runqueue clock updat.

   - Do not unlock the tunables mutex in the error path of the cpu
     frequency scheduler utils. Its not held.

   - Enforce proper alignement for 'struct util_est' in sched_avg to
     prevent a misalignment fault on IA64"

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/core: Force proper alignment of 'struct util_est'
  sched/core: Simplify helpers for rq clock update skip requests
  sched/rt: Fix rq->clock_update_flags < RQCF_ACT_SKIP warning
  sched/cpufreq/schedutil: Fix error path mutex unlock

6 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 15 Apr 2018 19:36:31 +0000 (12:36 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull more perf updates from Thomas Gleixner:
 "A rather large set of perf updates:

  Kernel:

   - Fix various initialization issues

   - Prevent creating [ku]probes for not CAP_SYS_ADMIN users

  Tooling:

   - Show only failing syscalls with 'perf trace --failure' (Arnaldo
     Carvalho de Melo)

            e.g: See what 'openat' syscalls are failing:

        # perf trace --failure -e openat
         762.323 ( 0.007 ms): VideoCapture/4566 openat(dfd: CWD, filename: /dev/video2) = -1 ENOENT No such file or directory
         <SNIP N /dev/videoN open attempts... sigh, where is that improvised camera lid?!? >
         790.228 ( 0.008 ms): VideoCapture/4566 openat(dfd: CWD, filename: /dev/video63) = -1 ENOENT No such file or directory
        ^C#

   - Show information about the event (freq, nr_samples, total
     period/nr_events) in the annotate --tui and --stdio2 'perf
     annotate' output, similar to the first line in the 'perf report
     --tui', but just for the samples for a the annotated symbol
     (Arnaldo Carvalho de Melo)

   - Introduce 'perf version --build-options' to show what features were
     linked, aliased as well as a shorter 'perf -vv' (Jin Yao)

   - Add a "dso_size" sort order (Kim Phillips)

   - Remove redundant ')' in the tracepoint output in 'perf trace'
     (Changbin Du)

   - Synchronize x86's cpufeatures.h, no effect on toolss (Arnaldo
     Carvalho de Melo)

   - Show group details on the title line in the annotate browser and
     'perf annotate --stdio2' output, so that the per-event columns can
     have headers (Arnaldo Carvalho de Melo)

   - Fixup vertical line separating metrics from instructions and
     cleaning unused lines at the bottom, both in the annotate TUI
     browser (Arnaldo Carvalho de Melo)

   - Remove duplicated 'samples' in lost samples warning in
     'perf report' (Arnaldo Carvalho de Melo)

   - Synchronize i915_drm.h, silencing the perf build process,
     automagically adding support for the new DRM_I915_QUERY ioctl
     (Arnaldo Carvalho de Melo)

   - Make auxtrace_queues__add_buffer() allocate struct buffer, from a
     patchkit already applied (Adrian Hunter)

   - Fix the --stdio2/TUI annotate output to include group details, be
     it for a recorded '{a,b,f}' explicit event group or when forcing
     group display using 'perf report --group' for a set of events not
     recorded as a group (Arnaldo Carvalho de Melo)

   - Fix display artifacts in the ui browser (base class for the
     annotate and main report/top TUI browser) related to the extra
     title lines work (Arnaldo Carvalho de Melo)

   - perf auxtrace refactorings, leftovers from a previously partially
     processed patchset (Adrian Hunter)

   - Fix the builtin clang build (Sandipan Das, Arnaldo Carvalho de
     Melo)

   - Synchronize i915_drm.h, silencing a perf build warning and in the
     process automagically adding support for a new ioctl command
     (Arnaldo Carvalho de Melo)

   - Fix a strncpy issue in uprobe tracing"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (36 commits)
  perf/core: Need CAP_SYS_ADMIN to create k/uprobe with perf_event_open()
  tracing/uprobe_event: Fix strncpy corner case
  perf/core: Fix perf_uprobe_init()
  perf/core: Fix perf_kprobe_init()
  perf/core: Fix use-after-free in uprobe_perf_close()
  perf tests clang: Fix function name for clang IR test
  perf clang: Add support for recent clang versions
  perf tools: Fix perf builds with clang support
  perf tools: No need to include namespaces.h in util.h
  perf hists browser: Remove leftover from row returned from refresh
  perf hists browser: Show extra_title_lines in the 'D' debug hotkey
  perf auxtrace: Make auxtrace_queues__add_buffer() do CPU filtering
  tools headers uapi: Synchronize i915_drm.h
  perf report: Remove duplicated 'samples' in lost samples warning
  perf ui browser: Fixup cleaning unused lines at the bottom
  perf annotate browser: Fixup vertical line separating metrics from instructions
  perf annotate: Show group details on the title line
  perf auxtrace: Make auxtrace_queues__add_buffer() allocate struct buffer
  perf/x86/intel: Move regs->flags EXACT bit init
  perf trace: Remove redundant ')'
  ...

