Clean up the KVM clock mess somewhat so that it is either based on the guest
TSC ("master clock" mode), or on the host CLOCK_MONOTONIC_RAW in cases where
the TSC isn't usable.
Eliminate the third variant where it was based directly on the *host* TSC,
due to bugs in e.g. __get_kvmclock().
Kill off the last vestiges of the KVM clock being based on CLOCK_MONOTONIC
instead of CLOCK_MONOTONIC_RAW and thus being subject to NTP skew.
Fix up migration support to allow the KVM clock to be saved/restored as an
arithmetic function of the guest TSC, since that's what it actually is in
the *common* case so it can be migrated precisely. Or at least to within
±1 ns which is good enough, as discussed in
https://lore.kernel.org/kvm/c8dca08bf848e663f192de6705bf04aa3966e856.camel@…
In v2 of this series, TSC synchronization is improved and simplified a bit
too, and we allow masterclock mode to be used even when the guest TSCs are
out of sync, as long as they're running at the same *rate*. The different
*offset* shouldn't matter.
And the kvm_get_time_scale() function annoyed me by being entirely opaque,
so I studied it until my brain hurt and then added some comments.
In v2 I also dropped the commits which were removing the periodic clock
syncs. Those are going to be needed still but *only* for non-masterclock
mode, which I'll do next. Along with ensuring that a masterclock update
while already in masterclock mode doesn't jump the clock, and just does
the same as KVM_SET_CLOCK_GUEST does to preserve it.
Needs a *lot* more testing. I think I'm almost done refactoring the code,
so should focus on building up the tests next.
(I do still hate that we're abusing KVM_GET_CLOCK just to get the tuple
of {host_tsc, CLOCK_REALTIME} without even *caring* about the eponymous
KVM clock. Especially as this information is (a) fundamentally what the
vDSO gettimeofday() exposes to us anyway, (b) using CLOCK_REALTIME not
TAI, (c) not available on other platforms, for example for migrating
the Arm arch counter.)
David Woodhouse (13):
KVM: x86/xen: Do not corrupt KVM clock in kvm_xen_shared_info_init()
KVM: x86: Improve accuracy of KVM clock when TSC scaling is in force
KVM: x86: Explicitly disable TSC scaling without CONSTANT_TSC
KVM: x86: Add KVM_VCPU_TSC_SCALE and fix the documentation on TSC migration
KVM: x86: Avoid NTP frequency skew for KVM clock on 32-bit host
KVM: x86: Fix KVM clock precision in __get_kvmclock()
KVM: x86: Fix software TSC upscaling in kvm_update_guest_time()
KVM: x86: Simplify and comment kvm_get_time_scale()
KVM: x86: Remove implicit rdtsc() from kvm_compute_l1_tsc_offset()
KVM: x86: Improve synchronization in kvm_synchronize_tsc()
KVM: x86: Kill cur_tsc_{nsec,offset,write} fields
KVM: x86: Allow KVM master clock mode when TSCs are offset from each other
KVM: x86: Factor out kvm_use_master_clock()
Jack Allister (2):
KVM: x86: Add KVM_[GS]ET_CLOCK_GUEST for accurate KVM clock migration
KVM: selftests: Add KVM/PV clock selftest to prove timer correction
Documentation/virt/kvm/api.rst | 37 ++
Documentation/virt/kvm/devices/vcpu.rst | 115 +++-
arch/x86/include/asm/kvm_host.h | 15 +-
arch/x86/include/uapi/asm/kvm.h | 6 +
arch/x86/kvm/svm/svm.c | 3 +-
arch/x86/kvm/vmx/vmx.c | 2 +-
arch/x86/kvm/x86.c | 687 +++++++++++++++-------
arch/x86/kvm/xen.c | 4 +-
include/uapi/linux/kvm.h | 3 +
tools/testing/selftests/kvm/Makefile | 1 +
tools/testing/selftests/kvm/x86_64/pvclock_test.c | 192 ++++++
11 files changed, 822 insertions(+), 243 deletions(-)
6.6-stable review patch. If anyone has any objections, please let me know.
------------------
From: Mark Rutland <mark.rutland(a)arm.com>
[ Upstream commit 8ecab2e64572f1aecdfc5a8feae748abda6e3347 ]
The event filter function test has been failing in our internal test
farm:
| # not ok 33 event filter function - test event filtering on functions
Running the test in verbose mode indicates that this is because the test
erroneously determines that kmem_cache_free() is the most common caller
of kmem_cache_free():
# # + cut -d: -f3 trace
# # + sed s/call_site=([^+]*)+0x.*/1/
# # + sort
# # + uniq -c
# # + sort
# # + tail -n 1
# # + sed s/^[ 0-9]*//
# # + target_func=kmem_cache_free
... and as kmem_cache_free() doesn't call itself, setting this as the
filter function for kmem_cache_free() results in no hits, and
consequently the test fails:
# # + grep kmem_cache_free trace
# # + grep kmem_cache_free
# # + wc -l
# # + hitcnt=0
# # + grep kmem_cache_free trace
# # + grep -v kmem_cache_free
# # + wc -l
# # + misscnt=0
# # + [ 0 -eq 0 ]
# # + exit_fail
This seems to be because the system in question has tasks with ':' in
their name (which a number of kernel worker threads have). These show up
in the trace, e.g.
test:.sh-1299 [004] ..... 2886.040608: kmem_cache_free: call_site=putname+0xa4/0xc8 ptr=000000000f4d22f4 name=names_cache
... and so when we try to extact the call_site with:
cut -d: -f3 trace | sed 's/call_site=\([^+]*\)+0x.*/\1/'
... the 'cut' command will extrace the column containing
'kmem_cache_free' rather than the column containing 'call_site=...', and
the 'sed' command will leave this unchanged. Consequently, the test will
decide to use 'kmem_cache_free' as the filter function, resulting in the
failure seen above.
Fix this by matching the 'call_site=<func>' part specifically to extract
the function name.
