llcp_sock_sendmsg() calls nfc_llcp_send_ui_frame() which in turn calls
nfc_alloc_send_skb(), which accesses the nfc_dev from the llcp_sock for
getting the headroom and tailroom needed for skb allocation.
Parallelly the nfc_dev can be freed, as the refcount is decreased via
nfc_free_device(), leading to a UAF reported by Syzkaller, which can
be summarized as follows:
(1) llcp_sock_sendmsg() -> nfc_llcp_send_ui_frame()
-> nfc_alloc_send_skb() -> Dereference *nfc_dev
(2) virtual_ncidev_close() -> nci_free_device() -> nfc_free_device()
-> put_device() -> nfc_release() -> Free *nfc_dev
When a reference to llcp_local is acquired, we do not acquire the same
for the nfc_dev. This leads to freeing even when the llcp_local is in
use, and this is the case with the UAF described above too.
Thus, when we acquire a reference to llcp_local, we should acquire a
reference to nfc_dev, and release the references appropriately later.
References for llcp_local is initialized in nfc_llcp_register_device()
(which is called by nfc_register_device()). Thus, we should acquire a
reference to nfc_dev there.
nfc_unregister_device() calls nfc_llcp_unregister_device() which in
turn calls nfc_llcp_local_put(). Thus, the reference to nfc_dev is
appropriately released later.
Reported-and-tested-by: syzbot+bbe84a4010eeea00982d@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=bbe84a4010eeea00982d
Fixes: c7aa12252f ("NFC: Take a reference on the LLCP local pointer when creating a socket")
Reviewed-by: Suman Ghosh <sumang@marvell.com>
Signed-off-by: Siddh Raman Pant <code@siddh.me>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 2f3ce7a56c ("net: sfp: rework the RollBall PHY waiting code")
changed the long wait before accessing RollBall / FS modules into
probing for PHY every 1 second, and trying 25 times.
Wei Lei reports that this does not work correctly on FS modules: when
initializing, they may report values different from 0xffff in PHY ID
registers for some MMDs, causing get_phy_c45_ids() to find some bogus
MMD.
Fix this by adding the module_t_wait member back, and setting it to 4
seconds for FS modules.
Fixes: 2f3ce7a56c ("net: sfp: rework the RollBall PHY waiting code")
Reported-by: Wei Lei <quic_leiwei@quicinc.com>
Signed-off-by: Marek Behún <kabel@kernel.org>
Tested-by: Lei Wei <quic_leiwei@quicinc.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
The BQ24296 is most similar to the BQ24196, but the:
1. OTG config is split from CHG config (REG01)
2. ICHG (Fast Charge Current limit) range is smaller (<=3008mA)
3. NTC fault is simplified to 2 bits
Signed-off-by: Hermes Zhang <chenhuiz@axis.com>
Link: https://lore.kernel.org/r/20231208034708.1248389-3-Hermes.Zhang@axis.com
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
The BQ24296 is most similar to the BQ24196, but the:
1. OTG config is split from CHG config (REG01)
2. ICHG (Fast Charge Current limit) range is smaller (<=3008mA)
3. NTC fault is simplified to 2 bits
Signed-off-by: Hermes Zhang <chenhuiz@axis.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://lore.kernel.org/r/20231208034708.1248389-2-Hermes.Zhang@axis.com
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Marek Report a possible irq lock inversion dependency warning when
commit 81a06f1d02 ("Revert "drm/rockchip: vop2: Use regcache_sync()
to fix suspend/resume"") lands linux-next.
I can reproduce this warning with:
CONFIG_PROVE_LOCKING=y
CONFIG_DEBUG_LOCKDEP=y
It seems than when use regmap_reinit_cache at runtime whith Mark's
commit 3d59c22bbb ("drm/rockchip: vop2: Convert to use maple tree
register cache"), it will trigger a possible irq lock inversion dependency
warning.