6 years agoMerge branch 'efi-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 15 Apr 2018 19:32:06 +0000 (12:32 -0700)]
Merge branch 'efi-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 EFI bootup fixlet from Thomas Gleixner:
 "A single fix for an early boot warning caused by invoking
  this_cpu_has() before SMP initialization"

* 'efi-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/mm: Fix bogus warning during EFI bootup, use boot_cpu_has() instead of this_cpu_has() in build_cr3_noflush()

6 years agoMerge branch 'irq-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 15 Apr 2018 19:29:46 +0000 (12:29 -0700)]
Merge branch 'irq-core-for-linus' of git://git./linux/kernel/git/tip/tip

Pull irq affinity fixes from Thomas Gleixner:

  - Fix error path handling in the affinity spreading code

  - Make affinity spreading smarter to avoid issues on systems which
    claim to have hotpluggable CPUs while in fact they can't hotplug
    anything.

    So instead of trying to spread the vectors (and thereby the
    associated device queues) to all possibe CPUs, spread them on all
    present CPUs first. If there are left over vectors after that first
    step they are spread among the possible, but not present CPUs which
    keeps the code backwards compatible for virtual decives and NVME
    which allocate a queue per possible CPU, but makes the spreading
    smarter for devices which have less queues than possible or present
    CPUs.

* 'irq-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  genirq/affinity: Spread irq vectors among present CPUs as far as possible
  genirq/affinity: Allow irq spreading from a given starting point
  genirq/affinity: Move actual irq vector spreading into a helper function
  genirq/affinity: Rename *node_to_possible_cpumask as *node_to_cpumask
  genirq/affinity: Don't return with empty affinity masks on error

6 years agoMerge tag 'for-linus' of git://github.com/openrisc/linux
Linus Torvalds [Sun, 15 Apr 2018 19:27:58 +0000 (12:27 -0700)]
Merge tag 'for-linus' of git://github.com/openrisc/linux

Pull OpenRISC fixlet from Stafford Horne:
 "Just one small thing here, it came in a while back but I didnt have
  anything in my 4.16 queue, still its the only thing for 4.17 so
  sending it alone.

  Small cleanup: remove unused __ARCH_HAVE_MMU define"

* tag 'for-linus' of git://github.com/openrisc/linux:
  openrisc: remove unused __ARCH_HAVE_MMU define

6 years agoMerge tag 'powerpc-4.17-2' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Sun, 15 Apr 2018 18:57:12 +0000 (11:57 -0700)]
Merge tag 'powerpc-4.17-2' of git://git./linux/kernel/git/powerpc/linux

Pull powerpc fixes from Michael Ellerman:

 - Fix crashes when loading modules built with a different
   CONFIG_RELOCATABLE value by adding CONFIG_RELOCATABLE to vermagic.

 - Fix busy loops in the OPAL NVRAM driver if we get certain error
   conditions from firmware.

 - Remove tlbie trace points from KVM code that's called in real mode,
   because it causes crashes.

 - Fix checkstops caused by invalid tlbiel on Power9 Radix.

 - Ensure the set of CPU features we "know" are always enabled is
   actually the minimal set when we build with support for firmware
   supplied CPU features.

Thanks to: Aneesh Kumar K.V, Anshuman Khandual, Nicholas Piggin.

* tag 'powerpc-4.17-2' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc/64s: Fix CPU_FTRS_ALWAYS vs DT CPU features
  powerpc/mm/radix: Fix checkstops caused by invalid tlbiel
  KVM: PPC: Book3S HV: trace_tlbie must not be called in realmode
  powerpc/8xx: Fix build with hugetlbfs enabled
  powerpc/powernv: Fix OPAL NVRAM driver OPAL_BUSY loops
  powerpc/powernv: define a standard delay for OPAL_BUSY type retry loops
  powerpc/fscr: Enable interrupts earlier before calling get_user()
  powerpc/64s: Fix section mismatch warnings from setup_rfi_flush()
  powerpc/modules: Fix crashes by adding CONFIG_RELOCATABLE to vermagic

6 years agoMerge branch 'akpm' (patches from Andrew)
Linus Torvalds [Sat, 14 Apr 2018 15:50:50 +0000 (08:50 -0700)]
Merge branch 'akpm' (patches from Andrew)

Merge yet more updates from Andrew Morton:

 - various hotfixes

 - kexec_file updates and feature work

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (27 commits)
  kernel/kexec_file.c: move purgatories sha256 to common code
  kernel/kexec_file.c: allow archs to set purgatory load address
  kernel/kexec_file.c: remove mis-use of sh_offset field during purgatory load
  kernel/kexec_file.c: remove unneeded variables in kexec_purgatory_setup_sechdrs
  kernel/kexec_file.c: remove unneeded for-loop in kexec_purgatory_setup_sechdrs
  kernel/kexec_file.c: split up __kexec_load_puragory
  kernel/kexec_file.c: use read-only sections in arch_kexec_apply_relocations*
  kernel/kexec_file.c: search symbols in read-only kexec_purgatory
  kernel/kexec_file.c: make purgatory_info->ehdr const
  kernel/kexec_file.c: remove checks in kexec_purgatory_load
  include/linux/kexec.h: silence compile warnings
  kexec_file, x86: move re-factored code to generic side
  x86: kexec_file: clean up prepare_elf64_headers()
  x86: kexec_file: lift CRASH_MAX_RANGES limit on crash_mem buffer
  x86: kexec_file: remove X86_64 dependency from prepare_elf64_headers()
  x86: kexec_file: purge system-ram walking from prepare_elf64_headers()
  kexec_file,x86,powerpc: factor out kexec_file_ops functions
  kexec_file: make use of purgatory optional
  proc: revalidate misc dentries
  mm, slab: reschedule cache_reap() on the same CPU
  ...