Signed-off-by: Mark Rutland <mark.rutland(a)arm.com>
Reported-by: Aishwarya TCV <aishwarya.tcv(a)arm.com>
Cc: Masami Hiramatsu <mhiramat(a)kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers(a)efficios.com>
Cc: Shuah Khan <shuah(a)kernel.org>
Cc: Steven Rostedt <rostedt(a)goodmis.org>
Cc: linux-kernel(a)vger.kernel.org
Cc: linux-kselftest(a)vger.kernel.org
Cc: linux-trace-kernel(a)vger.kernel.org
Acked-by: Masami Hiramatsu (Google) <mhiramat(a)kernel.org>
Signed-off-by: Shuah Khan <skhan(a)linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal(a)kernel.org>
---
.../selftests/ftrace/test.d/filter/event-filter-function.tc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/testing/selftests/ftrace/test.d/filter/event-filter-function.tc b/tools/testing/selftests/ftrace/test.d/filter/event-filter-function.tc
index 2de7c61d1ae30..3f74c09c56b62 100644
--- a/tools/testing/selftests/ftrace/test.d/filter/event-filter-function.tc
+++ b/tools/testing/selftests/ftrace/test.d/filter/event-filter-function.tc
@@ -24,7 +24,7 @@ echo 0 > events/enable
echo "Get the most frequently calling function"
sample_events
-target_func=`cut -d: -f3 trace | sed 's/call_site=\([^+]*\)+0x.*/\1/' | sort | uniq -c | sort | tail -n 1 | sed 's/^[ 0-9]*//'`
+target_func=`cat trace | grep -o 'call_site=\([^+]*\)' | sed 's/call_site=//' | sort | uniq -c | sort | tail -n 1 | sed 's/^[ 0-9]*//'`
if [ -z "$target_func" ]; then
exit_fail
fi
--
2.43.0
6.8-stable review patch. If anyone has any objections, please let me know.
------------------
From: Mark Rutland <mark.rutland(a)arm.com>
[ Upstream commit 8ecab2e64572f1aecdfc5a8feae748abda6e3347 ]
The event filter function test has been failing in our internal test
farm:
| # not ok 33 event filter function - test event filtering on functions
Running the test in verbose mode indicates that this is because the test
erroneously determines that kmem_cache_free() is the most common caller
of kmem_cache_free():
# # + cut -d: -f3 trace
# # + sed s/call_site=([^+]*)+0x.*/1/
# # + sort
# # + uniq -c
# # + sort
# # + tail -n 1
# # + sed s/^[ 0-9]*//
# # + target_func=kmem_cache_free
... and as kmem_cache_free() doesn't call itself, setting this as the
filter function for kmem_cache_free() results in no hits, and
consequently the test fails:
# # + grep kmem_cache_free trace
# # + grep kmem_cache_free
# # + wc -l
# # + hitcnt=0
# # + grep kmem_cache_free trace
# # + grep -v kmem_cache_free
# # + wc -l
# # + misscnt=0
# # + [ 0 -eq 0 ]
# # + exit_fail
This seems to be because the system in question has tasks with ':' in
their name (which a number of kernel worker threads have). These show up
in the trace, e.g.
test:.sh-1299 [004] ..... 2886.040608: kmem_cache_free: call_site=putname+0xa4/0xc8 ptr=000000000f4d22f4 name=names_cache
... and so when we try to extact the call_site with:
cut -d: -f3 trace | sed 's/call_site=\([^+]*\)+0x.*/\1/'
... the 'cut' command will extrace the column containing
'kmem_cache_free' rather than the column containing 'call_site=...', and
the 'sed' command will leave this unchanged. Consequently, the test will
decide to use 'kmem_cache_free' as the filter function, resulting in the
failure seen above.
Fix this by matching the 'call_site=<func>' part specifically to extract
the function name.
Signed-off-by: Mark Rutland <mark.rutland(a)arm.com>
Reported-by: Aishwarya TCV <aishwarya.tcv(a)arm.com>
Cc: Masami Hiramatsu <mhiramat(a)kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers(a)efficios.com>
Cc: Shuah Khan <shuah(a)kernel.org>
Cc: Steven Rostedt <rostedt(a)goodmis.org>
Cc: linux-kernel(a)vger.kernel.org
Cc: linux-kselftest(a)vger.kernel.org
Cc: linux-trace-kernel(a)vger.kernel.org
Acked-by: Masami Hiramatsu (Google) <mhiramat(a)kernel.org>
Signed-off-by: Shuah Khan <skhan(a)linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal(a)kernel.org>
---
.../selftests/ftrace/test.d/filter/event-filter-function.tc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tools/testing/selftests/ftrace/test.d/filter/event-filter-function.tc b/tools/testing/selftests/ftrace/test.d/filter/event-filter-function.tc
index 2de7c61d1ae30..3f74c09c56b62 100644
--- a/tools/testing/selftests/ftrace/test.d/filter/event-filter-function.tc
+++ b/tools/testing/selftests/ftrace/test.d/filter/event-filter-function.tc
@@ -24,7 +24,7 @@ echo 0 > events/enable
echo "Get the most frequently calling function"
sample_events
-target_func=`cut -d: -f3 trace | sed 's/call_site=\([^+]*\)+0x.*/\1/' | sort | uniq -c | sort | tail -n 1 | sed 's/^[ 0-9]*//'`
+target_func=`cat trace | grep -o 'call_site=\([^+]*\)' | sed 's/call_site=//' | sort | uniq -c | sort | tail -n 1 | sed 's/^[ 0-9]*//'`
if [ -z "$target_func" ]; then
exit_fail
fi
--
2.43.0
After this change the single SAN device (ns3eth1) is now replaced with
two SAN devices - respectively ns4eth1 and ns5eth1.
It is possible to extend this script to have more SAN devices connected
by adding them to ns3br1 bridge.
Signed-off-by: Lukasz Majewski <lukma(a)denx.de>
---
tools/testing/selftests/net/hsr/hsr_redbox.sh | 71 +++++++++++++------
1 file changed, 49 insertions(+), 22 deletions(-)
diff --git a/tools/testing/selftests/net/hsr/hsr_redbox.sh b/tools/testing/selftests/net/hsr/hsr_redbox.sh
index db69be95ecb3..1f36785347c0 100755
--- a/tools/testing/selftests/net/hsr/hsr_redbox.sh
+++ b/tools/testing/selftests/net/hsr/hsr_redbox.sh
@@ -8,12 +8,19 @@ source ./hsr_common.sh
do_complete_ping_test()
{
echo "INFO: Initial validation ping (HSR-SAN/RedBox)."
- # Each node has to be able each one.
+ # Each node has to be able to reach each one.
do_ping "${ns1}" 100.64.0.2
do_ping "${ns2}" 100.64.0.1
- # Ping from SAN to hsr1 (via hsr2)
+ # Ping between SANs (test bridge)
+ do_ping "${ns4}" 100.64.0.51
+ do_ping "${ns5}" 100.64.0.41
+ # Ping from SANs to hsr1 (via hsr2) (and opposite)
do_ping "${ns3}" 100.64.0.1
do_ping "${ns1}" 100.64.0.3
+ do_ping "${ns1}" 100.64.0.41
+ do_ping "${ns4}" 100.64.0.1
+ do_ping "${ns1}" 100.64.0.51
+ do_ping "${ns5}" 100.64.0.1
stop_if_error "Initial validation failed."