One solution is switch back to REGCACHE_RBTREE, but it seems that
REGCACHE_MAPLE is the future, so I avoid using regmap_reinit_cache,
and drop all the regcache when vop is disabled, then we get a fresh
start at next enbable time.
Fixes: 81a06f1d02 ("Revert "drm/rockchip: vop2: Use regcache_sync() to fix suspend/resume"")
Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
Closes: https://lore.kernel.org/all/98a9f15d-30ac-47bf-9b93-3aa2c9900f7b@samsung.com/
Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
[dropped the large kernel log of the lockdep report from the message]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20231217084415.2373043-1-andyshrk@163.com
Currently, a not-yet-entirely-initialized battmgr (e.g. with pd-mapper
not having yet started or ADSP not being up etc.) results in a couple of
zombie power supply devices hanging around.
This is particularly noticeable when trying to suspend the device (even
s2idle): the PSY-internal thermal zone is inaccessible and returns
-ENODEV, which causes log spam.
Register the power supplies only after we received some notification
indicating battmgr is ready to take off.
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Tested-by: Luca Weiss <luca.weiss@fairphone.com>
Link: https://lore.kernel.org/r/20231218-topic-battmgr_fixture_attempt-v1-1-6145745f34fe@linaro.org
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
The current PON example is a bit of a mess after converting the binding
document to yaml and in the process updating parts of the example to
match the pmk8350 binding while leaving parts from the older pm8998
example in place.
Clean up the example and make it consistent by adding some newline
separators; dropping labels; removing stray spaces; fixing the PON node
name; and fixing the unit address so that it matches the interrupt
specifiers (which re-encodes the PON base address, 0x800 => 0x8).
Fixes: 76ba1900cb ("dt-bindings: power: reset: qcom-pon: Convert qcom PON binding to yaml")
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20231130173017.12723-1-johan+linaro@kernel.org
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
devm_kasprintf and devm_kzalloc return a pointer to dynamically
allocated memory which can be NULL upon failure.
Fixes: 8648aeb5d7 ("power: supply: add Qualcomm PMI8998 SMB2 Charger driver")
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
Link: https://lore.kernel.org/r/20231124075021.1335289-1-chentao@kylinos.cn
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
When building with clang, there are two section mismatch warnings:
WARNING: modpost: vmlinux: section mismatch in reference: at91_poweroff_probe+0x7c (section: .text) -> at91_wakeup_status (section: .init.text)
WARNING: modpost: vmlinux: section mismatch in reference: at91_shdwc_probe+0xcc (section: .text) -> at91_wakeup_status (section: .init.text)
Drop '__init' from at91_wakeup_status() to clear up the mismatch.
Fixes: dde74a5de8 ("power: reset: at91-sama5d2_shdwc: Stop using module_platform_driver_probe()")
Fixes: 099806de68 ("power: reset: at91-poweroff: Stop using module_platform_driver_probe()")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Acked-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
This resolves checkpatch warning "quoted string split across lines" on:
1640: WARNING: quoted string split across lines
1641: WARNING: quoted string split across lines
The motive to use multiple MODULE_AUTHOR statements came from this
comment from "include/linux/module.h":
/*
* Author(s), use "Name <email>" or just "Name", for multiple
* authors use multiple MODULE_AUTHOR() statements/lines.
*/
#define MODULE_AUTHOR(_author) MODULE_INFO(author, _author)
Signed-off-by: Charalampos Mitrodimas <charmitro@posteo.net>
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
These were mentioned by checkpatch:
Errors:
(1) code indent should use tabs where possible
(2) switch and case should be at the same indent
Warnings:
(1) Missing a blank line after declarations
Signed-off-by: Charalampos Mitrodimas <charmitro@posteo.net>
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Use device life-cycle managed register function to simplify probe error
path and eliminate need for explicit remove function.
Signed-off-by: Andrew Davis <afd@ti.com>
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
When the region driver fails while allocating HPA space for a
new region it can be because the parent resource, the CXL Window,
has no more available space.
