staging:rtl8192u: Rename RxLastFragNum - Style

Rename the member variable RxLastFragNum to rx_last_frag_num. This
change clears the checkpatch issue with CamelCase naming.

The resulting edits are all coding style changes which should have
no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename RxLastSeqNum - Style

Rename the member variable RxLastSeqNum to rx_last_seq_num. This change
clears the checkpatch issue with CamelCase naming.

The resulting changes are purely coding style in nature so should have
no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename RxAdmittedBARecord - Style

Rename the member variable RxAdmittedBARecord to rx_admitted_ba_record.
This change clears the checkpatch issue with CamelCase naming.

The resulting changes are coding style in nature and as such should have
no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename RxPktPendingTimer - Style

Rename the member variable RxPktPendingTimer to rx_pkt_pending_timer.
This change clears the checkpatch issue with CamelCase naming.

The resulting changes are coding style changes in nature so should
have no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename RxPendingPktList - Style

Rename the member variable RxPendingPktList to rx_pending_pkt_list.
This change resolves the checkpatch issue with CamelCase naming.

The resulting changes are purely coding style changes which should have
no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename RxTimeoutIndicateSeq _Style

Rename member variable RxTimeoutIndicateSeq to rx_timeout_indicate_seq.
This change clears the checkpatch issue with CamelCase naming.

The resulting changes are purely coding style in nature and should have
no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename RxIndicateSeq - Style

Rename the member variable RxIndicateSeq to rx_indicate_seq. This change
clears the checkpatch issue with CamelCase naming.

The resulting changes are purely coding style changes which should have
no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove unused timer values - Style

Remove the unused timer constants TS_SETUP_TIMEOUT and TS_INACT_TIMEOUT
as neither are used in code.

This is a style change which should have no impact on runtime code
execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove commented out code - Style

Remove commented out sections of code, specifically for a structure,
QOS_DL_RECORD, which no longer exists in the module.

This change is a coding style change which should have no impact on
runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove typedef and rename RX_TS_RECORD - Style

Remove the typedef from structure RX_TS_RECORD, this change clears the
checkpatch issue with creation of new types.

Additionally the structure is renamed from RX_TS_RECORD to rx_ts_record.
Whilst this is not raised as a checkpatch error structures are meant to
be named in lower case.

These changes are coding style changes which should have no impact on
runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename TsAddBaTimer - Style

Rename the member variable TsAddBaTimer to ts_add_ba_timer. This change
clears the checkpatch issue with CamelCase naming.

The resulting changes are coding style in nature and as such should not
have any impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename bUsingBa - Style

Rename the member variable bUsingBa to using_ba. This change clears the
checkpatch issue with CamelCase naming.

The resulting changes are purely coding style and should not impact
runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename bAddBaReqDelayed - Style

Rename the member variable bAddBaReqDelayed to add_ba_req_delayed
This change clears the checkpatch issue with CamelCase naming.

The resulting changes are purely coding style changes which should
have no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename bAddBaReqInProgress - Style

Rename the member variable bAddBaReqInProgress to add_ba_req_in_progress
This change clears the checkpatch issue with CamelCase naming.

Changes are purely coding style and should have no impact on runtime
code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename TxAdmittedBARecord - Style

Rename the member variable TxAdmittedBARecord to tx_admitted_ba_record
This change clears the checkpatch issue with CamelCase naming.

The resulting changes are purely coding style changes which should not
impact runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename TxPendingBARecord - Style

Rename the member variable TxPendingBARecord to tx_pending_ba_record.
This change clears the checkpatch issue with CamelCase naming.

The changes are purely coding style changes and should not impact
runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename TxCurSeq - Style

Rename the member variable TxCurSeq to tx_cur_seq. This change clears
the checkpatch issue with CamelCase naming.

The changes are coding style changes which should not impact runtime
code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename TsCommonInfo - Style

Rename the member variable TsCommonInfo in two structures, both
tx_ts_record and RX_TS_RECORD. This member variable is used in both
structures and in both cases causes a checkpatch issue with CamelCase
naming.