# Wait for MGNT HSR frames being received and nodes being
@@ -23,8 +30,12 @@ do_complete_ping_test()
echo "INFO: Longer ping test (HSR-SAN/RedBox)."
# Ping from SAN to hsr1 (via hsr2)
do_ping_long "${ns3}" 100.64.0.1
- # Ping from hsr1 (via hsr2) to SAN
+ # Ping from hsr1 (via hsr2) to SANs (and opposite)
do_ping_long "${ns1}" 100.64.0.3
+ do_ping_long "${ns1}" 100.64.0.41
+ do_ping_long "${ns4}" 100.64.0.1
+ do_ping_long "${ns1}" 100.64.0.51
+ do_ping_long "${ns5}" 100.64.0.1
stop_if_error "Longer ping test failed."
echo "INFO: All good."
@@ -35,22 +46,26 @@ setup_hsr_interfaces()
local HSRv="$1"
echo "INFO: preparing interfaces for HSRv${HSRv} (HSR-SAN/RedBox)."
-
-# |NS1 |
-# | |
-# | /-- hsr1 --\ |
-# | ns1eth1 ns1eth2 |
-# |------------------------|
-# | |
-# | |
-# | |
-# |------------------------| |-----------|
-# | ns2eth1 ns2eth2 | | |
-# | \-- hsr2 --/ | | |
-# | \ | | |
-# | ns2eth3 |--------| ns3eth1 |
-# | (interlink)| | |
-# |NS2 (RedBOX) | |NS3 (SAN) |
+#
+# IPv4 addresses (100.64.X.Y/24), and [X.Y] is presented on below diagram:
+#
+#
+# |NS1 | |NS4 |
+# | [0.1] | | |
+# | /-- hsr1 --\ | | [0.41] |
+# | ns1eth1 ns1eth2 | | ns4eth1 (SAN) |
+# |------------------------| |-------------------|
+# | | |
+# | | |
+# | | |
+# |------------------------| |-------------------------------|
+# | ns2eth1 ns2eth2 | | ns3eth2 |
+# | \-- hsr2 --/ | | / |
+# | [0.2] \ | | / | |------------|
+# | ns2eth3 |---| ns3eth1 -- ns3br1 -- ns3eth3--|--| ns5eth1 |
+# | (interlink)| | [0.3] [0.11] | | [0.51] |
+# |NS2 (RedBOX) | |NS3 (BR) | | NS5 (SAN) |
+#
#
# Check if iproute2 supports adding interlink port to hsrX device
ip link help hsr | grep -q INTERLINK
@@ -59,7 +74,9 @@ setup_hsr_interfaces()
# Create interfaces for name spaces
ip link add ns1eth1 netns "${ns1}" type veth peer name ns2eth1 netns "${ns2}"
ip link add ns1eth2 netns "${ns1}" type veth peer name ns2eth2 netns "${ns2}"
- ip link add ns3eth1 netns "${ns3}" type veth peer name ns2eth3 netns "${ns2}"
+ ip link add ns2eth3 netns "${ns2}" type veth peer name ns3eth1 netns "${ns3}"
+ ip link add ns3eth2 netns "${ns3}" type veth peer name ns4eth1 netns "${ns4}"
+ ip link add ns3eth3 netns "${ns3}" type veth peer name ns5eth1 netns "${ns5}"
sleep 1
@@ -70,21 +87,31 @@ setup_hsr_interfaces()
ip -n "${ns2}" link set ns2eth2 up
ip -n "${ns2}" link set ns2eth3 up
- ip -n "${ns3}" link set ns3eth1 up
+ ip -n "${ns3}" link add name ns3br1 type bridge
+ ip -n "${ns3}" link set ns3br1 up
+ ip -n "${ns3}" link set ns3eth1 master ns3br1 up
+ ip -n "${ns3}" link set ns3eth2 master ns3br1 up
+ ip -n "${ns3}" link set ns3eth3 master ns3br1 up
+
+ ip -n "${ns4}" link set ns4eth1 up
+ ip -n "${ns5}" link set ns5eth1 up
ip -net "${ns1}" link add name hsr1 type hsr slave1 ns1eth1 slave2 ns1eth2 supervision 45 version ${HSRv} proto 0
ip -net "${ns2}" link add name hsr2 type hsr slave1 ns2eth1 slave2 ns2eth2 interlink ns2eth3 supervision 45 version ${HSRv} proto 0
ip -n "${ns1}" addr add 100.64.0.1/24 dev hsr1
ip -n "${ns2}" addr add 100.64.0.2/24 dev hsr2
+ ip -n "${ns3}" addr add 100.64.0.11/24 dev ns3br1
ip -n "${ns3}" addr add 100.64.0.3/24 dev ns3eth1
+ ip -n "${ns4}" addr add 100.64.0.41/24 dev ns4eth1
+ ip -n "${ns5}" addr add 100.64.0.51/24 dev ns5eth1
ip -n "${ns1}" link set hsr1 up
ip -n "${ns2}" link set hsr2 up
}
check_prerequisites
-setup_ns ns1 ns2 ns3
+setup_ns ns1 ns2 ns3 ns4 ns5
trap cleanup_all_ns EXIT
--
2.20.1
Joachim kindly merged the IPv6 support in
https://github.com/troglobit/mtools/pull/2, so we can just use his
version now. A few more fixes subsequently came in for IPv6, so even
better.
Check that the deployed mtools version is 3.0 or above. Note that the
version check breaks compatibility with my fork where I didn't bump the
version, but I assume that won't be a problem.
