Factor out a little helper to make the checks for the noiommu driver less
ugly.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20210924155705.4258-3-hch@lst.de
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
We don't need to hold a reference to the group in the driver as well as
obtain a reference to the same group as the first thing
vfio_register_group_dev() does.
Since the drivers never use the group move this all into the core code.
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20210924155705.4258-2-hch@lst.de
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Cross-architecture update to move task_struct::cpu back into thread_info
on arm64, x86, s390, powerpc, and riscv. All Acked by arch maintainers.
Quoting Ard Biesheuvel:
"Move task_struct::cpu back into thread_info
Keeping CPU in task_struct is problematic for architectures that define
raw_smp_processor_id() in terms of this field, as it requires
linux/sched.h to be included, which causes a lot of pain in terms of
circular dependencies (aka 'header soup')
This series moves it back into thread_info (where it came from) for all
architectures that enable THREAD_INFO_IN_TASK, addressing the header
soup issue as well as some pointless differences in the implementations
of task_cpu() and set_task_cpu()."
Remove ntfs_sb_info members sector_size and sector_bits.
Print details why mount failed.
Reviewed-by: Kari Argillander <kari.argillander@gmail.com>
Signed-off-by: Konstantin Komarov <almaz.alexandrovich@paragon-software.com>
If we continue to work in this case, then we can corrupt fs.
Fixes: 82cae269cf ("fs/ntfs3: Add initialization of super block").
Reviewed-by: Kari Argillander <kari.argillander@gmail.com>
Signed-off-by: Konstantin Komarov <almaz.alexandrovich@paragon-software.com>
init_nfsd() should not unregister pernet subsys if the register fails
but should instead unwind from the last successful operation which is
register_filesystem().
Unregistering a failed register_pernet_subsys() call can result in
a kernel GPF as revealed by programmatically injecting an error in
register_pernet_subsys().
Verified the fix handled failure gracefully with no lingering nfsd
entry in /proc/filesystems. This change was introduced by the commit
bd5ae9288d ("nfsd: register pernet ops last, unregister first"),
the original error handling logic was correct.
Fixes: bd5ae9288d ("nfsd: register pernet ops last, unregister first")
Cc: stable@vger.kernel.org
Signed-off-by: Patrick Ho <Patrick.Ho@netapp.com>
Acked-by: J. Bruce Fields <bfields@redhat.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Validate that the transform and smb request headers are present
before checking OriginalMessageSize and SessionId fields.
Cc: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Cc: Ralph Böhme <slow@samba.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Reviewed-by: Tom Talpey <tom@talpey.com>
Acked-by: Hyunchul Lee <hyc.lee@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Add buffer validation for SMB2_CREATE_CONTEXT.
Cc: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Signed-off-by: Hyunchul Lee <hyc.lee@gmail.com>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
This patch add validation to check request buffer check in smb2
negotiate and fix null pointer deferencing oops in smb3_preauth_hash_rsp()
that found from manual test.
Cc: Tom Talpey <tom@talpey.com>
Cc: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Cc: Ralph Böhme <slow@samba.org>
Cc: Hyunchul Lee <hyc.lee@gmail.com>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Add buffer validation in smb2_set_info, and remove unused variable
in set_file_basic_info. and smb2_set_info infolevel functions take
structure pointer argument.
Cc: Tom Talpey <tom@talpey.com>
Cc: Ronnie Sahlberg <ronniesahlberg@gmail.com>
Cc: Ralph Böhme <slow@samba.org>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>
Acked-by: Hyunchul Lee <hyc.lee@gmail.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
Use correct basic info level in set/get_file_basic_info().
Reviewed-by: Ralph Boehme <slow@samba.org>
Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Signed-off-by: Steve French <stfrench@microsoft.com>
The DAI clock is only used in I2S mode, to make it clear
and to fix clock resource release issue, we move CCF clock
related code to rt5682_i2c_probe to fix clock
register/unregister issue.