6 years agokernel/kexec_file.c: move purgatories sha256 to common code
Philipp Rudo [Fri, 13 Apr 2018 22:36:46 +0000 (15:36 -0700)]
kernel/kexec_file.c: move purgatories sha256 to common code

The code to verify the new kernels sha digest is applicable for all
architectures.  Move it to common code.

One problem is the string.c implementation on x86.  Currently sha256
includes x86/boot/string.h which defines memcpy and memset to be gcc
builtins.  By moving the sha256 implementation to common code and
changing the include to linux/string.h both functions are no longer
defined.  Thus definitions have to be provided in x86/purgatory/string.c

Link: http://lkml.kernel.org/r/20180321112751.22196-12-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokernel/kexec_file.c: allow archs to set purgatory load address
Philipp Rudo [Fri, 13 Apr 2018 22:36:43 +0000 (15:36 -0700)]
kernel/kexec_file.c: allow archs to set purgatory load address

For s390 new kernels are loaded to fixed addresses in memory before they
are booted.  With the current code this is a problem as it assumes the
kernel will be loaded to an 'arbitrary' address.  In particular,
kexec_locate_mem_hole searches for a large enough memory region and sets
the load address (kexec_bufer->mem) to it.

Luckily there is a simple workaround for this problem.  By returning 1
in arch_kexec_walk_mem, kexec_locate_mem_hole is turned off.  This
allows the architecture to set kbuf->mem by hand.  While the trick works
fine for the kernel it does not for the purgatory as here the
architectures don't have access to its kexec_buffer.

Give architectures access to the purgatories kexec_buffer by changing
kexec_load_purgatory to take a pointer to it.  With this change
architectures have access to the buffer and can edit it as they need.

A nice side effect of this change is that we can get rid of the
purgatory_info->purgatory_load_address field.  As now the information
stored there can directly be accessed from kbuf->mem.

Link: http://lkml.kernel.org/r/20180321112751.22196-11-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Reviewed-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokernel/kexec_file.c: remove mis-use of sh_offset field during purgatory load
Philipp Rudo [Fri, 13 Apr 2018 22:36:39 +0000 (15:36 -0700)]
kernel/kexec_file.c: remove mis-use of sh_offset field during purgatory load

The current code uses the sh_offset field in purgatory_info->sechdrs to
store a pointer to the current load address of the section.  Depending
whether the section will be loaded or not this is either a pointer into
purgatory_info->purgatory_buf or kexec_purgatory.  This is not only a
violation of the ELF standard but also makes the code very hard to
understand as you cannot tell if the memory you are using is read-only
or not.

Remove this misuse and store the offset of the section in
pugaroty_info->purgatory_buf in sh_offset.

Link: http://lkml.kernel.org/r/20180321112751.22196-10-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokernel/kexec_file.c: remove unneeded variables in kexec_purgatory_setup_sechdrs
Philipp Rudo [Fri, 13 Apr 2018 22:36:35 +0000 (15:36 -0700)]
kernel/kexec_file.c: remove unneeded variables in kexec_purgatory_setup_sechdrs

The main loop currently uses quite a lot of variables to update the
section headers.  Some of them are unnecessary.  So clean them up a
little.

Link: http://lkml.kernel.org/r/20180321112751.22196-9-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokernel/kexec_file.c: remove unneeded for-loop in kexec_purgatory_setup_sechdrs
Philipp Rudo [Fri, 13 Apr 2018 22:36:32 +0000 (15:36 -0700)]
kernel/kexec_file.c: remove unneeded for-loop in kexec_purgatory_setup_sechdrs

To update the entry point there is an extra loop over all section
headers although this can be done in the main loop.  So move it there
and eliminate the extra loop and variable to store the 'entry section
index'.

Also, in the main loop, move the usual case, i.e.  non-bss section, out
of the extra if-block.

Link: http://lkml.kernel.org/r/20180321112751.22196-8-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Reviewed-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokernel/kexec_file.c: split up __kexec_load_puragory
Philipp Rudo [Fri, 13 Apr 2018 22:36:28 +0000 (15:36 -0700)]
kernel/kexec_file.c: split up __kexec_load_puragory

When inspecting __kexec_load_purgatory you find that it has two tasks

1) setting up the kexec_buffer for the new kernel and,
2) setting up pi->sechdrs for the final load address.

The two tasks are independent of each other.  To improve readability
split up __kexec_load_purgatory into two functions, one for each task,
and call them directly from kexec_load_purgatory.

Link: http://lkml.kernel.org/r/20180321112751.22196-7-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokernel/kexec_file.c: use read-only sections in arch_kexec_apply_relocations*
Philipp Rudo [Fri, 13 Apr 2018 22:36:24 +0000 (15:36 -0700)]
kernel/kexec_file.c: use read-only sections in arch_kexec_apply_relocations*

When the relocations are applied to the purgatory only the section the
relocations are applied to is writable.  The other sections, i.e.  the
symtab and .rel/.rela, are in read-only kexec_purgatory.  Highlight this
by marking the corresponding variables as 'const'.

While at it also change the signatures of arch_kexec_apply_relocations* to
take section pointers instead of just the index of the relocation section.
This removes the second lookup and sanity check of the sections in arch
code.