Signed-off-by: Vladimir Oltean <vladimir.oltean(a)nxp.com>
---
tools/testing/selftests/net/forwarding/lib.sh | 19 +++++++++++++++++--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/net/forwarding/lib.sh b/tools/testing/selftests/net/forwarding/lib.sh
index 4fe28ab5d8b9..aa925c0954a5 100644
--- a/tools/testing/selftests/net/forwarding/lib.sh
+++ b/tools/testing/selftests/net/forwarding/lib.sh
@@ -309,6 +309,21 @@ require_command()
fi
}
+# IPv6 support was added in v3.0
+check_mtools_version()
+{
+ local version="$(msend -v)"
+ local major
+
+ version=${version##msend version }
+ major=$(echo $version | cut -d. -f1)
+
+ if [ $major -lt 3 ]; then
+ echo "SKIP: expected mtools version 3.0, got $version"
+ exit $ksft_skip
+ fi
+}
+
if [[ "$REQUIRE_JQ" = "yes" ]]; then
require_command jq
fi
@@ -316,10 +331,10 @@ if [[ "$REQUIRE_MZ" = "yes" ]]; then
require_command $MZ
fi
if [[ "$REQUIRE_MTOOLS" = "yes" ]]; then
- # https://github.com/vladimiroltean/mtools/
- # patched for IPv6 support
+ # https://github.com/troglobit/mtools
require_command msend
require_command mreceive
+ check_mtools_version
fi
##############################################################################
--
2.34.1
Hi Linus,
Without reply from Shuah, and given the importance of these fixes [1], here is
a PR to fix Kselftest (broken since v6.9-rc1) for at least KVM, pidfd, and
Landlock. I cannot test against all kselftests though. This has been in
linux-next since the beginning of this week, and so far only one issue has been
reported [2] and fixed [3].
Feel free to take this PR if you see fit.
Regards,
Mickaël
[1] https://lore.kernel.org/r/Zjo1xyhjmehsRhZ2@google.com
[2] https://lore.kernel.org/r/202405100339.vfBe0t9C-lkp@intel.com
[3] https://lore.kernel.org/r/20240511171445.904356-1-mic@digikod.net
--
The following changes since commit e67572cd2204894179d89bd7b984072f19313b03:
Linux 6.9-rc6 (2024-04-28 13:47:24 -0700)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/mic/linux.git tags/kselftest-fix-vfork-2024-05-12
for you to fetch changes up to 323feb3bdb67649bfa5614eb24ec9cb92a60cf33:
selftests/harness: Handle TEST_F()'s explicit exit codes (2024-05-11 19:18:47 +0200)
----------------------------------------------------------------
Fix Kselftest's vfork() side effects
See https://lore.kernel.org/r/20240511171445.904356-1-mic@digikod.net
----------------------------------------------------------------
Mickaël Salaün (10):
selftests/pidfd: Fix config for pidfd_setns_test
selftests/landlock: Fix FS tests when run on a private mount point
selftests/harness: Fix fixture teardown
selftests/harness: Fix interleaved scheduling leading to race conditions
selftests/landlock: Do not allocate memory in fixture data
selftests/harness: Constify fixture variants
selftests/pidfd: Fix wrong expectation
selftests/harness: Share _metadata between forked processes
selftests/harness: Fix vfork() side effects
selftests/harness: Handle TEST_F()'s explicit exit codes
tools/testing/selftests/kselftest_harness.h | 127 +++++++++++++++++------
tools/testing/selftests/landlock/fs_test.c | 83 +++++++++------
tools/testing/selftests/pidfd/config | 2 +
tools/testing/selftests/pidfd/pidfd_setns_test.c | 2 +-
4 files changed, 147 insertions(+), 67 deletions(-)
The selftest/sgx/main.c didn't compile with [-Werror=implicit-function-declaration]
[edited]:
make[3]: Entering directory 'tools/testing/selftests/sgx'
gcc -Wall -Werror -g -Itools/testing/selftests/../../../tools/include -fPIC -c main.c \
-o tools/testing/selftests/sgx/main.o
In file included from main.c:21:
../kselftest_harness.h: In function ‘__run_test’:
../kselftest_harness.h:1169:13: error: implicit declaration of function ‘asprintf’; \
did you mean ‘vsprintf’? [-Werror=implicit-function-declaration]
1169 | if (asprintf(&test_name, "%s%s%s.%s", f->name,
| ^~~~~~~~
| vsprintf
cc1: all warnings being treated as errors
make[3]: *** [Makefile:36: tools/testing/selftests/sgx/main.o] Error 1
The cause is in the included <stdio.h> on Ubuntu 22.04 LTS:
19 /*
20 * ISO C99 Standard: 7.19 Input/output <stdio.h>
21 */
.
.
.
387 #if __GLIBC_USE (LIB_EXT2)
388 /* Write formatted output to a string dynamically allocated with `malloc'.
389 Store the address of the string in *PTR. */
390 extern int vasprintf (char **__restrict __ptr, const char *__restrict __f,
391 __gnuc_va_list __arg)
392 __THROWNL __attribute__ ((__format__ (__printf__, 2, 0))) __wur;
393 extern int __asprintf (char **__restrict __ptr,
394 const char *__restrict __fmt, ...)
395 __THROWNL __attribute__ ((__format__ (__printf__, 2, 3))) __wur;
396 extern int asprintf (char **__restrict __ptr,
397 const char *__restrict __fmt, ...)
398 __THROWNL __attribute__ ((__format__ (__printf__, 2, 3))) __wur;
399 #endif
__GLIBC_USE (LIB_EXT2) expands into __GLIBC_USE_LIB_EXT2 as defined here:
/usr/include/features.h:186:#define __GLIBC_USE(F) __GLIBC_USE_ ## F
Now, what is unobvious is that <stdio.h> includes
/usr/include/x86_64-linux-gnu/bits/libc-header-start.h:
------------------------------------------------------
35 /* ISO/IEC TR 24731-2:2010 defines the __STDC_WANT_LIB_EXT2__
36 macro. */
37 #undef __GLIBC_USE_LIB_EXT2
38 #if (defined __USE_GNU \
39 || (defined __STDC_WANT_LIB_EXT2__ && __STDC_WANT_LIB_EXT2__ > 0))
40 # define __GLIBC_USE_LIB_EXT2 1
41 #else
42 # define __GLIBC_USE_LIB_EXT2 0
43 #endif
This makes <stdio.h> exclude line 396 and asprintf() prototype from normal
include file processing.
The fix defines __USE_GNU before including <stdio.h> in case it isn't already
defined. After this intervention the module compiles OK.
Converting snprintf() to asprintf() in selftests/kselftest_harness.h:1169
created this new dependency and the implicit declaration broke the compilation.
