Currently, guard regions are not visible to users except through
/proc/$pid/pagemap, with no explicit visibility at the VMA level.
This makes the feature less useful, as it isn't entirely apparent which
VMAs may have these entries present, especially when performing actions
which walk through memory regions such as those performed by CRIU.
This series addresses this issue by introducing the VM_MAYBE_GUARD flag
which fulfils this role, updating the smaps logic to display an entry for
these.
The semantics of this flag are that a guard region MAY be present if set
(we cannot be sure, as we can't efficiently track whether an
MADV_GUARD_REMOVE finally removes all the guard regions in a VMA) - but if
not set the VMA definitely does NOT have any guard regions present.
It's problematic to establish this flag without further action, because
that means that VMAs with guard regions in them become non-mergeable with
adjacent VMAs for no especially good reason.
To work around this, this series also introduces the concept of 'sticky'
VMA flags - that is flags which:
a. if set in one VMA and not in another still permit those VMAs to be
merged (if otherwise compatible).
b. When they are merged, the resultant VMA must have the flag set.
The VMA logic is updated to propagate these flags correctly.
Additionally, VM_MAYBE_GUARD being an explicit VMA flag allows us to solve
an issue with file-backed guard regions - previously these established an
anon_vma object for file-backed mappings solely to have vma_needs_copy()
correctly propagate guard region mappings to child processes.
We introduce a new flag alias VM_COPY_ON_FORK (which currently only
specifies VM_MAYBE_GUARD) and update vma_needs_copy() to check explicitly
for this flag and to copy page tables if it is present, which resolves this
issue.
Additionally, we add the ability for allow-listed VMA flags to be
atomically writable with only mmap/VMA read locks held.
The only flag we allow so far is VM_MAYBE_GUARD, which we carefully ensure
does not cause any races by being allowed to do so.
This allows us to maintain guard region installation as a read-locked
operation and not endure the overhead of obtaining a write lock here.
Finally we introduce extensive VMA userland tests to assert that the sticky
VMA logic behaves correctly as well as guard region self tests to assert
that smaps visibility is correctly implemented.
v2:
* Separated out userland VMA tests for sticky behaviour as per Suren.
* Added the concept of atomic writable VMA flags as per Pedro and Vlastimil.
* Made VM_MAYBE_GUARD an atomic writable flag so we don't have to take a VMA
write lock in madvise() as per Pedro and Vlastimil.
v1:
https://lore.kernel.org/all/cover.1761756437.git.lorenzo.stoakes@oracle.com/
Lorenzo Stoakes (5):
mm: introduce VM_MAYBE_GUARD and make visible in /proc/$pid/smaps
mm: add atomic VMA flags, use VM_MAYBE_GUARD as such
mm: implement sticky, copy on fork VMA flags
tools/testing/vma: add VMA sticky userland tests
selftests/mm/guard-regions: add smaps visibility test
Documentation/filesystems/proc.rst | 1 +
fs/proc/task_mmu.c | 1 +
include/linux/mm.h | 58 ++++++++++
include/trace/events/mmflags.h | 1 +
mm/madvise.c | 22 ++--
mm/memory.c | 3 +
mm/vma.c | 22 ++--
tools/testing/selftests/mm/guard-regions.c | 120 +++++++++++++++++++++
tools/testing/selftests/mm/vm_util.c | 5 +
tools/testing/selftests/mm/vm_util.h | 1 +
tools/testing/vma/vma.c | 89 +++++++++++++--
tools/testing/vma/vma_internal.h | 35 ++++++
12 files changed, 330 insertions(+), 28 deletions(-)
--
2.51.0
This small series proposes the removal of the
BPF_RI_F_RF_NO_DIRECT XDP flag in favour of page_pool's
internal page_pool_napi_local() check which can override
a non-direct recycle into a direct one if the right
conditions are met.,
This was discussed on the mailing list on several occasions
[1][2].
The first patch adds additional benchmarking code to the page_pool
benchmark.
The second patch has the actual change with a proper explanation
and measurements. It remains to be debated if the whole
BPF_RI_F_RF_NO_DIRECT mechanism should be deleted or only
its use in xdp_return_frame_rx_napi().
There is still the unresolved issue of drivers that don't support
page_pool NAPI recycling. This series could be extended to add
that support. Otherwise those drivers would end up with slow
path recycling for XDP.