In that case, the debug user sees this message:
cxl_core:alloc_hpa:555: cxl region2: failed to allocate HPA: -34
Expand the message like this:
cxl_core:alloc_hpa:555: cxl region8: HPA allocation error (-34) for size:0x20000000 in CXL Window 0 [mem 0xf010000000-0xf04fffffff flags 0x200]
Now the debug user can examine /proc/iomem and consider actions
like removing other allocations in that space or reducing the
size of their region request.
Suggested-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Alison Schofield <alison.schofield@intel.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Reviewed-by: Vishal Verma <vishal.l.verma@intel.com>
Reviewed-by: Fan Ni <fan.ni@samsung.com>
Link: https://lore.kernel.org/r/20231223004740.1401858-1-alison.schofield@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Some ioctl commands do not require ioctl permission, but are routed to
other permissions such as FILE_GETATTR or FILE_SETATTR. This routing is
done by comparing the ioctl cmd to a set of 64-bit flags (FS_IOC_*).
However, if a 32-bit process is running on a 64-bit kernel, it emits
32-bit flags (FS_IOC32_*) for certain ioctl operations. These flags are
being checked erroneously, which leads to these ioctl operations being
routed to the ioctl permission, rather than the correct file
permissions.
This was also noted in a RED-PEN finding from a while back -
"/* RED-PEN how should LSM module know it's handling 32bit? */".
This patch introduces a new hook, security_file_ioctl_compat(), that is
called from the compat ioctl syscall. All current LSMs have been changed
to support this hook.
Reviewing the three places where we are currently using
security_file_ioctl(), it appears that only SELinux needs a dedicated
compat change; TOMOYO and SMACK appear to be functional without any
change.
Cc: stable@vger.kernel.org
Fixes: 0b24dcb7f2 ("Revert "selinux: simplify ioctl checking"")
Signed-off-by: Alfred Piccioni <alpic@google.com>
Reviewed-by: Stephen Smalley <stephen.smalley.work@gmail.com>
[PM: subject tweak, line length fixes, and alignment corrections]
Signed-off-by: Paul Moore <paul@paul-moore.com>
The QoS blocks saved/restored when toggling the PD_USB power domain are
clocked by ACLK_USB. Attempting to access these memory regions without
that clock running will result in an indefinite CPU stall.
The PD_USB node wasn't specifying this clock dependency, resulting in
hangs when trying to toggle the power domain (either on or off), unless
we get "lucky" and have ACLK_USB running for another reason at the time.
This "luck" can result from the bootloader leaving USB powered/clocked,
and if no built-in driver wants USB, Linux will disable the unused
PD+CLK on boot when {pd,clk}_ignore_unused aren't given. This can also
be unlucky because the two cleanup tasks run in parallel and race: if
the CLK is disabled first, the PD deactivation stalls the boot. In any
case, the PD cannot then be reenabled (if e.g. the driver loads later)
once the clock has been stopped.
Fix this by specifying a dependency on ACLK_USB, instead of only
ACLK_USB_ROOT. The child-parent relationship means the former implies
the latter anyway.
Fixes: c9211fa260 ("arm64: dts: rockchip: Add base DT for rk3588 SoC")
Cc: stable@vger.kernel.org
Signed-off-by: Sam Edwards <CFSworks@gmail.com>
Link: https://lore.kernel.org/r/20231216021019.1543811-1-CFSworks@gmail.com
[changed to only include the missing clock, not dropping the root-clocks]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The RK806 on the NanoPC-T6 can be used to power on/off the whole board.
Mark it as the system power controller.
Signed-off-by: Hugh Cole-Baker <sigmaris@gmail.com>
Link: https://lore.kernel.org/r/20231216212134.23314-1-sigmaris@gmail.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Cool Pi CM5 EVB works as a mother board connect with CM5.