Fixes: 809216233555 ("selftests/harness: remove use of LINE_MAX")
Cc: Edward Liaw <edliaw(a)google.com>
Cc: Jarkko Sakkinen <jarkko(a)kernel.org>
Cc: Dave Hansen <dave.hansen(a)linux.intel.com>
Cc: Shuah Khan <shuah(a)kernel.org>
Cc: linux-sgx(a)vger.kernel.org
Cc: linux-kselftest(a)vger.kernel.org
Cc: linux-kernel(a)vger.kernel.org
Signed-off-by: Mirsad Todorovac <mtodorov69(a)gmail.com>
---
tools/testing/selftests/sgx/main.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/tools/testing/selftests/sgx/main.c b/tools/testing/selftests/sgx/main.c
index 9820b3809c69..f5cb426bd797 100644
--- a/tools/testing/selftests/sgx/main.c
+++ b/tools/testing/selftests/sgx/main.c
@@ -6,6 +6,9 @@
#include <errno.h>
#include <fcntl.h>
#include <stdbool.h>
+#ifndef __USE_GNU
+#define __USE_GNU
+#endif
#include <stdio.h>
#include <stdint.h>
#include <stdlib.h>
--
2.34.1
Currently, if at runtime we are not able to allocate a huge page, the
test will trivially pass on Aarch64 due to no exception being raised on
division by zero while computing compaction_index. Fix that by checking
for nr_hugepages == 0. Anyways, in general, avoid a division by zero by
exiting the program beforehand. While at it, fix a typo.
Signed-off-by: Dev Jain <dev.jain(a)arm.com>
---
tools/testing/selftests/mm/compaction_test.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/tools/testing/selftests/mm/compaction_test.c b/tools/testing/selftests/mm/compaction_test.c
index 533999b6c284..df1b76f9c734 100644
--- a/tools/testing/selftests/mm/compaction_test.c
+++ b/tools/testing/selftests/mm/compaction_test.c
@@ -134,6 +134,10 @@ int check_compaction(unsigned long mem_free, unsigned int hugepage_size)
/* We should have been able to request at least 1/3 rd of the memory in
huge pages */
+ if (!atoi(nr_hugepages)) {
+ ksft_print_msg("ERROR: No memory is available as huge pages\n");
+ goto close_fd;
+ }
compaction_index = mem_free/(atoi(nr_hugepages) * hugepage_size);
lseek(fd, 0, SEEK_SET);
@@ -149,7 +153,7 @@ int check_compaction(unsigned long mem_free, unsigned int hugepage_size)
atoi(nr_hugepages));
if (compaction_index > 3) {
- ksft_print_msg("ERROR: Less that 1/%d of memory is available\n"
+ ksft_print_msg("ERROR: Less than 1/%d of memory is available\n"
"as huge pages\n", compaction_index);
goto close_fd;
}
--
2.39.2
From: Kunwu Chan <chentao(a)kylinos.cn>
The "malloc" call may not be successful.Add the malloc
failure checking to avoid possible null dereference.
Changes since v1 [1]:
- As Daniel Borkmann suggested, change patch 4/4 only
- Other 3 patches no changes.
[1] https://lore.kernel.org/all/20240424020444.2375773-1-chentao@kylinos.cn/
Kunwu Chan (4):
selftests/bpf: Add some null pointer checks
selftests/bpf/sockopt: Add a null pointer check for the run_test
selftests/bpf: Add a null pointer check for the load_btf_spec
selftests/bpf: Add a null pointer check for the
serial_test_tp_attach_query
tools/testing/selftests/bpf/prog_tests/sockopt.c | 6 ++++++
tools/testing/selftests/bpf/prog_tests/tp_attach_query.c | 3 +++
tools/testing/selftests/bpf/test_progs.c | 7 +++++++
tools/testing/selftests/bpf/test_verifier.c | 2 ++
4 files changed, 18 insertions(+)
--
2.40.1
From: Geliang Tang <tanggeliang(a)kylinos.cn>
The strdup() function returns a pointer to a new string which is a
duplicate of the string "ifname". Memory for the new string is obtained
with malloc(), and need to be freed with free().
This patch adds this missing "free(saved_hwtstamp_ifname)" in cleanup()
to avoid a potential memory leak in xdp_hw_metadata.c.
Signed-off-by: Geliang Tang <tanggeliang(a)kylinos.cn>
---
tools/testing/selftests/bpf/xdp_hw_metadata.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/tools/testing/selftests/bpf/xdp_hw_metadata.c b/tools/testing/selftests/bpf/xdp_hw_metadata.c
index 0859fe727da7..6f9956eed797 100644
--- a/tools/testing/selftests/bpf/xdp_hw_metadata.c
+++ b/tools/testing/selftests/bpf/xdp_hw_metadata.c
@@ -581,6 +581,8 @@ static void cleanup(void)
if (bpf_obj)
xdp_hw_metadata__destroy(bpf_obj);
+
+ free((void *)saved_hwtstamp_ifname);
}
static void handle_signal(int sig)
--
2.43.0
In this series from Geliang, modifying MPTCP BPF selftests, we have:
- SIGINT support
- A new macro to reduce duplicated code
- A new MPTCP subflow BPF program setting socket options per subflow: it
looks better to have this old test program in the BPF selftests to
track regressions and to serve as example.
Note: Nicolas is no longer working for Tessares, but he did this work
while working for them, and his email address is no longer available.
- A new MPTCP BPF subtest validating the new BPF program.
Signed-off-by: Matthieu Baerts (NGI0) <matttbe(a)kernel.org>
---
Geliang Tang (3):
selftests/bpf: Handle SIGINT when creating netns
selftests/bpf: Add RUN_MPTCP_TEST macro
selftests/bpf: Add mptcp subflow subtest
Nicolas Rybowski (1):
selftests/bpf: Add mptcp subflow example
tools/testing/selftests/bpf/prog_tests/mptcp.c | 127 +++++++++++++++++++++-
tools/testing/selftests/bpf/progs/mptcp_subflow.c | 70 ++++++++++++
2 files changed, 193 insertions(+), 4 deletions(-)
---
base-commit: 329a6720a3ebbc041983b267981ab2cac102de93
change-id: 20240506-upstream-bpf-next-20240506-mptcp-subflow-test-faef6654bfa3
Best regards,
--
Matthieu Baerts (NGI0) <matttbe(a)kernel.org>
On slow machines the SND timestamp sometimes doesn't arrive before
we quit. The test only waits as long as the packet delay, so it's
easy for a race condition to happen.
Double the wait but do a bit of polling, once the SND timestamp
arrives there's no point to wait any longer.