[1] https://lore.kernel.org/all/8d165026-1477-46cb-94d4-a01e1da40833@kernel.org/
[2] https://lore.kernel.org/all/20250918084823.372000-1-dtatulea@nvidia.com/
Dragos Tatulea (2):
page_pool: add benchmarking for napi-based recycling
xdp: Delegate fast path return decision to page_pool
drivers/net/veth.c | 2 -
include/linux/filter.h | 22 -----
include/net/xdp.h | 2 +-
kernel/bpf/cpumap.c | 2 -
net/bpf/test_run.c | 2 -
net/core/filter.c | 2 +-
net/core/xdp.c | 24 ++---
.../bench/page_pool/bench_page_pool_simple.c | 92 ++++++++++++++++++-
8 files changed, 104 insertions(+), 44 deletions(-)
--
2.50.1
Hi all! I wanted to resend out this series to respark the discussion
on KTAP version 2. Many of the features proposed are already in use by
KUnit. This would add these features to the KTAP documentation. Note
that all the features of KTAP v2 are backwards compatible. Also, today
is my last day at Google so I will be responding with my personal
email afterwards.
--
This patch series represents the final release of KTAP version 2.
There have been open discussions on version 2 for just over 2
years. This patch series marks the end of KTAP version 2 development
and beginning of the KTAP version 3 development.
The largest component of KTAP version 2 release is the addition of test
metadata to the specification. KTAP metadata could include any test
information that is pertinent for user interaction before or after the
running of the test. For example, the test file path or the test speed.
Example of KTAP Metadata:
KTAP version 2
#:ktap_test: main
#:ktap_arch: uml
1..1
KTAP version 2
#:ktap_test: suite_1
#:ktap_subsystem: example
#:ktap_test_file: lib/test.c
1..2
ok 1 test_1
#:ktap_test: test_2
#:ktap_speed: very_slow
# test_2 has begun
#:custom_is_flaky: true
ok 2 test_2
# suite_1 has passed
ok 1 suite_1
The release also includes some formatting fixes and changes to update
the specification to version 2.
Frank Rowand (2):
ktap_v2: change version to 2-rc in KTAP specification
ktap_v2: change "version 1" to "version 2" in examples
Rae Moar (3):
ktap_v2: add test metadata
ktap_v2: formatting fixes to ktap spec
ktap_v2: change version to 2 in KTAP specification
Documentation/dev-tools/ktap.rst | 273 +++++++++++++++++++++++++++++--
1 file changed, 257 insertions(+), 16 deletions(-)
base-commit: 9de5f847ef8fa205f4fd704a381d32ecb5b66da9
--
2.51.2.1041.gc1ab5b90ca-goog
This series improves the CPU cost of RX token management by adding a
socket option that configures the socket to avoid the xarray allocator
and instead use an niov array and a uref field in niov.
Improvement is ~13% cpu util per RX user thread.
Using kperf, the following results were observed:
Before:
Average RX worker idle %: 13.13, flows 4, test runs 11
After:
Average RX worker idle %: 26.32, flows 4, test runs 11
Two other approaches were tested, but with no improvement. Namely, 1)
using a hashmap for tokens and 2) keeping an xarray of atomic counters
but using RCU so that the hotpath could be mostly lockless. Neither of
these approaches proved better than the simple array in terms of CPU.
The sockopt SO_DEVMEM_AUTORELEASE is added to toggle the optimization.
It defaults to 0 (i.e., optimization on).
Note that prior revs reported only a 5% gain. This lower gain was
measured with cpu frequency boosting (unknowingly) disabled. A
consistent ~13% is measured for both kperf and nccl workloads with cpu
frequency boosting on.