CM5 Specification:
- Rockchip RK3588
- LPDDR4 2/4/8/16 GB
- TF scard slot
- eMMC 8/32/64/128 GB module
- Gigabit ethernet x 1 with PHY YT8531
- Gigabit ethernet x 1 drived by PCIE with YT6801S
CM5 EVB Specification:
- HDMI Type A out x 2
- HDMI Type D in x 1
- USB 2.0 Host x 2
- USB 3.0 OTG x 1
- USB 3.0 Host x 1
- PCIE M.2 E Key for Wireless connection
- PCIE M.2 M Key for NVME connection
- 40 pin header
Signed-off-by: Andy Yan <andyshrk@163.com>
Link: https://lore.kernel.org/r/20231212124407.1897604-1-andyshrk@163.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Add Cool Pi CM5, a board powered by RK3588
CM5 EVB works with a mother board connect with
CM5
Signed-off-by: Andy Yan <andyshrk@163.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20231212124340.1897502-1-andyshrk@163.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
CoolPi 4B is a rk3588s based SBC.
Specification:
- Rockchip RK3588S
- LPDDR4 2/4/8/16 GB
- TF scard slot
- eMMC 8/32/64/128 GB module
- Gigabit ethernet drived by PCIE with RTL8111HS
- HDMI Type D out
- Mini DP out
- USB 2.0 Host x 2
- USB 3.0 OTG x 1
- USB 3.0 Host x 1
- WIFI/BT module AIC8800
- 40 pin header
Signed-off-by: Andy Yan <andyshrk@163.com>
arm64: dts: rockchip: Add support for rk3588s based board Cool Pi 4B
Link: https://lore.kernel.org/r/20231212124253.1897438-1-andyshrk@163.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Add Cool Pi 4B, a SBC powered by RK3588S
Signed-off-by: Andy Yan <andyshrk@163.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20231212124237.1897378-1-andyshrk@163.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Add names to the pins of the general-purpose expansion header as given
in the Radxa GPIO page[1] following the conventions in the kernel
documentation[2] to make it easier for users to correlate the pins with
functions when using utilities such as 'gpioinfo'.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Link: https://lore.kernel.org/r/20231213160556.14424-1-twoerner@gmail.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This should be a per board property, should not be put in
a soc core dtsi.
And when this property convert from default-sample-phase
in linux-5.7 by commit 8a385eb572 ("ARM: dts: rockchip: fix
rockchip,default-sample-phase property names"), the emmc
on rk3036 kylin board get a initialising error:
[ 4.512797] Freeing unused kernel memory: 8192K
[ 4.519500] mmc_host mmc1: Bus speed (slot 0) = 37125000Hz (slot req 37500000Hz, actual 37125000HZ div = 0)
[ 4.530971] mmc1: error -84 whilst initialising MMC card
[ 4.537277] Run /init as init process
[ 4.550932] mmc_host mmc1: Bus speed (slot 0) = 300000Hz (slot req 300000Hz, actual 300000HZ div = 0)
[ 4.664717] mmc_host mmc1: Bus speed (slot 0) = 37125000Hz (slot req 37500000Hz, actual 37125000HZ div = 0)
[ 4.676156] mmc1: error -84 whilst initialising MMC card
I think the reason why the emmc on rk3036 kylin board was able
to work before linux-5.7 was that the illegal property was not
correctly identified by the rockchip dw_mmc driver.
Fixes: faea098e18 ("ARM: dts: rockchip: add core rk3036 dtsi")
Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
Reviewed-by: Shawn Lin <shawn.lin@rock-chips.com>
Link: https://lore.kernel.org/r/20231218105523.2478315-4-andyshrk@163.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Fold the afs_addr_cursor struct into the afs_operation struct and the
afs_vl_cursor struct and fold its operations into their callers also.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Use the rxrpc_peer plus the service ID as the call address instead of
passing in a sockaddr_srx down to rxrpc. The peer record is obtained by
using rxrpc_kernel_get_peer(). This avoids the need to repeatedly look up
the peer and allows rxrpc to hold on to resources for it.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Rename the ->index and ->untried fields of the afs_vl_cursor and
afs_operation struct to ->server_index and ->untried_servers to avoid
confusion with address iteration fields when those get folded in.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Add a tracepoint to track the lifetime of the afs_addr_list struct.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Simplify error handling a bit by moving it from the afs_addr_cursor struct
to the afs_operation and afs_vl_cursor structs and using the error
prioritisation function for accumulating errors from multiple sources (AFS
tries to rotate between multiple fileservers, some of which may be
inaccessible or in some state of offlinedness).