This fixes the "TXTIME abs" failures on debug kernels, like:
Case ICMPv4 - TXTIME abs returned '', expected 'OK'
Signed-off-by: Jakub Kicinski <kuba(a)kernel.org>
---
CC: shuah(a)kernel.org
CC: linux-kselftest(a)vger.kernel.org
---
tools/testing/selftests/net/cmsg_sender.c | 20 +++++++++++++++-----
1 file changed, 15 insertions(+), 5 deletions(-)
diff --git a/tools/testing/selftests/net/cmsg_sender.c b/tools/testing/selftests/net/cmsg_sender.c
index c79e65581dc3..f25268504937 100644
--- a/tools/testing/selftests/net/cmsg_sender.c
+++ b/tools/testing/selftests/net/cmsg_sender.c
@@ -333,16 +333,17 @@ static const char *cs_ts_info2str(unsigned int info)
return "unknown";
}
-static void
+static unsigned long
cs_read_cmsg(int fd, struct msghdr *msg, char *cbuf, size_t cbuf_sz)
{
struct sock_extended_err *see;
struct scm_timestamping *ts;
+ unsigned int ts_seen = 0;
struct cmsghdr *cmsg;
int i, err;
if (!opt.ts.ena)
- return;
+ return 0;
msg->msg_control = cbuf;
msg->msg_controllen = cbuf_sz;
@@ -396,8 +397,11 @@ cs_read_cmsg(int fd, struct msghdr *msg, char *cbuf, size_t cbuf_sz)
printf(" %5s ts%d %lluus\n",
cs_ts_info2str(see->ee_info),
i, rel_time);
+ ts_seen |= 1 << see->ee_info;
}
}
+
+ return ts_seen;
}
static void ca_set_sockopts(int fd)
@@ -509,10 +513,16 @@ int main(int argc, char *argv[])
err = ERN_SUCCESS;
if (opt.ts.ena) {
- /* Make sure all timestamps have time to loop back */
- usleep(opt.txtime.delay);
+ unsigned long seen;
+ int i;
- cs_read_cmsg(fd, &msg, cbuf, sizeof(cbuf));
+ /* Make sure all timestamps have time to loop back */
+ for (i = 0; i < 40; i++) {
+ seen = cs_read_cmsg(fd, &msg, cbuf, sizeof(cbuf));
+ if (seen & (1 << SCM_TSTAMP_SND))
+ break;
+ usleep(opt.txtime.delay / 20);
+ }
}
err_out:
--
2.45.0
The selftest/sgx/main.c didn't compile with [-Werror=implicit-function-declaration]
[edited]:
make[3]: Entering directory 'tools/testing/selftests/sgx'
gcc -Wall -Werror -g -Itools/testing/selftests/../../../tools/include -fPIC -c main.c \
-o tools/testing/selftests/sgx/main.o
In file included from main.c:21:
../kselftest_harness.h: In function ‘__run_test’:
../kselftest_harness.h:1169:13: error: implicit declaration of function ‘asprintf’; \
did you mean ‘vsprintf’? [-Werror=implicit-function-declaration]
1169 | if (asprintf(&test_name, "%s%s%s.%s", f->name,
| ^~~~~~~~
| vsprintf
cc1: all warnings being treated as errors
make[3]: *** [Makefile:36: tools/testing/selftests/sgx/main.o] Error 1
The cause is in the included <stdio.h> on Ubuntu 22.04 LTS:
19 /*
20 * ISO C99 Standard: 7.19 Input/output <stdio.h>
21 */
.
.
.
387 #if __GLIBC_USE (LIB_EXT2)
388 /* Write formatted output to a string dynamically allocated with `malloc'.
389 Store the address of the string in *PTR. */
390 extern int vasprintf (char **__restrict __ptr, const char *__restrict __f,
391 __gnuc_va_list __arg)
392 __THROWNL __attribute__ ((__format__ (__printf__, 2, 0))) __wur;
393 extern int __asprintf (char **__restrict __ptr,
394 const char *__restrict __fmt, ...)
395 __THROWNL __attribute__ ((__format__ (__printf__, 2, 3))) __wur;
396 extern int asprintf (char **__restrict __ptr,
397 const char *__restrict __fmt, ...)
398 __THROWNL __attribute__ ((__format__ (__printf__, 2, 3))) __wur;
399 #endif
__GLIBC_USE (LIB_EXT2) expands into __GLIBC_USE_LIB_EXT2 as defined here:
/usr/include/features.h:186:#define __GLIBC_USE(F) __GLIBC_USE_ ## F
Now, what is unobvious is that <stdio.h> includes
/usr/include/x86_64-linux-gnu/bits/libc-header-start.h:
------------------------------------------------------
35 /* ISO/IEC TR 24731-2:2010 defines the __STDC_WANT_LIB_EXT2__
36 macro. */
37 #undef __GLIBC_USE_LIB_EXT2
38 #if (defined __USE_GNU \
39 || (defined __STDC_WANT_LIB_EXT2__ && __STDC_WANT_LIB_EXT2__ > 0))
40 # define __GLIBC_USE_LIB_EXT2 1
41 #else
42 # define __GLIBC_USE_LIB_EXT2 0
43 #endif
This makes <stdio.h> exclude line 396 and asprintf() prototype from normal
include file processing.
The fix defines __USE_GNU before including <stdio.h> in case it isn't already
defined. After this intervention the module compiles OK.
Converting snprintf() to asprintf() in selftests/kselftest_harness.h:1169
created this new dependency and the implicit declaration broke the compilation.
Fixes: 809216233555 ("selftests/harness: remove use of LINE_MAX")
Cc: Edward Liaw <edliaw(a)google.com>
Cc: Jarkko Sakkinen <jarkko(a)kernel.org>
Cc: Dave Hansen <dave.hansen(a)linux.intel.com>
Cc: Shuah Khan <shuah(a)kernel.org>
Cc: linux-sgx(a)vger.kernel.org
Cc: linux-kselftest(a)vger.kernel.org
Cc: linux-kernel(a)vger.kernel.org
Signed-off-by: Mirsad Todorovac <mtodorov69(a)gmail.com>
---
tools/testing/selftests/sgx/main.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/tools/testing/selftests/sgx/main.c b/tools/testing/selftests/sgx/main.c
index 9820b3809c69..f5cb426bd797 100644
--- a/tools/testing/selftests/sgx/main.c
+++ b/tools/testing/selftests/sgx/main.c
@@ -6,6 +6,9 @@
#include <errno.h>
#include <fcntl.h>
#include <stdbool.h>
+#ifndef __USE_GNU
+#define __USE_GNU
+#endif
#include <stdio.h>
#include <stdint.h>
#include <stdlib.h>
--
2.34.1
The cb fields network_offset and inner_network_offset are used instead of
skb->network_header throughout GRO.