Signed-off-by: Jack Yu <jack.yu@realtek.com>
Link: https://lore.kernel.org/r/20210929054344.12112-1-jack.yu@realtek.com
Signed-off-by: Mark Brown <broonie@kernel.org>
In prealloc_elems_and_freelist(), the multiplication to calculate the
size passed to bpf_map_area_alloc() could lead to an integer overflow.
As a result, out-of-bounds write could occur in pcpu_freelist_populate()
as reported by KASAN:
[...]
[ 16.968613] BUG: KASAN: slab-out-of-bounds in pcpu_freelist_populate+0xd9/0x100
[ 16.969408] Write of size 8 at addr ffff888104fc6ea0 by task crash/78
[ 16.970038]
[ 16.970195] CPU: 0 PID: 78 Comm: crash Not tainted 5.15.0-rc2+ #1
[ 16.970878] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014
[ 16.972026] Call Trace:
[ 16.972306] dump_stack_lvl+0x34/0x44
[ 16.972687] print_address_description.constprop.0+0x21/0x140
[ 16.973297] ? pcpu_freelist_populate+0xd9/0x100
[ 16.973777] ? pcpu_freelist_populate+0xd9/0x100
[ 16.974257] kasan_report.cold+0x7f/0x11b
[ 16.974681] ? pcpu_freelist_populate+0xd9/0x100
[ 16.975190] pcpu_freelist_populate+0xd9/0x100
[ 16.975669] stack_map_alloc+0x209/0x2a0
[ 16.976106] __sys_bpf+0xd83/0x2ce0
[...]
The possibility of this overflow was originally discussed in [0], but
was overlooked.
Fix the integer overflow by changing elem_size to u64 from u32.
[0] https://lore.kernel.org/bpf/728b238e-a481-eb50-98e9-b0f430ab01e7@gmail.com/
Fixes: 557c0c6e7d ("bpf: convert stackmap to pre-allocation")
Signed-off-by: Tatsuhiko Yasumatsu <th.yasumatsu@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20210930135545.173698-1-th.yasumatsu@gmail.com
Now that the core code switched back to using thread_info::cpu to keep
a task's CPU number, we no longer need to keep it in sync explicitly. So
just drop the code that does this.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Palmer Dabbelt <palmerdabbelt@google.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Instead of relying on awful hacks to obtain the offset of the cpu field
in struct task_struct, move it back into struct thread_info, which does
not create the same level of circular dependency hell when trying to
include the header file that defines it.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Michael Ellerman <mpe@ellerman.id.au>
THREAD_INFO_IN_TASK moved the CPU field out of thread_info, but this
causes some issues on architectures that define raw_smp_processor_id()
in terms of this field, due to the fact that #include'ing linux/sched.h
to get at struct task_struct is problematic in terms of circular
dependencies.
Given that thread_info and task_struct are the same data structure
anyway when THREAD_INFO_IN_TASK=y, let's move it back so that having
access to the type definition of struct thread_info is sufficient to
reference the CPU number of the current task.
Note that this requires THREAD_INFO_IN_TASK's definition of the
task_thread_info() helper to be updated, as task_cpu() takes a
pointer-to-const, whereas task_thread_info() (which is used to generate
lvalues as well), needs a non-const pointer. So make it a macro instead.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Michael Ellerman <mpe@ellerman.id.au>
The CPU field will be moved back into thread_info even when
THREAD_INFO_IN_TASK is enabled, so add it back to powerpc's definition
of struct thread_info.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Michael Ellerman <mpe@ellerman.id.au>
The CPU field will be moved back into thread_info even when
THREAD_INFO_IN_TASK is enabled, so add it back to s390's definition of
struct thread_info.
Note that s390 always has CONFIG_SMP=y so there is no point in guarding
the CPU field with an #ifdef.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Heiko Carstens <hca@linux.ibm.com>
NOTE: This patch intentionally omits any "Fixes:" metadata or stable
tagging since it removes a SELinux access control check; while
removing the control point is the right thing to do moving forward,
removing it in stable kernels could be seen as a regression.