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Don't put the afs_call struct in afs_wait_for_call_to_complete() but rather
have the caller do it. This will allow the caller to fish stuff out of the
afs_call struct rather than the afs_addr_cursor struct, thereby allowing a
subsequent patch to subsume it.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Wrap most op->error accesses with inline funcs which will make it easier
for a subsequent patch to replace op->error with something else. Two
functions are added to this end:
(1) afs_op_error() - Get the error code.
(2) afs_op_set_error() - Set the error code.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Set op->nr_iterations to -1 to indicate that we need to begin fileserver
iteration rather than setting error to SHRT_MAX. This makes it easier to
eliminate the address cursor.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
When processing the result of a call, handle the VIO and UAEIO abort
specifically rather than leaving it to a default case. Rather than
erroring out unconditionally, see if there's another server if the volume
has more than one server available, otherwise return -EREMOTEIO.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Rename the failed member of struct addr_list to probe_failed as it's
specifically related to probe failures.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
In the rotation algorithms for iterating over volume location servers and
file servers, don't skip servers from which we got a valid response to a
probe (either a reply DATA packet or an ABORT) even if we didn't manage to
get an RTT reading.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Change rxrpc's API such that:
(1) A new function, rxrpc_kernel_lookup_peer(), is provided to look up an
rxrpc_peer record for a remote address and a corresponding function,
rxrpc_kernel_put_peer(), is provided to dispose of it again.
(2) When setting up a call, the rxrpc_peer object used during a call is
now passed in rather than being set up by rxrpc_connect_call(). For
afs, this meenat passing it to rxrpc_kernel_begin_call() rather than
the full address (the service ID then has to be passed in as a
separate parameter).
(3) A new function, rxrpc_kernel_remote_addr(), is added so that afs can
get a pointer to the transport address for display purposed, and
another, rxrpc_kernel_remote_srx(), to gain a pointer to the full
rxrpc address.
(4) The function to retrieve the RTT from a call, rxrpc_kernel_get_srtt(),
is then altered to take a peer. This now returns the RTT or -1 if
there are insufficient samples.
(5) Rename rxrpc_kernel_get_peer() to rxrpc_kernel_call_get_peer().
(6) Provide a new function, rxrpc_kernel_get_peer(), to get a ref on a
peer the caller already has.
This allows the afs filesystem to pin the rxrpc_peer records that it is
using, allowing faster lookups and pointer comparisons rather than
comparing sockaddr_rxrpc contents. It also makes it easier to get hold of
the RTT. The following changes are made to afs:
(1) The addr_list struct's addrs[] elements now hold a peer struct pointer
and a service ID rather than a sockaddr_rxrpc.
(2) When displaying the transport address, rxrpc_kernel_remote_addr() is
used.
(3) The port arg is removed from afs_alloc_addrlist() since it's always
overridden.
(4) afs_merge_fs_addr4() and afs_merge_fs_addr6() do peer lookup and may
now return an error that must be handled.
(5) afs_find_server() now takes a peer pointer to specify the address.
(6) afs_find_server(), afs_compare_fs_alists() and afs_merge_fs_addr[46]{}
now do peer pointer comparison rather than address comparison.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Turn the afs_addr_list address array into an array of structs, thereby
allowing per-address (such as RTT) info to be added.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Add some comments on AFS abort code handling in the rotation algorithm and
adjust the errors produced to match.