These fields are then leveraged in the next commit to remove flush_id state
from napi_gro_cb, and stateful code in {ipv6,inet}_gro_receive which may be
unnecessarily complicated due to encapsulation support in GRO. These fields
are checked in L4 instead.
3rd patch adds tests for different flush_id flows in GRO.
v8 -> v9:
- rename skb_gro_network_offset to skb_gro_receive_network_offset for
clarification
- improved code readability in tests and gro_network_flush functions
- v8:
https://lore.kernel.org/all/20240506093550.128210-1-richardbgobert@gmail.co…
v7 -> v8:
- Remove network_header use in gro
- Re-send commits after the dependent patch to net was applied
- v7:
https://lore.kernel.org/all/20240412155533.115507-1-richardbgobert@gmail.co…
v6 -> v7:
- Moved bug fixes to a separate submission in net
- Added UDP fwd benchmark
- v6:
https://lore.kernel.org/all/20240410153423.107381-1-richardbgobert@gmail.co…
v5 -> v6:
- Write inner_network_offset in vxlan and geneve
- Ignore is_atomic when DF=0
- v5:
https://lore.kernel.org/all/20240408141720.98832-1-richardbgobert@gmail.com/
v4 -> v5:
- Add 1st commit - flush id checks in udp_gro_receive segment which can be
backported by itself
- Add TCP measurements for the 5th commit
- Add flush id tests to ensure flush id logic is preserved in GRO
- Simplify gro_inet_flush by removing a branch
- v4:
https://lore.kernel.org/all/202420325182543.87683-1-richardbgobert@gmail.co…
v3 -> v4:
- Fix code comment and commit message typos
- v3:
https://lore.kernel.org/all/f939c84a-2322-4393-a5b0-9b1e0be8ed8e@gmail.com/
v2 -> v3:
- Use napi_gro_cb instead of skb->{offset}
- v2:
https://lore.kernel.org/all/2ce1600b-e733-448b-91ac-9d0ae2b866a4@gmail.com/
v1 -> v2:
- Pass p_off in *_gro_complete to fix UDP bug
- Remove more conditionals and memory fetches from inet_gro_flush
- v1:
https://lore.kernel.org/netdev/e1d22505-c5f8-4c02-a997-64248480338b@gmail.c…
Richard Gobert (3):
net: gro: use cb instead of skb->network_header
net: gro: move L3 flush checks to tcp_gro_receive and udp_gro_receive_segment
selftests/net: add flush id selftests
include/net/gro.h | 87 ++++++++++++++++---
net/core/gro.c | 3 -
net/ipv4/af_inet.c | 45 +---------
net/ipv4/tcp_offload.c | 20 ++---
net/ipv4/udp_offload.c | 10 +--
net/ipv6/ip6_offload.c | 16 +---
net/ipv6/tcpv6_offload.c | 3 +-
tools/testing/selftests/net/gro.c | 138 ++++++++++++++++++++++++++++++
8 files changed, 227 insertions(+), 95 deletions(-)
--
2.36.1
When building with clang, via:
make LLVM=1 -C tools/testing/selftests
...two types of warnings occur:
warning: absolute value function 'abs' given an argument of type
'long' but has parameter of type 'int' which may cause truncation of
value
warning: taking the absolute value of unsigned type 'unsigned long'
has no effect
Fix these by:
a) using labs() in place of abs(), when long integers are involved, and
b) Change to use signed integer data types, in places where subtraction
is used (and could end up with negative values).
c) Remove a duplicate abs() call in cmt_test.c.
Cc: Reinette Chatre <reinette.chatre(a)intel.com>
Cc: Ilpo Järvinen <ilpo.jarvinen(a)linux.intel.com>
Signed-off-by: John Hubbard <jhubbard(a)nvidia.com>
---
tools/testing/selftests/resctrl/cmt_test.c | 4 ++--
tools/testing/selftests/resctrl/mba_test.c | 2 +-
tools/testing/selftests/resctrl/mbm_test.c | 2 +-
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/tools/testing/selftests/resctrl/cmt_test.c b/tools/testing/selftests/resctrl/cmt_test.c
index a81f91222a89..05a241519ae8 100644
--- a/tools/testing/selftests/resctrl/cmt_test.c
+++ b/tools/testing/selftests/resctrl/cmt_test.c
@@ -40,11 +40,11 @@ static int show_results_info(unsigned long sum_llc_val, int no_of_bits,
int ret;
avg_llc_val = sum_llc_val / num_of_runs;
- avg_diff = (long)abs(cache_span - avg_llc_val);
+ avg_diff = (long)(cache_span - avg_llc_val);
diff_percent = ((float)cache_span - avg_llc_val) / cache_span * 100;
ret = platform && abs((int)diff_percent) > max_diff_percent &&
- abs(avg_diff) > max_diff;
+ labs(avg_diff) > max_diff;
ksft_print_msg("%s Check cache miss rate within %lu%%\n",
ret ? "Fail:" : "Pass:", max_diff_percent);
diff --git a/tools/testing/selftests/resctrl/mba_test.c b/tools/testing/selftests/resctrl/mba_test.c
index 7946e32e85c8..5fffbc9ff6a4 100644
--- a/tools/testing/selftests/resctrl/mba_test.c
+++ b/tools/testing/selftests/resctrl/mba_test.c
@@ -60,7 +60,7 @@ static bool show_mba_info(unsigned long *bw_imc, unsigned long *bw_resc)
/* Memory bandwidth from 100% down to 10% */
for (allocation = 0; allocation < ALLOCATION_MAX / ALLOCATION_STEP;
allocation++) {
- unsigned long avg_bw_imc, avg_bw_resc;
+ long avg_bw_imc, avg_bw_resc;
unsigned long sum_bw_imc = 0, sum_bw_resc = 0;
int avg_diff_per;
float avg_diff;
diff --git a/tools/testing/selftests/resctrl/mbm_test.c b/tools/testing/selftests/resctrl/mbm_test.c
index d67ffa3ec63a..a4c3ea49b0e8 100644
--- a/tools/testing/selftests/resctrl/mbm_test.c
+++ b/tools/testing/selftests/resctrl/mbm_test.c
@@ -17,7 +17,7 @@
static int
show_bw_info(unsigned long *bw_imc, unsigned long *bw_resc, size_t span)
{
- unsigned long avg_bw_imc = 0, avg_bw_resc = 0;
+ long avg_bw_imc = 0, avg_bw_resc = 0;
unsigned long sum_bw_imc = 0, sum_bw_resc = 0;
int runs, ret, avg_diff_per;
float avg_diff = 0;
base-commit: 45db3ab70092637967967bfd8e6144017638563c
prerequisite-patch-id: b901ece2a5b78503e2fb5480f20e304d36a0ea27
prerequisite-patch-id: 8d96c4b8c3ed6d9ea2588ef7f594ae0f9f83c279
--
2.45.0
Since kselftest_harness.h introduces asprintf()[1], many selftests have
compilation warnings or errors due to missing _GNU_SOURCE definitions.