The changes are purely coding style and should not impact runtime
code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove typedef and rename TX_TS_RECORD - Style

Remove the typdef from structure TX_TS_RECORD and rename to tx_ts_record.
The removal of the typedef clears the checkpatch issue with creating new
types in code. The name change, whilst not specifically flagged by
checkpatch, is an issue since types are meant to be named in lowercase.

These changes are purely coding style changes and should have no impact
on runtime execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename eRFPath - Style

Rename the variable eRFPath to e_rfpath. This change resolves a
checkpatch issue with CamelCase naming. The variable name is used both
as a parameter name in function definitions, function prototypes and as
a local variable.

This change is purely a coding style change which should have no impact
on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove repeated definitions - Style

Remove bit field masks which are defined in two files: r819xU_phy.h
and r819xU_phyreg.h. Or in the case of bMaskByte3 defined but never used.

The definitions have been removed from the file r819xU_phy.h.

This is a coding style change, which should have no impact on runtime
code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename RF90_RADIO_PATH_E - Style

Rename the enumerated type RF90_RADIO_PATH_E to rf90_radio_path_e.
Whilst it is not flagged as an issue by checkpatch, types are meant
to be named in lowercase.

This change is purely a coding style change which should have no
impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename HW90_BLOCK_E - Style

Rename the enumerated type HW90_BLOCK_E to hw90_block_e. Whilst this is
not flagged by checkpatch types are meant to be in lowercase.

This is a coding style change which should have no impact on runtime
code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove proxy rtl819XRadioD_Array - Style

Remove the array rtl819XRadioD_Array which is only acting as a proxy to
the real array Rtl8192UsbRadioD_Array.

This is a coding style change which should have no impact on runtime
code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove proxy rtl819XRadioC_Array - Style

Remove the array rtl819XRadioC_Array which is only serving as a proxy for
the real array Rtl8192UsbRadioC_Array. This is a coding style change and
should have no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove proxy rtl819XRadioB_Array - Style

Remove the array rtl819XRadioB_Array Rtl8192UsbRadioB_Array which acts
as a proxy the actual array Rtl8192UsbRadioB_Array. The only purpose, I
can see for this implementation is to obscure the actual array being
used Rtl8192UsbRadioB_Array.

This is a coding style change which should have no impact on runtime
code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove proxy rtl819XRadioA_Array - Style

The array rtl819XRadioA_Array serves as a proxy array for the actual
array Rtl8192UsbRadioA_Array. This implementation seems to serve no
other purpose then to obscure the actually array Rtl8192UsbRadioA_Array.

This is a coding style change which should have no impact on runtime
code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove proxy array rtl819XAGCTAB_Array - Style

Remove the array rtl819XAGCTAB_Array which is only a proxy to the
array Rtl8192UsbAGCTAB_Array. There appears to be no purpose
served by this implementation, other then to obscure the
Rtl8192UsbAGCTAB_Array structure, which is actually written to and
read from.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove proxy struct rtl819XPHY_REG_1T2RArray - Style

Remove the struct rtl819XPHY_REG_1T2RArray which is simply a proxy for
the struct Rtl8192UsbPHY_REG_1T2RArray. There appears to be no purpose
served by this implementation, other then to obscure the
Rtl8192UsbPHY_REG_1T2RArray structure.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove proxy struct rtl819XMACPHY_Array_PG - Style

Remove the struct rtl819XMACPHY_Array_PG which is simply a proxy for
the struct Rtl8192UsbMACPHY_Array_PG. There appears to be no purpose
served by this implementation, other then to obscure the
Rtl8192UsbMACPHY_Array_PG structure.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename msDelay to ms_delay - Style

Rename the member variable msDelay to ms_delay. This change clears the
checkpatch issue with CamelCase naming. The change is purely a coding
style change which should not impact runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename Para2 to para_2 - style

Rename member variable Para2 to para_2. This change clears the checkpatch
issue with CamelCase naming. The change is purely a coding style change
and should have no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename Para1 > para_1 - Style

Rename the member variable Para1 to para_1. This change clears the
checkpatch issue with CamelCase naming. The change is purely a coding
style change and should have no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename CmdID - Style