Link: http://lkml.kernel.org/r/20180321112751.22196-6-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokernel/kexec_file.c: search symbols in read-only kexec_purgatory
Philipp Rudo [Fri, 13 Apr 2018 22:36:21 +0000 (15:36 -0700)]
kernel/kexec_file.c: search symbols in read-only kexec_purgatory

The stripped purgatory does not contain a symtab.  So when looking for
symbols this is done in read-only kexec_purgatory.  Highlight this by
marking the corresponding variables as 'const'.

Link: http://lkml.kernel.org/r/20180321112751.22196-5-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokernel/kexec_file.c: make purgatory_info->ehdr const
Philipp Rudo [Fri, 13 Apr 2018 22:36:17 +0000 (15:36 -0700)]
kernel/kexec_file.c: make purgatory_info->ehdr const

The kexec_purgatory buffer is read-only.  Thus all pointers into
kexec_purgatory are read-only, too.  Point this out by explicitly
marking purgatory_info->ehdr as 'const' and update the comments in
purgatory_info.

Link: http://lkml.kernel.org/r/20180321112751.22196-4-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokernel/kexec_file.c: remove checks in kexec_purgatory_load
Philipp Rudo [Fri, 13 Apr 2018 22:36:13 +0000 (15:36 -0700)]
kernel/kexec_file.c: remove checks in kexec_purgatory_load

Before the purgatory is loaded several checks are done whether the ELF
file in kexec_purgatory is valid or not.  These checks are incomplete.
For example they don't check for the total size of the sections defined
in the section header table or if the entry point actually points into
the purgatory.

On the other hand the purgatory, although an ELF file on its own, is
part of the kernel.  Thus not trusting the purgatory means not trusting
the kernel build itself.

So remove all validity checks on the purgatory and just trust the kernel
build.

Link: http://lkml.kernel.org/r/20180321112751.22196-3-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoinclude/linux/kexec.h: silence compile warnings
Philipp Rudo [Fri, 13 Apr 2018 22:36:10 +0000 (15:36 -0700)]
include/linux/kexec.h: silence compile warnings

Patch series "kexec_file: Clean up purgatory load", v2.

Following the discussion with Dave and AKASHI, here are the common code
patches extracted from my recent patch set (Add kexec_file_load support
to s390) [1].  The patches were extracted to allow upstream integration
together with AKASHI's common code patches before the arch code gets
adjusted to the new base.

The reason for this series is to prepare common code for adding
kexec_file_load to s390 as well as cleaning up the mis-use of the
sh_offset field during purgatory load.  In detail this series contains:

Patch #1&2: Minor cleanups/fixes.

Patch #3-9: Clean up the purgatory load/relocation code.  Especially
remove the mis-use of the purgatory_info->sechdrs->sh_offset field,
currently holding a pointer into either kexec_purgatory (ro) or
purgatory_buf (rw) depending on the section.  With these patches the
section address will be calculated verbosely and sh_offset will contain
the offset of the section in the stripped purgatory binary
(purgatory_buf).

Patch #10: Allows architectures to set the purgatory load address.  This
patch is important for s390 as the kernel and purgatory have to be
loaded to fixed addresses.  In current code this is impossible as the
purgatory load is opaque to the architecture.

Patch #11: Moves x86 purgatories sha implementation to common lib/
directory to allow reuse in other architectures.

This patch (of 11)

When building the kernel with CONFIG_KEXEC_FILE enabled gcc prints a
compile warning multiple times.

  In file included from <path>/linux/init/initramfs.c:526:0:
  <path>/include/linux/kexec.h:120:9: warning: `struct kimage' declared inside parameter list [enabled by default]
           unsigned long cmdline_len);
           ^

This is because the typedefs for kexec_file_load uses struct kimage
before it is declared.  Fix this by simply forward declaring struct
kimage.

Link: http://lkml.kernel.org/r/20180321112751.22196-2-prudo@linux.vnet.ibm.com
Signed-off-by: Philipp Rudo <prudo@linux.vnet.ibm.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: AKASHI Takahiro <takahiro.akashi@linaro.org>
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokexec_file, x86: move re-factored code to generic side
AKASHI Takahiro [Fri, 13 Apr 2018 22:36:06 +0000 (15:36 -0700)]
kexec_file, x86: move re-factored code to generic side

In the previous patches, commonly-used routines, exclude_mem_range() and
prepare_elf64_headers(), were carved out.  Now place them in kexec
common code.  A prefix "crash_" is given to each of their names to avoid
possible name collisions.

Link: http://lkml.kernel.org/r/20180306102303.9063-8-takahiro.akashi@linaro.org
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Acked-by: Dave Young <dyoung@redhat.com>
Tested-by: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Baoquan He <bhe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agox86: kexec_file: clean up prepare_elf64_headers()
AKASHI Takahiro [Fri, 13 Apr 2018 22:36:03 +0000 (15:36 -0700)]
x86: kexec_file: clean up prepare_elf64_headers()

Removing bufp variable in prepare_elf64_headers() makes the code simpler
and more understandable.

Link: http://lkml.kernel.org/r/20180306102303.9063-7-takahiro.akashi@linaro.org
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Acked-by: Dave Young <dyoung@redhat.com>
Tested-by: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Baoquan He <bhe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agox86: kexec_file: lift CRASH_MAX_RANGES limit on crash_mem buffer
AKASHI Takahiro [Fri, 13 Apr 2018 22:35:59 +0000 (15:35 -0700)]
x86: kexec_file: lift CRASH_MAX_RANGES limit on crash_mem buffer

While CRASH_MAX_RANGES (== 16) seems to be good enough, fixed-number
array is not a good idea in general.