The original SELinux lockdown implementation in 59438b4647
("security,lockdown,selinux: implement SELinux lockdown") used the
current task's credentials as both the subject and object in the
SELinux lockdown hook, selinux_lockdown(). Unfortunately that
proved to be incorrect in a number of cases as the core kernel was
calling the LSM lockdown hook in places where the credentials from
the "current" task_struct were not the correct credentials to use
in the SELinux access check.
Attempts were made to resolve this by adding a credential pointer
to the LSM lockdown hook as well as suggesting that the single hook
be split into two: one for user tasks, one for kernel tasks; however
neither approach was deemed acceptable by Linus. Faced with the
prospect of either changing the subj/obj in the access check to a
constant context (likely the kernel's label) or removing the SELinux
lockdown check entirely, the SELinux community decided that removing
the lockdown check was preferable.
The supporting changes to the general LSM layer are left intact, this
patch only removes the SELinux implementation.
Acked-by: Ondrej Mosnacek <omosnace@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
If CONFIG_OF is disabled, devm_drm_of_get_bridge won't be compiled in
and drivers using that function will fail to build.
Add an inline stub so that we can still build-test those cases.
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested
Link: https://patchwork.freedesktop.org/patch/msgid/20210928181333.1176840-1-maxime@cerno.tech
On i.MX8ULP, for some of the PCCs, it has a peripheral SW RST bit
resides in the same registers as the clock controller. So add this
SW RST controller support alongs with the pcc clock initialization.
the reset and clock shared the same register, to avoid accessing
the same register by reset control and clock control concurrently,
locking is necessary, so reuse the imx_ccm_lock spinlock to simplify
the code.
Suggested-by: Liu Ying <victor.liu@nxp.com>
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/20210914065208.3582128-10-ping.bai@nxp.com
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
Add clock driver for i.MX8ULP.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/20210914065208.3582128-9-ping.bai@nxp.com
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
On i.MX8ULP, the 'CLK_SET_RATE_PARENT' flag should NOT be
set and according to the laest RM, the PFD divider value range
seems will be changed in the future, so update the pfdv2 to
include the specific support for i.MX8ULP.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/20210914065208.3582128-8-ping.bai@nxp.com
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
It is possible that a PFD is enabled in HW but not in SW. That
means the enable count & prepare count of the PFD clock is '0',
so the 'CLK_SET_RATE' flag can do nothing when the rate is changed
while the PFD is hw enabled. In order to safely change the pfd
rate, we can disable the PFD directly if it is hw enabled but not
used by SW end user.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/20210914065208.3582128-7-ping.bai@nxp.com
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
For the imx_composite-7ulp clock type, The clock parent should
be changed explicitly by end user of this clock, if the the
'CLK_SET_RATE_NO_REPARENT' flag is not set, when user want to
set a clock frequency that can NOT get from HW accurately, then
the clock's parent will be switch to another clock parent sometimes.
This is NOT what we expected and introduced some additional debug
effort, so add the 'CLK_SET_RATE_NO_REPARENT' to avoid such unexpected
result.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/20210914065208.3582128-6-ping.bai@nxp.com
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
i.MX7ULP peripheral clock ONLY allow parent/rate to be changed
with clock gated, however, during clock tree initialization, the
peripheral clock could be enabled by bootloader, but the prepare
count in clock tree is still zero, so clock core driver will allow
parent/rate changed even with CLK_SET_RATE_GATE/CLK_SET_PARENT_GATE
set, but the change will fail due to HW NOT allow parent/rate change
with clock enabled. It will cause clock HW status mismatch with
clock tree info and lead to function issue. Below is an example:
usdhc0's pcc clock value is 0xC5000000 during kernel boot up, it
means usdhc0 clock is enabled, its parent is APLL_PFD1. In DT file,
the usdhc0 clock settings are as below:
assigned-clocks = <&pcc2 IMX7ULP_CLK_USDHC0>;
assigned-clock-parents = <&scg1 IMX7ULP_CLK_NIC1_DIV>;
when kernel boot up, the clock tree info is as below, but the usdhc0
PCC register is still 0xC5000000, which means its parent is still
from APLL_PFD1, which is incorrect and cause usdhc0 NOT work.