Rename the member variable CmdId to cmd_id. This change clears the
checkpatch issue with CamelCase naming. The change is purely a coding
style change which should have no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename SwChnlCmd - Style

Rename the structure SwChnlCmd to sw_chnl_cmd. This change clears the
checkpatch issue with CamelCase naming. The change is purely a coding
style change and should have no impact on runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Remove multiple blank lines - Style

Remove multiple blank lines, raise a checkpatch issue. This is purely
a coding style change which should not impact runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename ISR_BcnTimerIntr - Style

Rename the bit definition ISR_BcnTimerIntr to ISR_BCN_TIMER_INTR. This
change clears the checkpatch issue with CamelCase naming. The change is
purely a style change and should not impact runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename ISR_TxBcnErr bit definition - Style

Rename the bit definition ISR_TxBcnErr to ISR_TX_BCN_ERR. This change
clears the checkpatch issue with CamelCase naming. The change is purely
a coding style change which should not impact runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:rtl8192u: Rename bit definition ISR_TxBcnOk - Style

Rename the bit definition ISR_TxBcnOk to ISR_TX_BCN_OK. This change
clears the checkpatch issue with CamelCase naming. The change is a
coding style change which should not impact runtime code execution.

Signed-off-by: John Whitmore <johnfwhitmore@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: remove unused rtw_get_oper_choffset()

The function rtw_get_oper_choffset() is never used, so remove it.
Discovered by cppcheck.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: remove unused rtw_get_oper_bw()

The function rtw_get_oper_bw() is never used, so remove it.
Discovered by cppcheck.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: remove unused CAM_empty_entry()

The function CAM_empty_entry() is never used, so remove it.
Discovered by cppcheck.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: remove unused ODM_RASupport_Init()

The function ODM_RASupport_Init() is never used, so remove it.
Discovered by cppcheck.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wlan-ng: correction to comment in hfa384x_usb

The comment for hfa384x_docmd incorrectly states that usercb_data
should be NULL for DOASYNC calls; in fact, it should be NULL for
DOWAIT calls (this is consistent with the other similar functions and
the rest of the comment text).

Signed-off-by: Tim Collier <osdevtc@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wlan-ng: remove volatile from reapable field in hfa384x_usbctlx

Fix checkpatch.pl warning:

"Use of volatile is usually wrong: see
Documentation/process/volatile-considered-harmful.rs"

reapable is not used to access h/w directly, so volatile is not
necessary.

>From reading the code, the contexts in which the field may be
updated/accessed across threads are protected by the hw->ctlxq.lock
spinlock, where hw is the device struct, so appears thread-safe.

Signed-off-by: Tim Collier <osdevtc@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8188eu: remove unused sreset_get_wifi_status()

The function sreset_get_wifi_status() is never used, so remove it.
Discovered by cppcheck.

Signed-off-by: Michael Straube <straube.linux@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: mt7621-mmc: Use BIT macro instead of explicit shifting in board.h

Replace explicit shifting with BIT macro in board.h.
Issue found by checkpatch.

Signed-off-by: Nishad Kamdar <nishadkamdar@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8192u: ieee80211: Convert from ahash to shash

This is an identical change to the wireless/lib80211 of the same name.
In preparing to remove all stack VLA usage from the kernel[1], this
removes the discouraged use of AHASH_REQUEST_ON_STACK in favor of
the smaller SHASH_DESC_ON_STACK by converting from ahash-wrapped-shash
to direct shash. By removing a layer of indirection this both improves
performance and reduces stack usage. The stack allocation will be made
a fixed size in a later patch to the crypto subsystem.

[1] https://lkml.kernel.org/r/CA+55aFzCG-zNmZwX4A2FQpadafLfEzK6CC=qPXydAacU1RqZWA@mail.gmail.com

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8192e: ieee80211: Convert from ahash to shash

This is an identical change to the wireless/lib80211 of the same name.
In preparing to remove all stack VLA usage from the kernel[1], this
removes the discouraged use of AHASH_REQUEST_ON_STACK in favor of
the smaller SHASH_DESC_ON_STACK by converting from ahash-wrapped-shash
to direct shash. By removing a layer of indirection this both improves
performance and reduces stack usage. The stack allocation will be made
a fixed size in a later patch to the crypto subsystem.