In this patch, size of crash_mem buffer is calculated as before and the
buffer is now dynamically allocated.  This change also allows removing
crash_elf_data structure.

Link: http://lkml.kernel.org/r/20180306102303.9063-6-takahiro.akashi@linaro.org
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Acked-by: Dave Young <dyoung@redhat.com>
Tested-by: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Baoquan He <bhe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agox86: kexec_file: remove X86_64 dependency from prepare_elf64_headers()
AKASHI Takahiro [Fri, 13 Apr 2018 22:35:56 +0000 (15:35 -0700)]
x86: kexec_file: remove X86_64 dependency from prepare_elf64_headers()

The code guarded by CONFIG_X86_64 is necessary on some architectures
which have a dedicated kernel mapping outside of linear memory mapping.
(arm64 is among those.)

In this patch, an additional argument, kernel_map, is added to enable/
disable the code removing #ifdef.

Link: http://lkml.kernel.org/r/20180306102303.9063-5-takahiro.akashi@linaro.org
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Acked-by: Dave Young <dyoung@redhat.com>
Tested-by: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Baoquan He <bhe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agox86: kexec_file: purge system-ram walking from prepare_elf64_headers()
AKASHI Takahiro [Fri, 13 Apr 2018 22:35:53 +0000 (15:35 -0700)]
x86: kexec_file: purge system-ram walking from prepare_elf64_headers()

While prepare_elf64_headers() in x86 looks pretty generic for other
architectures' use, it contains some code which tries to list crash
memory regions by walking through system resources, which is not always
architecture agnostic.  To make this function more generic, the related
code should be purged.

In this patch, prepare_elf64_headers() simply scans crash_mem buffer
passed and add all the listed regions to elf header as a PT_LOAD
segment.  So walk_system_ram_res(prepare_elf64_headers_callback) have
been moved forward before prepare_elf64_headers() where the callback,
prepare_elf64_headers_callback(), is now responsible for filling up
crash_mem buffer.

Meanwhile exclude_elf_header_ranges() used to be called every time in
this callback it is rather redundant and now called only once in
prepare_elf_headers() as well.

Link: http://lkml.kernel.org/r/20180306102303.9063-4-takahiro.akashi@linaro.org
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Acked-by: Dave Young <dyoung@redhat.com>
Tested-by: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Baoquan He <bhe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokexec_file,x86,powerpc: factor out kexec_file_ops functions
AKASHI Takahiro [Fri, 13 Apr 2018 22:35:49 +0000 (15:35 -0700)]
kexec_file,x86,powerpc: factor out kexec_file_ops functions

As arch_kexec_kernel_image_{probe,load}(),
arch_kimage_file_post_load_cleanup() and arch_kexec_kernel_verify_sig()
are almost duplicated among architectures, they can be commonalized with
an architecture-defined kexec_file_ops array.  So let's factor them out.

Link: http://lkml.kernel.org/r/20180306102303.9063-3-takahiro.akashi@linaro.org
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Acked-by: Dave Young <dyoung@redhat.com>
Tested-by: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Baoquan He <bhe@redhat.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Thiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokexec_file: make use of purgatory optional
AKASHI Takahiro [Fri, 13 Apr 2018 22:35:45 +0000 (15:35 -0700)]
kexec_file: make use of purgatory optional

Patch series "kexec_file, x86, powerpc: refactoring for other
architecutres", v2.

This is a preparatory patchset for adding kexec_file support on arm64.

It was originally included in a arm64 patch set[1], but Philipp is also
working on their kexec_file support on s390[2] and some changes are now
conflicting.

So these common parts were extracted and put into a separate patch set
for better integration.  What's more, my original patch#4 was split into
a few small chunks for easier review after Dave's comment.

As such, the resulting code is basically identical with my original, and
the only *visible* differences are:

 - renaming of _kexec_kernel_image_probe() and  _kimage_file_post_load_cleanup()

 - change one of types of arguments at prepare_elf64_headers()

Those, unfortunately, require a couple of trivial changes on the rest
(#1, #6 to #13) of my arm64 kexec_file patch set[1].

Patch #1 allows making a use of purgatory optional, particularly useful
for arm64.

Patch #2 commonalizes arch_kexec_kernel_{image_probe, image_load,
verify_sig}() and arch_kimage_file_post_load_cleanup() across
architectures.

Patches #3-#7 are also intended to generalize parse_elf64_headers(),
along with exclude_mem_range(), to be made best re-use of.

[1] http://lists.infradead.org/pipermail/linux-arm-kernel/2018-February/561182.html
[2] http://lkml.iu.edu//hypermail/linux/kernel/1802.1/02596.html

This patch (of 7):

On arm64, crash dump kernel's usable memory is protected by *unmapping*
it from kernel virtual space unlike other architectures where the region
is just made read-only.  It is highly unlikely that the region is
accidentally corrupted and this observation rationalizes that digest
check code can also be dropped from purgatory.  The resulting code is so
simple as it doesn't require a bit ugly re-linking/relocation stuff,
i.e.  arch_kexec_apply_relocations_add().