To: David S. Miller <davem(a)davemloft.net>
To: Eric Dumazet <edumazet(a)google.com>
To: Jakub Kicinski <kuba(a)kernel.org>
To: Paolo Abeni <pabeni(a)redhat.com>
To: Simon Horman <horms(a)kernel.org>
To: Kuniyuki Iwashima <kuniyu(a)google.com>
To: Willem de Bruijn <willemb(a)google.com>
To: Neal Cardwell <ncardwell(a)google.com>
To: David Ahern <dsahern(a)kernel.org>
To: Mina Almasry <almasrymina(a)google.com>
To: Arnd Bergmann <arnd(a)arndb.de>
To: Jonathan Corbet <corbet(a)lwn.net>
To: Andrew Lunn <andrew+netdev(a)lunn.ch>
To: Shuah Khan <shuah(a)kernel.org>
Cc: Stanislav Fomichev <sdf(a)fomichev.me>
Cc: netdev(a)vger.kernel.org
Cc: linux-kernel(a)vger.kernel.org
Cc: linux-arch(a)vger.kernel.org
Cc: linux-doc(a)vger.kernel.org
Cc: linux-kselftest(a)vger.kernel.org
Signed-off-by: Bobby Eshleman <bobbyeshleman(a)meta.com>
Changes in v6:
- renamed 'net: devmem: use niov array for token management' to refer to
optionality of new config
- added documentation and tests
- make autorelease flag per-socket sockopt instead of binding
field / sysctl
- many per-patch changes (see Changes sections per-patch)
- Link to v5: https://lore.kernel.org/r/20251023-scratch-bobbyeshleman-devmem-tcp-token-u…
Changes in v5:
- add sysctl to opt-out of performance benefit, back to old token release
- Link to v4: https://lore.kernel.org/all/20250926-scratch-bobbyeshleman-devmem-tcp-token…
Changes in v4:
- rebase to net-next
- Link to v3: https://lore.kernel.org/r/20250926-scratch-bobbyeshleman-devmem-tcp-token-u…
Changes in v3:
- make urefs per-binding instead of per-socket, reducing memory
footprint
- fallback to cleaning up references in dmabuf unbind if socket
leaked tokens
- drop ethtool patch
- Link to v2: https://lore.kernel.org/r/20250911-scratch-bobbyeshleman-devmem-tcp-token-u…
Changes in v2:
- net: ethtool: prevent user from breaking devmem single-binding rule
(Mina)
- pre-assign niovs in binding->vec for RX case (Mina)
- remove WARNs on invalid user input (Mina)
- remove extraneous binding ref get (Mina)
- remove WARN for changed binding (Mina)
- always use GFP_ZERO for binding->vec (Mina)
- fix length of alloc for urefs
- use atomic_set(, 0) to initialize sk_user_frags.urefs
- Link to v1: https://lore.kernel.org/r/20250902-scratch-bobbyeshleman-devmem-tcp-token-u…
---
Bobby Eshleman (6):
net: devmem: rename tx_vec to vec in dmabuf binding
net: devmem: refactor sock_devmem_dontneed for autorelease split
net: devmem: prepare for autorelease rx token management
net: devmem: add SO_DEVMEM_AUTORELEASE for autorelease control
net: devmem: document SO_DEVMEM_AUTORELEASE socket option
net: devmem: add tests for SO_DEVMEM_AUTORELEASE socket option
Documentation/networking/devmem.rst | 70 +++++++++-
include/net/netmem.h | 1 +
include/net/sock.h | 13 +-
include/uapi/asm-generic/socket.h | 2 +
net/core/devmem.c | 54 +++++---
net/core/devmem.h | 4 +-
net/core/sock.c | 152 ++++++++++++++++++----
net/ipv4/tcp.c | 69 ++++++++--
net/ipv4/tcp_ipv4.c | 11 +-
net/ipv4/tcp_minisocks.c | 5 +-
tools/include/uapi/asm-generic/socket.h | 2 +
tools/testing/selftests/drivers/net/hw/devmem.py | 115 +++++++++++++++-
tools/testing/selftests/drivers/net/hw/ncdevmem.c | 20 ++-
13 files changed, 453 insertions(+), 65 deletions(-)
---
base-commit: 255d75ef029f33f75fcf5015052b7302486f7ad2
change-id: 20250829-scratch-bobbyeshleman-devmem-tcp-token-upstream-292be174d503
Best regards,
--
Bobby Eshleman <bobbyeshleman(a)meta.com>
Hello,
This small series is another follow-up to [1], in which I misunderstood
Martin's initial feedback (see [2]). I proposed to make tc-tunnel apply
SO_REUSEPORT once server is brought up. This series updates
start_server_addr to really apply Martin's proposal after his
clarification [3]
[1] https://lore.kernel.org/bpf/20251031-tc_tunnel_improv-v1-0-0ffe44d27eda@boo…
[2] https://lore.kernel.org/bpf/efa3540a-1f52-46ca-9f49-e631a5e3e48c@linux.dev/
[3] https://lore.kernel.org/bpf/4cbabdf1-af2c-490a-a41a-b40c1539c1cb@linux.dev/
Signed-off-by: Alexis Lothoré (eBPF Foundation) <alexis.lothore(a)bootlin.com>
---
Alexis Lothoré (eBPF Foundation) (2):
selftests/bpf: systematically add SO_REUSEADDR in start_server_addr
selftests/bpf: use start_server_str rather than start_reuseport_server in tc_tunnel
tools/testing/selftests/bpf/network_helpers.c | 9 +++++++-
.../selftests/bpf/prog_tests/test_tc_tunnel.c | 27 ++++++++++++----------
2 files changed, 23 insertions(+), 13 deletions(-)
---
base-commit: de0745f7cc98146c70a020bc3a1b73c7f3405282
change-id: 20251104-start-server-soreuseaddr-e442446e2d37
Best regards,
--
Alexis Lothoré, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
From: Steven Rostedt <rostedt(a)goodmis.org>
Commit 64cf7d058a00 ("tracing: Have trace_marker use per-cpu data to read
user space") made an update that fixed both trace_marker and
trace_marker_raw. But the small difference made to trace_marker_raw had a
blatant bug in it that any basic testing would have uncovered.