[1] https://lkml.kernel.org/r/CA+55aFzCG-zNmZwX4A2FQpadafLfEzK6CC=qPXydAacU1RqZWA@mail.gmail.com

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: use NULL instead of 0 for null pointer

Fixes sparse warning: Using plain integer as NULL pointer

Signed-off-by: Dmitriy Cherkasov <dmitriy@oss-tech.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: core: hold reference on device while in use

Hold a reference on the struct device while a pointer to that
device is in use by gasket.

Reported-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: fix conditional uninitialized `pcn' in z_erofs_map_blocks_iter

This patch adds error handling code for
z_erofs_map_blocks_iter to fix the compiler blame.

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: fix compile error without built-in decompression support

This patch fixes incorrect code snippets due to spilt code
into small patches by mistake.

Link: https://lists.01.org/pipermail/kbuild-all/2018-July/050747.html
Link: https://lists.01.org/pipermail/kbuild-all/2018-July/050750.html
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: axis-fifo: fix return value check in axis_fifo_probe()

In case of error, the function device_create() returns ERR_PTR() and
never returns NULL. The NULL test in the return value check should be
replaced with IS_ERR().

Fixes: 4a965c5f89de ("staging: add driver for Xilinx AXI-Stream FIFO v4.1 IP core")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: page table: remove code for "no dma_ops"

Remove code with TODOs on it for working around apparent problems
previously seen in a qemu environment where dma_ops was not set
correctly.  There is no user of this in the current code.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: sysfs: remove unnecessary NULL check on device ptr

The device pointer passed into get_mapping() will never be NULL; the
check is unnecessary.

Reported-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: apex: fixup undefined PCI class

Apex chips with class 0 (PCI_CLASS_NOT_DEFINED) fixed up to
PCI_CLASS_SYSTEM_OTHER to enable PCI resource assignments.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: sysfs: remove check for refcount already zero

Remove the check for refcount already zero, which shouldn't be
necessary.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: fix a compile warning of Z_EROFS_VLE_VMAP_ONSTACK_PAGES

There is a type mismatch in the definition of
Z_EROFS_VLE_VMAP_ONSTACK_PAGES, let's fix it.

Link: https://lists.01.org/pipermail/kbuild-all/2018-July/050707.html
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add a TODO and update MAINTAINERS for staging

This patch adds a TODO to list the things to be done, and
the relevant info to MAINTAINERS so we can take all the blame :)

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: introduce cached decompression

This patch adds an optional choice which can be
enabled by users in order to cache both incomplete
ends of compressed clusters as a complement to
the in-place decompression in order to boost random
read, but it costs more memory than the in-place
decompression only.

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: introduce VLE decompression support

This patch introduces the basic in-place VLE decompression
implementation for the erofs file system.

Compared with fixed-sized input compression, it implements
what we call 'the variable-length extent compression' which
specifies the same output size for each compression block
to make the full use of IO bandwidth (which means almost
all data from block device can be directly used for decomp-
ression), improve the real (rather than just via data caching,
which costs more memory) random read and keep the relatively
lower compression ratios (it saves more storage space than
fixed-sized input compression which is also configured with
the same input block size), as illustrated below:

        |---  variable-length extent ---|------ VLE ------|---  VLE ---|
         /> clusterofs                  /> clusterofs     /> clusterofs /> clusterofs
   ++---|-------++-----------++---------|-++-----------++-|---------++-|
...||   |       ||           ||         | ||           || |         || | ... original data
   ++---|-------++-----------++---------|-++-----------++-|---------++-|
   ++->cluster<-++->cluster<-++->cluster<-++->cluster<-++->cluster<-++
        size         size         size         size         size
         \                             /                 /            /
          \                      /              /            /
           \               /            /            /
            ++-----------++-----------++-----------++
        ... ||           ||           ||           || ... compressed clusters
            ++-----------++-----------++-----------++
            ++->cluster<-++->cluster<-++->cluster<-++
                 size         size         size

The main point of 'in-place' refers to the decompression mode:
Instead of allocating independent compressed pages and data
structures, it reuses the allocated file cache pages at most
to store its compressed data and the corresponding pagevec in
a time-sharing approach by default, which will be useful for
low memory scenario.