Please see:

   http://lists.infradead.org/pipermail/linux-arm-kernel/2017-December/545428.html

All that the purgatory does is to shuffle arguments and jump into a new
kernel, while we still need to have some space for a hash value
(purgatory_sha256_digest) which is never checked against.

As such, it doesn't make sense to have trampline code between old kernel
and new kernel on arm64.

This patch introduces a new configuration, ARCH_HAS_KEXEC_PURGATORY, and
allows related code to be compiled in only if necessary.

[takahiro.akashi@linaro.org: fix trivial screwup]
Link: http://lkml.kernel.org/r/20180309093346.GF25863@linaro.org
Link: http://lkml.kernel.org/r/20180306102303.9063-2-takahiro.akashi@linaro.org
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Acked-by: Dave Young <dyoung@redhat.com>
Tested-by: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Baoquan He <bhe@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: revalidate misc dentries
Alexey Dobriyan [Fri, 13 Apr 2018 22:35:42 +0000 (15:35 -0700)]
proc: revalidate misc dentries

If module removes proc directory while another process pins it by
chdir'ing to it, then subsequent recreation of proc entry and all
entries down the tree will not be visible to any process until pinning
process unchdir from directory and unpins everything.

Steps to reproduce:

proc_mkdir("aaa", NULL);
proc_create("aaa/bbb", ...);

chdir("/proc/aaa");

remove_proc_entry("aaa/bbb", NULL);
remove_proc_entry("aaa", NULL);

proc_mkdir("aaa", NULL);
# inaccessible because "aaa" dentry still points
# to the original "aaa".
proc_create("aaa/bbb", ...);

Fix is to implement ->d_revalidate and ->d_delete.

Link: http://lkml.kernel.org/r/20180312201938.GA4871@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm, slab: reschedule cache_reap() on the same CPU
Vlastimil Babka [Fri, 13 Apr 2018 22:35:38 +0000 (15:35 -0700)]
mm, slab: reschedule cache_reap() on the same CPU

cache_reap() is initially scheduled in start_cpu_timer() via
schedule_delayed_work_on(). But then the next iterations are scheduled
via schedule_delayed_work(), i.e. using WORK_CPU_UNBOUND.

Thus since commit ef557180447f ("workqueue: schedule WORK_CPU_UNBOUND
work on wq_unbound_cpumask CPUs") there is no guarantee the future
iterations will run on the originally intended cpu, although it's still
preferred.  I was able to demonstrate this with
/sys/module/workqueue/parameters/debug_force_rr_cpu.  IIUC, it may also
happen due to migrating timers in nohz context.  As a result, some cpu's
would be calling cache_reap() more frequently and others never.

This patch uses schedule_delayed_work_on() with the current cpu when
scheduling the next iteration.

Link: http://lkml.kernel.org/r/20180411070007.32225-1-vbabka@suse.cz
Fixes: ef557180447f ("workqueue: schedule WORK_CPU_UNBOUND work on wq_unbound_cpumask CPUs")
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Pekka Enberg <penberg@kernel.org>
Acked-by: Christoph Lameter <cl@linux.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Stephen Boyd <sboyd@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokexec: export PG_swapbacked to VMCOREINFO
Petr Tesarik [Fri, 13 Apr 2018 22:35:34 +0000 (15:35 -0700)]
kexec: export PG_swapbacked to VMCOREINFO

Since commit 6326fec1122c ("mm: Use owner_priv bit for PageSwapCache,
valid when PageSwapBacked"), PG_swapcache is an alias for
PG_owner_priv_1, which may be also used for other purposes.

To know whether the bit indeed has the PG_swapcache meaning, it is
necessary to check PG_swapbacked, hence this bit must be exported.

Link: http://lkml.kernel.org/r/20180410161345.142e142d@ezekiel.suse.cz
Signed-off-by: Petr Tesarik <ptesarik@suse.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Dave Young <dyoung@redhat.com>
Cc: Xunlei Pang <xlpang@redhat.com>
Cc: Baoquan He <bhe@redhat.com>
Cc: Hari Bathini <hbathini@linux.vnet.ibm.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: "Marc-Andr Lureau" <marcandre.lureau@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoipc/shm: fix use-after-free of shm file via remap_file_pages()
Eric Biggers [Fri, 13 Apr 2018 22:35:30 +0000 (15:35 -0700)]
ipc/shm: fix use-after-free of shm file via remap_file_pages()

syzbot reported a use-after-free of shm_file_data(file)->file->f_op in
shm_get_unmapped_area(), called via sys_remap_file_pages().

Unfortunately it couldn't generate a reproducer, but I found a bug which
I think caused it.  When remap_file_pages() is passed a full System V
shared memory segment, the memory is first unmapped, then a new map is
created using the ->vm_file.  Between these steps, the shm ID can be
removed and reused for a new shm segment.  But, shm_mmap() only checks
whether the ID is currently valid before calling the underlying file's
->mmap(); it doesn't check whether it was reused.  Thus it can use the
wrong underlying file, one that was already freed.

Fix this by making the "outer" shm file (the one that gets put in
->vm_file) hold a reference to the real shm file, and by making
__shm_open() require that the file associated with the shm ID matches
the one associated with the "outer" file.

Taking the reference to the real shm file is needed to fully solve the
problem, since otherwise sfd->file could point to a freed file, which
then could be reallocated for the reused shm ID, causing the wrong shm
segment to be mapped (and without the required permission checks).