Reported-by: Jeffrey E Altman <jaltman@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Jeffrey Altman <jaltman@auristor.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
rxrpc_find_service_conn_rcu() should make the "seq" counter odd on the
second pass, otherwise read_seqbegin_or_lock() never takes the lock.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Link: https://lore.kernel.org/r/20231117164846.GA10410@redhat.com/
David Howells says:
(3) afs_check_validity().
(4) afs_getattr().
These are both pretty short, so your solution is probably good for them.
That said, afs_vnode_commit_status() can spend a long time under the
write lock - and pretty much every file RPC op returns a status update.
Change these functions to use read_seqbegin(). This simplifies the code
and doesn't change the current behaviour, the "seq" counter is always even
so read_seqbegin_or_lock() can never take the lock.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Link: https://lore.kernel.org/r/20231130115617.GA21584@redhat.com/
David Howells says:
(5) afs_find_server().
There could be a lot of servers in the list and each server can have
multiple addresses, so I think this would be better with an exclusive
second pass.
The server list isn't likely to change all that often, but when it does
change, there's a good chance several servers are going to be
added/removed one after the other. Further, this is only going to be
used for incoming cache management/callback requests from the server,
which hopefully aren't going to happen too often - but it is remotely
drivable.
(6) afs_find_server_by_uuid().
Similarly to (5), there could be a lot of servers to search through, but
they are in a tree not a flat list, so it should be faster to process.
Again, it's not likely to change that often and, again, when it does
change it's likely to involve multiple changes. This can be driven
remotely by an incoming cache management request but is mostly going to
be driven by setting up or reconfiguring a volume's server list -
something that also isn't likely to happen often.
Make the "seq" counter odd on the 2nd pass, otherwise read_seqbegin_or_lock()
never takes the lock.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Link: https://lore.kernel.org/r/20231130115614.GA21581@redhat.com/
David Howells says:
(2) afs_lookup_volume_rcu().
There can be a lot of volumes known by a system. A thousand would
require a 10-step walk and this is drivable by remote operation, so I
think this should probably take a lock on the second pass too.
Make the "seq" counter odd on the 2nd pass, otherwise read_seqbegin_or_lock()
never takes the lock.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Link: https://lore.kernel.org/r/20231130115606.GA21571@redhat.com/
Add three iov_iter structs:
(1) Add an iov_iter (->iter) to the I/O request to describe the
unencrypted-side buffer.
(2) Add an iov_iter (->io_iter) to the I/O request to describe the
encrypted-side I/O buffer. This may be a different size to the buffer
in (1).
(3) Add an iov_iter (->io_iter) to the I/O subrequest to describe the part
of the I/O buffer for that subrequest.
This will allow future patches to point to a bounce buffer instead for
purposes of handling oversize writes, decryption (where we want to save the
encrypted data to the cache) and decompression.
These iov_iters persist for the lifetime of the (sub)request, and so can be
accessed multiple times without worrying about them being deallocated upon
return to the caller.
The network filesystem must appropriately advance the iterator before
terminating the request.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Jeff Layton <jlayton@kernel.org>
cc: linux-cachefs@redhat.com
cc: linux-fsdevel@vger.kernel.org
cc: linux-mm@kvack.org
Borrow NFS's direct-vs-buffered I/O locking into netfslib. Similar code is
also used in ceph.
Modify it to have the correct checker annotations for i_rwsem lock
acquisition/release and to return -ERESTARTSYS if waits are interrupted.
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
cc: linux-cachefs@redhat.com
cc: linux-fsdevel@vger.kernel.org
cc: linux-mm@kvack.org
Provide default invalidate_folio and release_folio calls. These will need
to interact with invalidation correctly at some point. They will be needed
if netfslib is to make use of folio->private for its own purposes.
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
cc: linux-cachefs@redhat.com
cc: linux-fsdevel@vger.kernel.org
cc: linux-mm@kvack.org