Unfortunately, the self tests have tests for trace_marker but nothing for
trace_marker_raw which allowed the bug to get upstream.
Add basic selftests to test trace_marker_raw so that this doesn't happen
again.
Signed-off-by: Steven Rostedt (Google) <rostedt(a)goodmis.org>
---
.../ftrace/test.d/00basic/trace_marker_raw.tc | 107 ++++++++++++++++++
1 file changed, 107 insertions(+)
create mode 100644 tools/testing/selftests/ftrace/test.d/00basic/trace_marker_raw.tc
diff --git a/tools/testing/selftests/ftrace/test.d/00basic/trace_marker_raw.tc b/tools/testing/selftests/ftrace/test.d/00basic/trace_marker_raw.tc
new file mode 100644
index 000000000000..7daf7292209e
--- /dev/null
+++ b/tools/testing/selftests/ftrace/test.d/00basic/trace_marker_raw.tc
@@ -0,0 +1,107 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0
+# description: Basic tests on writing to trace_marker_raw
+# requires: trace_marker_raw
+# flags: instance
+
+is_little_endian() {
+ if lscpu | grep -q 'Little Endian'; then
+ echo 1;
+ else
+ echo 0;
+ fi
+}
+
+little=`is_little_endian`
+
+make_str() {
+ id=$1
+ cnt=$2
+
+ if [ $little -eq 1 ]; then
+ val=`printf "\\%03o\\%03o\\%03o\\%03o" \
+ $(($id & 0xff)) \
+ $((($id >> 8) & 0xff)) \
+ $((($id >> 16) & 0xff)) \
+ $((($id >> 24) & 0xff))`
+ else
+ val=`printf "\\%03o\\%03o\\%03o\\%03o" \
+ $((($id >> 24) & 0xff)) \
+ $((($id >> 16) & 0xff)) \
+ $((($id >> 8) & 0xff)) \
+ $(($id & 0xff))`
+ fi
+
+ data=`printf -- 'X%.0s' $(seq $cnt)`
+
+ printf "${val}${data}"
+}
+
+write_buffer() {
+ id=$1
+ size=$2
+
+ # write the string into the raw marker
+ make_str $id $size > trace_marker_raw
+}
+
+
+test_multiple_writes() {
+
+ # Write a bunch of data where the id is the count of
+ # data to write
+ for i in `seq 1 10` `seq 101 110` `seq 1001 1010`; do
+ write_buffer $i $i
+ done
+
+ # add a little buffer
+ echo stop > trace_marker
+
+ # Check to make sure the number of entries is the id (rounded up by 4)
+ awk '/.*: # [0-9a-f]* / {
+ print;
+ cnt = -1;
+ for (i = 0; i < NF; i++) {
+ # The counter is after the "#" marker
+ if ( $i == "#" ) {
+ i++;
+ cnt = strtonum("0x" $i);
+ num = NF - (i + 1);
+ # The number of items is always rounded up by 4
+ cnt2 = int((cnt + 3) / 4) * 4;
+ if (cnt2 != num) {
+ exit 1;
+ }
+ break;
+ }
+ }
+ }
+ // { if (NR > 30) { exit 0; } } ' trace_pipe;
+}
+
+
+get_buffer_data_size() {
+ sed -ne 's/^.*data.*size:\([0-9][0-9]*\).*/\1/p' events/header_page
+}
+
+test_buffer() {
+
+ # The id must be four bytes, test that 3 bytes fails a write
+ if echo -n abc > ./trace_marker_raw ; then
+ echo "Too small of write expected to fail but did not"
+ exit_fail
+ fi
+
+ size=`get_buffer_data_size`
+ echo size = $size
+
+ # Now add a little more than what it can handle
+
+ if write_buffer 0xdeadbeef $size ; then
+ echo "Too big of write expected to fail but did not"
+ exit_fail
+ fi
+}
+
+test_buffer
+test_multiple_writes
--
2.51.0