nic1_clk 2 2 0 176000000 0 0 50000
usdhc0 0 0 0 176000000 0 0 50000
After making sure the peripheral clock is disabled during clock tree
initialization, the usdhc0 is working, and this change is necessary
for all i.MX7ULP peripheral clocks.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/20210914065208.3582128-5-ping.bai@nxp.com
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
On i.MX8ULP, some peripherals have a sw_rst control resides
in the per device PCC clock control register, all others are
same as i.MX7ULP, so update the 7ulp clock composite driver to
support i.MX8ULP to maxmimize the code reuse.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/20210914065208.3582128-4-ping.bai@nxp.com
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
The PLLs used on i.MX8ULP is mostly the same as on i.MX7ULP,
except the PLL register offset is changed. Change the PLLv4
driver for code reuse on i.MX7ULP and i.MX8ULP.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/20210914065208.3582128-3-ping.bai@nxp.com
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
Add the clock dt-binding file for i.MX8ULP.
For pcc node, it will also be used as a reset controller,
so add the '#reset-cells' property description and add the
pcc reset IDs.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/20210914065208.3582128-2-ping.bai@nxp.com
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
It looks much cleaner to just have a macro compared to having
a function that passes NULL as dev to the lower-level
imx_dev_clk_hw_pll14xx.
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/1631521490-17171-9-git-send-email-abel.vesa@nxp.com
Reviewed-by: Stephen Boyd <sboyd@kernel.org>
Rather than having multiple different macros for each different type
of imx8m_clk_hw_composite, implement them in such a way so we can
take advantage the most of the already defined simpler types. Basically,
we end up having one low-level __imx8m_clk_hw_composite function, a
wrapper to simplify the parents related arguments called
_imx8m_clk_hw_composite and then all the types can use those for each
specific case.
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/1631521490-17171-8-git-send-email-abel.vesa@nxp.com
Reviewed-by: Stephen Boyd <sboyd@kernel.org>
Instead of having multiple inline functions that were calling
clk_hw_register_divider, implement a generic low-level
__imx_clk_hw_divider and implement the rest as macros that
pass on as arguments whatever is needed in each case.
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/1631521490-17171-7-git-send-email-abel.vesa@nxp.com
Reviewed-by: Stephen Boyd <sboyd@kernel.org>
Instead of having multiple inline functions that were calling
clk_hw_register_mux, implement a generic low-level __imx_clk_hw_mux
and implement the rest as macros that pass on as arguments whatever
is needed in each case.
Signed-off-by: Abel Vesa <abel.vesa@nxp.com>
Link: https://lore.kernel.org/r/1631521490-17171-6-git-send-email-abel.vesa@nxp.com
Reviewed-by: Stephen Boyd <sboyd@kernel.org>
Jann Horn reported that SO_PEERCRED and SO_PEERGROUPS implementations
are racy, as af_unix can concurrently change sk_peer_pid and sk_peer_cred.
In order to fix this issue, this patch adds a new spinlock that needs
to be used whenever these fields are read or written.
Jann also pointed out that l2cap_sock_get_peer_pid_cb() is currently
reading sk->sk_peer_pid which makes no sense, as this field
is only possibly set by AF_UNIX sockets.
We will have to clean this in a separate patch.
This could be done by reverting b48596d1dc "Bluetooth: L2CAP: Add get_peer_pid callback"
or implementing what was truly expected.
Fixes: 109f6e39fa ("af_unix: Allow SO_PEERCRED to work across namespaces.")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Jann Horn <jannh@google.com>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Cc: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet says:
====================
net: snmp: minor optimizations
Fetching many SNMP counters on hosts with large number of cpus
takes a lot of time. mptcp still uses the old non-batched
fashion which is not cache friendly.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Using snmp_get_cpu_field_batch() allows for better cpu cache
utilization, especially on hosts with large number of cpus.
Also remove special handling when mptcp mibs where not yet
allocated.