In the end, unlike the other filesystems with (de)compression
support using a relatively large compression block size, which
reads and decompresses >= 128KB at once, and gains a more
good-looking random read (In fact it collects small random reads
into large sequential reads and caches all decompressed data
in memory, but it is unacceptable especially for embedded devices
with limited memory, and it is not the real random read), we
select a universal small-sized 4KB compressed cluster, which is
the smallest page size for most architectures, and all compressed
clusters can be read and decompressed independently, which ensures
random read number for all use cases.

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: introduce workstation for decompression

This patch introduces another concept used by the unzip
subsystem called 'workstation'. It can be seen as a sparse
array that stores pointers pointed to data structures
related to the corresponding physical blocks.

All lookup cases are protected by RCU read lock. Besides,
reference count and spin_lock are also introduced to
manage its lifetime and serialize all update operations.

'workstation' is currently implemented on the in-kernel
radix tree approach for backward compatibility.
With the evolution of linux kernel, it could be migrated
into XArray implementation in the future.

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: introduce erofs shrinker

This patch adds a dedicated shrinker targeting to free unneeded
memory consumed by a number of erofs in-memory data structures.

Like F2FS and UBIFS, it also adds:
  - sbi->umount_mutex to avoid races on shrinker and put_super
  - sbi->shrinker_run_no to not revisit recently scaned objects

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: introduce superblock registration

In order to introducing shrinker solution for erofs,
let's manage all mounted erofs instances at first.

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add a generic z_erofs VLE decompressor

Currently, this patch only simply implements LZ4
decompressor due to its development priority.

In the future, erofs will support more compression
algorithm and format other than LZ4, thus a generic
decompressor interface will be needed.

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: introduce a customized LZ4 decompression

We have to reduce the memory cost as much as possible,
so we don't want to decompress more data beyond
the output buffer size, however "LZ4_decompress_safe_partial"
doesn't guarantee to stop at the arbitary end position,
but stop just after its current LZ4 "sequence" is completed.

Link: https://groups.google.com/forum/#!topic/lz4c/_3kkz5N6n00
Therefore, I hacked the LZ4 decompression logic by hand,
probably NOT the fastest approach, and hope for better
implementation.

Signed-off-by: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: globalize prepare_bio and __submit_bio

The unzip subsystem also uses these functions,
let's export them to internal.h.

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add erofs_allocpage

This patch introduces an temporary _on-stack_ page
pool to reuse the freed page directly as much as
it can for better performance and release all pages
at a time, it also slightly reduces the possibility of
the potential memory allocation failure.

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add erofs_map_blocks_iter

This patch introduces an iterable L2P mapping
operation 'erofs_map_blocks_iter'.
Compared with 'erofs_map_blocks', it avoids
a number of redundant 'release and regrab'
processes if they request the same meta page.

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: introduce pagevec for unzip subsystem

For each compressed cluster, there is a straight-forward
way of allocating a fixed or variable-sized (for VLE) array
to record the corresponding file pages for its decompression
if we decide to decompress these pages asynchronously (eg.
read-ahead case), however it could take much extra on-heap
memory compared with traditional uncompressed filesystems.

This patch introduces a pagevec solution to reuse some
allocated file page in the time-sharing approach storing
parts of the array itself in order to minimize the extra
memory overhead, thus only a constant and small-sized array
used for booting the whole array itself up will be needed.

Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: <linux/tagptr.h>: introduce tagged pointer

Currently kernel has scattered tagged pointer usages hacked
by hand in plain code, without a unique and portable functionset
to highlight the tagged pointer itself and wrap these hacked code
in order to clean up all over meaningless magic masks.

Therefore, this patch introduces simple generic methods to fold
tags into a pointer integer. It currently supports the last n bits
of the pointer for tags, which can be selected by users.

In addition, it will also be used for the upcoming EROFS filesystem,
which heavily uses tagged pointer approach for high performance
and reducing extra memory allocation.

Link: https://en.wikipedia.org/wiki/Tagged_pointer
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: support tracepoint

Add basic tracepoints for ->readpage{,s}, ->lookup,
->destroy_inode, fill_inode and map_blocks.