Commit 1ac0b6dec656 ("ipc/shm: handle removed segments gracefully in
shm_mmap()") almost fixed this bug, but it didn't go far enough because
it didn't consider the case where the shm ID is reused.

The following program usually reproduces this bug:

#include <stdlib.h>
#include <sys/shm.h>
#include <sys/syscall.h>
#include <unistd.h>

int main()
{
int is_parent = (fork() != 0);
srand(getpid());
for (;;) {
int id = shmget(0xF00F, 4096, IPC_CREAT|0700);
if (is_parent) {
void *addr = shmat(id, NULL, 0);
usleep(rand() % 50);
while (!syscall(__NR_remap_file_pages, addr, 4096, 0, 0, 0));
} else {
usleep(rand() % 50);
shmctl(id, IPC_RMID, NULL);
}
}
}

It causes the following NULL pointer dereference due to a 'struct file'
being used while it's being freed.  (I couldn't actually get a KASAN
use-after-free splat like in the syzbot report.  But I think it's
possible with this bug; it would just take a more extraordinary race...)

BUG: unable to handle kernel NULL pointer dereference at 0000000000000058
PGD 0 P4D 0
Oops: 0000 [#1] SMP NOPTI
CPU: 9 PID: 258 Comm: syz_ipc Not tainted 4.16.0-05140-gf8cf2f16a7c95 #189
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.0-20171110_100015-anatol 04/01/2014
RIP: 0010:d_inode include/linux/dcache.h:519 [inline]
RIP: 0010:touch_atime+0x25/0xd0 fs/inode.c:1724
[...]
Call Trace:
 file_accessed include/linux/fs.h:2063 [inline]
 shmem_mmap+0x25/0x40 mm/shmem.c:2149
 call_mmap include/linux/fs.h:1789 [inline]
 shm_mmap+0x34/0x80 ipc/shm.c:465
 call_mmap include/linux/fs.h:1789 [inline]
 mmap_region+0x309/0x5b0 mm/mmap.c:1712
 do_mmap+0x294/0x4a0 mm/mmap.c:1483
 do_mmap_pgoff include/linux/mm.h:2235 [inline]
 SYSC_remap_file_pages mm/mmap.c:2853 [inline]
 SyS_remap_file_pages+0x232/0x310 mm/mmap.c:2769
 do_syscall_64+0x64/0x1a0 arch/x86/entry/common.c:287
 entry_SYSCALL_64_after_hwframe+0x42/0xb7

[ebiggers@google.com: add comment]
Link: http://lkml.kernel.org/r/20180410192850.235835-1-ebiggers3@gmail.com
Link: http://lkml.kernel.org/r/20180409043039.28915-1-ebiggers3@gmail.com
Reported-by: syzbot+d11f321e7f1923157eac80aa990b446596f46439@syzkaller.appspotmail.com
Fixes: c8d78c1823f4 ("mm: replace remap_file_pages() syscall with emulation")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Manfred Spraul <manfred@colorfullife.com>
Cc: "Eric W . Biederman" <ebiederm@xmission.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/filemap.c: provide dummy filemap_page_mkwrite() for NOMMU
Arnd Bergmann [Fri, 13 Apr 2018 22:35:27 +0000 (15:35 -0700)]
mm/filemap.c: provide dummy filemap_page_mkwrite() for NOMMU

Building orangefs on MMU-less machines now results in a link error
because of the newly introduced use of the filemap_page_mkwrite()
function:

  ERROR: "filemap_page_mkwrite" [fs/orangefs/orangefs.ko] undefined!

This adds a dummy version for it, similar to the existing
generic_file_mmap and generic_file_readonly_mmap stubs in the same file,
to avoid the link error without adding #ifdefs in each file system that
uses these.

Link: http://lkml.kernel.org/r/20180409105555.2439976-1-arnd@arndb.de
Fixes: a5135eeab2e5 ("orangefs: implement vm_ops->fault")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Martin Brandenburg <martin@omnibond.com>
Cc: Mike Marshall <hubcap@omnibond.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/gup.c: document return value
Michael S. Tsirkin [Fri, 13 Apr 2018 22:35:23 +0000 (15:35 -0700)]
mm/gup.c: document return value

__get_user_pages_fast handles errors differently from
get_user_pages_fast: the former always returns the number of pages
pinned, the later might return a negative error code.

Link: http://lkml.kernel.org/r/1522962072-182137-6-git-send-email-mst@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thorsten Leemhuis <regressions@leemhuis.info>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoget_user_pages_fast(): return -EFAULT on access_ok failure
Michael S. Tsirkin [Fri, 13 Apr 2018 22:35:20 +0000 (15:35 -0700)]
get_user_pages_fast(): return -EFAULT on access_ok failure

get_user_pages_fast is supposed to be a faster drop-in equivalent of
get_user_pages.  As such, callers expect it to return a negative return
code when passed an invalid address, and never expect it to return 0
when passed a positive number of pages, since its documentation says:

 * Returns number of pages pinned. This may be fewer than the number
 * requested. If nr_pages is 0 or negative, returns 0. If no pages
 * were pinned, returns -errno.