I chose to use temporary storage on the stack to keep this patch simple.
We might in the future use the storage allocated in netstat_seq_show().
Combined with prior patch (inlining snmp_get_cpu_field)
time to fetch and output mptcp counters on a 256 cpu host [1]
goes from 75 usec to 16 usec.
[1] L1 cache size is 32KB, it is not big enough to hold all dataset.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This trivial function is called ~90,000 times on 256 cpus hosts,
when reading /proc/net/netstat. And this number keeps inflating.
Inlining it saves many cycles.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add counters for XDP REDIRECT success and failure. This brings the
redirect path in line with metrics gathered via the other XDP paths.
Signed-off-by: Joshua Roys <roysjosh@gmail.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When STMMAC is paired with Energy-Efficient Ethernet(EEE) capable PHY,
and the PHY is advertising EEE by default, we need to enable EEE on the
xPCS side too, instead of having user to manually trigger the enabling
config via ethtool.
Fixed this by adding xpcs_config_eee() call in stmmac_eee_init().
Fixes: 7617af3d1a ("net: pcs: Introducing support for DWC xpcs Energy Efficient Ethernet")
Cc: Michael Sit Wei Hong <michael.wei.hong.sit@intel.com>
Signed-off-by: Wong Vee Khee <vee.khee.wong@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The CPU field will be moved back into thread_info even when
THREAD_INFO_IN_TASK is enabled, so add it back to x86's definition of
struct thread_info.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Borislav Petkov <bp@suse.de>
Acked-by: Mark Rutland <mark.rutland@arm.com>
The CPU field will be moved back into thread_info even when
THREAD_INFO_IN_TASK is enabled, so add it back to arm64's definition of
struct thread_info.
Note that arm64 always has CONFIG_SMP=y so there is no point in guarding
the CPU field with an #ifdef.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Originally, ixgbe driver doesn't allow the mounting of xdpdrv if the
server is equipped with more than 64 cpus online. So it turns out that
the loading of xdpdrv causes the "NOMEM" failure.
Actually, we can adjust the algorithm and then make it work through
mapping the current cpu to some xdp ring with the protect of @tx_lock.
Here are some numbers before/after applying this patch with xdp-example
loaded on the eth0X:
As client (tx path):
Before After
TCP_STREAM send-64 734.14 714.20
TCP_STREAM send-128 1401.91 1395.05
TCP_STREAM send-512 5311.67 5292.84
TCP_STREAM send-1k 9277.40 9356.22 (not stable)
TCP_RR send-1 22559.75 21844.22
TCP_RR send-128 23169.54 22725.13
TCP_RR send-512 21670.91 21412.56
As server (rx path):
Before After
TCP_STREAM send-64 1416.49 1383.12
TCP_STREAM send-128 3141.49 3055.50
TCP_STREAM send-512 9488.73 9487.44
TCP_STREAM send-1k 9491.17 9356.22 (not stable)
TCP_RR send-1 23617.74 23601.60
...
Notice: the TCP_RR mode is unstable as the official document explains.
I tested many times with different parameters combined through netperf.
Though the result is not that accurate, I cannot see much influence on
this patch. The static key is places on the hot path, but it actually
shouldn't cause a huge regression theoretically.
Co-developed-by: Shujin Li <lishujin@kuaishou.com>
Signed-off-by: Shujin Li <lishujin@kuaishou.com>
Signed-off-by: Jason Xing <xingwanli@kuaishou.com>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Wei Wang says:
====================
net: add new socket option SO_RESERVE_MEM
This patch series introduces a new socket option SO_RESERVE_MEM.
This socket option provides a mechanism for users to reserve a certain
amount of memory for the socket to use. When this option is set, kernel
charges the user specified amount of memory to memcg, as well as
sk_forward_alloc. This amount of memory is not reclaimable and is
available in sk_forward_alloc for this socket.
With this socket option set, the networking stack spends less cycles
doing forward alloc and reclaim, which should lead to better system
performance, with the cost of an amount of pre-allocated and
unreclaimable memory, even under memory pressure.