Reviewed-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: introduce error injection infrastructure

This patch introduces error injection infrastructure, with it, we can
inject error in any kernel exported common functions which erofs used,
so that it can force erofs running into error paths, it turns out that
tests can cover real rare paths more easily to find bugs.

Reviewed-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: support special inode

This patch adds to support special inode, such as block dev, char,
socket, pipe inode.

Reviewed-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: introduce xattr & acl support

This implements xattr and acl functionalities.

Inline and shared xattrs are introduced for flexibility.
Specifically, if the same xattr occurs for many times
in a large number of inodes or the value of a xattr is so large
that it isn't suitable to be inlined, a shared xattr
kept in the xattr meta will be used instead.

Signed-off-by: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: update Kconfig and Makefile

This commit adds Makefile and Kconfig for erofs, and
updates Makefile and Kconfig files in the fs directory.

Signed-off-by: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add namei functions

This commit adds functions that transfer names to inodes.

Signed-off-by: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add directory operations

This adds functions for directory, mainly readdir.

Signed-off-by: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add inode operations

This adds core functions to get, read an inode.

Signed-off-by: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add raw address_space operations

This commit adds functions for meta and raw data, and also
provides address_space_operations for raw data access.

Signed-off-by: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add super block operations

This commit adds erofs super block operations, including (u)mount,
remount_fs, show_options, statfs, in addition to some private
icache management functions.

Signed-off-by: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add erofs in-memory stuffs

 - erofs_sb_info:
   contains erofs-specific in-memory information.

 - erofs_vnode:
   contains vfs_inode and other fs-specific information.
   same as super block, the only one in-memory definition exists.

 - erofs_map_blocks
   plays a role in the file L2P mapping

Signed-off-by: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: erofs: add on-disk layout

This commit adds the on-disk layout header file of erofs.

Note that the on-disk layout is still WIP, and some fields are
reserved for the future use by design.

Any comments are welcome.

Thanks-to: Li Guifu <liguifu2@huawei.com>
Thanks-to: Sun Qiuyang <sunqiuyang@huawei.com>
Signed-off-by: Miao Xie <miaoxie@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: fsl-dpaa2/ethsw: document nested structs as per kernel-doc

Document nested structs per kernel-doc requirements by moving
all comments before the actual struct.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: fsl-dpaa2/eth: document nested structs as per kernel-doc

Document nested structs per kernel-doc requirements by moving
all comments before the actual struct.

Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: fsl-dpaa2/eth: Remove unused driver version

We never really used the driver version, so no point
in keeping it around.

Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: fsl-dpaa2/eth: Update default hash key

In our documentation, we claim to use a 5-tuple key for Rx hash
distribution of flows. The code however configures a key composed
of all supported header fields.

Update the Rx hash key to contain only the documented fields:
{IP src, IP dst, IP nextproto, L4 src, L4 dst}, which was the
original intention and makes most sense as a default.

Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: fsl-dpaa2/ethsw: Fix error message

Error message was referencing wrong function, fix it.

Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: octeon-usb: Adding SPDX license identifier

Adding appropriate SPDX-License-Identifier (GPL-2) that were missing
from code, header and make files.

Signed-off-by: Georgios Tsotsos <tsotsos@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: don't print device addresses as kernel pointers

Print device addresses as unsigned long, not as kernel pointers.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: TODO: remove entry for convert to standard logging

Gasket/apex drivers now use standard logging, remove TODO entry for
this.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: remove gasket logging header

Gasket logging functions no longer used.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: apex: convert to standard logging

Drop gasket logging calls in favor of standard logging.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: sysfs: convert to standard logging

Drop gasket logging calls in favor of standard logging.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: page table: convert to standard logging

Replace gasket logging calls with standard logging calls.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: ioctl: convert to standard logging

Replace gasket logging calls with standard logging calls.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: interrupt: convert to standard logging

Convert gasket logging calls to standard functions.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: core: convert to standard logging

Use standard logging functions, drop use of gasket log functions.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gasket: save struct device for a gasket device

Save the struct device pointer to a gasket device in gasket's metadata,
to facilitate use of standard logging calls and in anticipation of
non-PCI gasket devices in the future.

Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>