When get_user_pages_fast fall back on get_user_pages this is exactly
what happens.  Unfortunately the implementation is inconsistent: it
returns 0 if passed a kernel address, confusing callers: for example,
the following is pretty common but does not appear to do the right thing
with a kernel address:

        ret = get_user_pages_fast(addr, 1, writeable, &page);
        if (ret < 0)
                return ret;

Change get_user_pages_fast to return -EFAULT when supplied a kernel
address to make it match expectations.

All callers have been audited for consistency with the documented
semantics.

Link: http://lkml.kernel.org/r/1522962072-182137-4-git-send-email-mst@redhat.com
Fixes: 5b65c4677a57 ("mm, x86/mm: Fix performance regression in get_user_pages_fast()")
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reported-by: syzbot+6304bf97ef436580fede@syzkaller.appspotmail.com
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thorsten Leemhuis <regressions@leemhuis.info>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/gup_benchmark: handle gup failures
Michael S. Tsirkin [Fri, 13 Apr 2018 22:35:16 +0000 (15:35 -0700)]
mm/gup_benchmark: handle gup failures

Patch series "mm/get_user_pages_fast fixes, cleanups", v2.

Turns out get_user_pages_fast and __get_user_pages_fast return different
values on error when given a single page: __get_user_pages_fast returns
0.  get_user_pages_fast returns either 0 or an error.

Callers of get_user_pages_fast expect an error so fix it up to return an
error consistently.

Stress the difference between get_user_pages_fast and
__get_user_pages_fast to make sure callers aren't confused.

This patch (of 3):

__gup_benchmark_ioctl does not handle the case where get_user_pages_fast
fails:

 - a negative return code will cause a buffer overrun

 - returning with partial success will cause use of uninitialized
   memory.

[akpm@linux-foundation.org: simplification]
Link: http://lkml.kernel.org/r/1522962072-182137-3-git-send-email-mst@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thorsten Leemhuis <regressions@leemhuis.info>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoresource: fix integer overflow at reallocation
Takashi Iwai [Fri, 13 Apr 2018 22:35:13 +0000 (15:35 -0700)]
resource: fix integer overflow at reallocation

We've got a bug report indicating a kernel panic at booting on an x86-32
system, and it turned out to be the invalid PCI resource assigned after
reallocation.  __find_resource() first aligns the resource start address
and resets the end address with start+size-1 accordingly, then checks
whether it's contained.  Here the end address may overflow the integer,
although resource_contains() still returns true because the function
validates only start and end address.  So this ends up with returning an
invalid resource (start > end).

There was already an attempt to cover such a problem in the commit
47ea91b4052d ("Resource: fix wrong resource window calculation"), but
this case is an overseen one.

This patch adds the validity check of the newly calculated resource for
avoiding the integer overflow problem.

Bugzilla: http://bugzilla.opensuse.org/show_bug.cgi?id=1086739
Link: http://lkml.kernel.org/r/s5hpo37d5l8.wl-tiwai@suse.de
Fixes: 23c570a67448 ("resource: ability to resize an allocated resource")
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Reported-by: Michael Henders <hendersm@shaw.ca>
Tested-by: Michael Henders <hendersm@shaw.ca>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Ram Pai <linuxram@us.ibm.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoMerge branch 'overlayfs-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszer...
Linus Torvalds [Fri, 13 Apr 2018 23:55:41 +0000 (16:55 -0700)]
Merge branch 'overlayfs-linus' of git://git./linux/kernel/git/mszeredi/vfs

Pull overlayfs updates from Miklos Szeredi:
 "In addition to bug fixes and cleanups there are two new features from
  Amir:

   - Consistent inode number support for the case when layers are not
     all on the same filesystem (feature is dubbed "xino").

   - Optimize overlayfs file handle decoding. This one touches the
     exportfs interface to allow detecting the disconnected directory
     case"

* 'overlayfs-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs:
  ovl: update documentation w.r.t "xino" feature
  ovl: add support for "xino" mount and config options
  ovl: consistent d_ino for non-samefs with xino
  ovl: consistent i_ino for non-samefs with xino
  ovl: constant st_ino for non-samefs with xino
  ovl: allocate anon bdev per unique lower fs
  ovl: factor out ovl_map_dev_ino() helper
  ovl: cleanup ovl_update_time()
  ovl: add WARN_ON() for non-dir redirect cases
  ovl: cleanup setting OVL_INDEX
  ovl: set d->is_dir and d->opaque for last path element
  ovl: Do not check for redirect if this is last layer
  ovl: lookup in inode cache first when decoding lower file handle
  ovl: do not try to reconnect a disconnected origin dentry
  ovl: disambiguate ovl_encode_fh()
  ovl: set lower layer st_dev only if setting lower st_ino
  ovl: fix lookup with middle layer opaque dir and absolute path redirects
  ovl: Set d->last properly during lookup
  ovl: set i_ino to the value of st_ino for NFS export

6 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/rzhang/linux
Linus Torvalds [Fri, 13 Apr 2018 23:52:26 +0000 (16:52 -0700)]
Merge branch 'next' of git://git./linux/kernel/git/rzhang/linux

Pull thermal management update from Zhang Rui:

 - Fix race condition in imx_thermal_probe() (Mikhail Lappo)

 - Add cooling device's statistics in sysfs (Viresh Kumar)

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/rzhang/linux:
  thermal: Add cooling device's statistics in sysfs
  thermal: imx: Fix race condition in imx_thermal_probe()