With a tcp_stream test with 10 flows running on a simulated 100ms RTT
link, I can see the cycles spent in __sk_mem_raise_allocated() dropping
by ~0.02%. Not a whole lot, since we already have logic in
sk_mem_uncharge() to only reclaim 1MB when sk_forward_alloc has more
than 2MB free space. But on a system suffering memory pressure
constently, the savings should be more.
The first patch is the implementation of this socket option. The
following 2 patches change the tcp stack to make use of this reserved
memory when under memory pressure. This makes the tcp stack behavior
more flexible when under memory pressure, and provides a way for user to
control the distribution of the memory among its sockets.
With a TCP connection on a simulated 100ms RTT link, the default
throughput under memory pressure is ~500Kbps. With SO_RESERVE_MEM set to
100KB, the throughput under memory pressure goes up to ~3.5Mbps.
Change since v2:
- Added description for new field added in struct sock in patch 1
Change since v1:
- Added performance stats in cover letter and rebased
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
When user sets SO_RESERVE_MEM socket option, in order to utilize the
reserved memory when in memory pressure state, we adjust rcv_ssthresh
according to the available reserved memory for the socket, instead of
using 4 * advmss always.
Signed-off-by: Wei Wang <weiwan@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If user sets SO_RESERVE_MEM socket option, in order to fully utilize the
reserved memory in memory pressure state on the tx path, we modify the
logic in sk_stream_moderate_sndbuf() to set sk_sndbuf according to
available reserved memory, instead of MIN_SOCK_SNDBUF, and adjust it
when new data is acked.
Signed-off-by: Wei Wang <weiwan@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This socket option provides a mechanism for users to reserve a certain
amount of memory for the socket to use. When this option is set, kernel
charges the user specified amount of memory to memcg, as well as
sk_forward_alloc. This amount of memory is not reclaimable and is
available in sk_forward_alloc for this socket.
With this socket option set, the networking stack spends less cycles
doing forward alloc and reclaim, which should lead to better system
performance, with the cost of an amount of pre-allocated and
unreclaimable memory, even under memory pressure.
Note:
This socket option is only available when memory cgroup is enabled and we
require this reserved memory to be charged to the user's memcg. We hope
this could avoid mis-behaving users to abused this feature to reserve a
large amount on certain sockets and cause unfairness for others.
Signed-off-by: Wei Wang <weiwan@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
struct dev_addr_list is used for device addresses, unicast addresses
and multicast addresses. The first of those needs special handling
of the main address - netdev->dev_addr points directly the data
of the entry and drivers write to it freely, so we can't maintain
it in the rbtree (for now, at least, to be fixed in net-next).
Current work around sprinkles special handling of the first
address on the list throughout the code but it missed the case
where address is being added. First address will not be visible
during subsequent adds.
Syzbot found a warning where unicast addresses are modified
without holding the rtnl lock, tl;dr is that team generates
the same modification multiple times, not necessarily when
right locks are held.
In the repro we have:
macvlan -> team -> veth
macvlan adds a unicast address to the team. Team then pushes
that address down to its memebers (veths). Next something unrelated
makes team sync member addrs again, and because of the bug
the addr entries get duplicated in the veths. macvlan gets
removed, removes its addr from team which removes only one
of the duplicated addresses from veths. This removal is done
under rtnl. Next syzbot uses iptables to add a multicast addr
to team (which does not hold rtnl lock). Team syncs veth addrs,
but because veths' unicast list still has the duplicate it will
also get sync, even though this update is intended for mc addresses.
Again, uc address updates need rtnl lock, boom.
Reported-by: syzbot+7a2ab2cdc14d134de553@syzkaller.appspotmail.com
Fixes: 406f42fa0d ("net-next: When a bond have a massive amount of VLANs with IPv6 addresses, performance of changing link state, attaching a VRF, changing an IPv6 address, etc. go down dramtically.")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
We lowered all the other constraint related log messages to debug level so
lower the logging of what supplies we're configuring to debug level too.
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20210929140717.3769-1-broonie@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