The issue stems from a lack of a LINE_MAX definition in Android (see
commit 38c957f07038), which is the reason why asprintf() was introduced.
We tried adding _GNU_SOURCE definitions to more selftests to fix, but
asprintf() may continue to cause problems, and since it is quite late in
the 6.9 cycle, we would like to revert 809216233555 first to provide
testing for forks[2].
[1] https://lore.kernel.org/all/20240411231954.62156-1-edliaw@google.com
[2] https://lore.kernel.org/linux-kselftest/ZjuA3aY_iHkjP7bQ@google.com
v1 -> v2:
- Stop defining _GNU_SOURCE in related selftests
- Revert commit 809216233555
- Use 1024 in place of LINE_MAX to fix 38c957f07038
v1: https://lore.kernel.org/all/20240507063534.4191447-1-tao1.su@linux.intel.co…
Tao Su (2):
Revert "selftests/harness: remove use of LINE_MAX"
selftests/harness: Use 1024 in place of LINE_MAX
tools/testing/selftests/kselftest_harness.h | 11 +++--------
tools/testing/selftests/mm/mdwe_test.c | 1 -
2 files changed, 3 insertions(+), 9 deletions(-)
base-commit: 45db3ab70092637967967bfd8e6144017638563c
--
2.34.1
When building with clang, via:
make LLVM=1 -C tools/testing/selftest
...clang warns about several cases of using a signed integer for the
priority argument to mq_receive(3), which expects an unsigned int.
Fix this by declaring the type as unsigned int in all cases.
Also, both input and output priority are unsigned, per the man pages, so
let's change the type of both priorities throughout, even though clang
did not warn about the prio_out variable.
Also, add an argument name to test->func(), in order to address another
warning from clang.
Cc: Ryan Roberts <ryan.roberts(a)arm.com>
Signed-off-by: John Hubbard <jhubbard(a)nvidia.com>
---
tools/testing/selftests/mqueue/mq_perf_tests.c | 15 ++++++++-------
1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/tools/testing/selftests/mqueue/mq_perf_tests.c b/tools/testing/selftests/mqueue/mq_perf_tests.c
index 5c16159d0bcd..9380c656581f 100644
--- a/tools/testing/selftests/mqueue/mq_perf_tests.c
+++ b/tools/testing/selftests/mqueue/mq_perf_tests.c
@@ -323,7 +323,8 @@ void *fake_cont_thread(void *arg)
void *cont_thread(void *arg)
{
char buff[MSG_SIZE];
- int i, priority;
+ int i;
+ unsigned int priority;
for (i = 0; i < num_cpus_to_pin; i++)
if (cpu_threads[i] == pthread_self())
@@ -373,27 +374,27 @@ void *cont_thread(void *arg)
struct test {
char *desc;
- void (*func)(int *);
+ void (*func)(unsigned int *prio);
};
-void const_prio(int *prio)
+void const_prio(unsigned int *prio)
{
return;
}
-void inc_prio(int *prio)
+void inc_prio(unsigned int *prio)
{
if (++*prio == mq_prio_max)
*prio = 0;
}
-void dec_prio(int *prio)
+void dec_prio(unsigned int *prio)
{
if (--*prio < 0)
*prio = mq_prio_max - 1;
}
-void random_prio(int *prio)
+void random_prio(unsigned int *prio)
{
*prio = random() % mq_prio_max;
}
@@ -425,7 +426,7 @@ struct test test2[] = {
void *perf_test_thread(void *arg)
{
char buff[MSG_SIZE];
- int prio_out, prio_in;
+ unsigned int prio_out, prio_in;
int i;
clockid_t clock;
pthread_t *t;
base-commit: 45db3ab70092637967967bfd8e6144017638563c
prerequisite-patch-id: b901ece2a5b78503e2fb5480f20e304d36a0ea27
--
2.45.0
There is a 'malloc' call in test_vmx_nested_state function, which can
be unsuccessful. This patch will add the malloc failure checking
to avoid possible null dereference and give more information
about test fail reasons.
Signed-off-by: Kunwu Chan <chentao(a)kylinos.cn>
---
tools/testing/selftests/kvm/x86_64/vmx_set_nested_state_test.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/tools/testing/selftests/kvm/x86_64/vmx_set_nested_state_test.c b/tools/testing/selftests/kvm/x86_64/vmx_set_nested_state_test.c
index 67a62a5a8895..18afc2000a74 100644
--- a/tools/testing/selftests/kvm/x86_64/vmx_set_nested_state_test.c
+++ b/tools/testing/selftests/kvm/x86_64/vmx_set_nested_state_test.c
@@ -91,6 +91,7 @@ void test_vmx_nested_state(struct kvm_vcpu *vcpu)
const int state_sz = sizeof(struct kvm_nested_state) + getpagesize();
struct kvm_nested_state *state =
(struct kvm_nested_state *)malloc(state_sz);
+ TEST_ASSERT(state, "-ENOMEM when allocating kvm state");
/* The format must be set to 0. 0 for VMX, 1 for SVM. */
set_default_vmx_state(state, state_sz);
--
2.40.1
The "malloc" call may not be successful.Add the malloc
failure checking to avoid possible null dereference.
Kunwu Chan (4):
selftests/bpf: Add some null pointer checks
selftests/bpf/sockopt: Add a null pointer check for the run_test
selftests/bpf: Add a null pointer check for the load_btf_spec
selftests/bpf: Add a null pointer check for the
serial_test_tp_attach_query
tools/testing/selftests/bpf/prog_tests/sockopt.c | 6 ++++++
tools/testing/selftests/bpf/prog_tests/tp_attach_query.c | 3 +++
tools/testing/selftests/bpf/test_progs.c | 7 +++++++
tools/testing/selftests/bpf/test_verifier.c | 2 ++
4 files changed, 18 insertions(+)
--
2.40.1
The series composes of two parts. The first part provides a quick fix for
the issue on a recent thread[1]. The issue happens when a platform has
ununified vector register length across multiple cores. Specifically,
patch 1 adds a comment at a callsite of riscv_setup_vsize to clarify how
vlenb is observed by the system. Patch 2 fixes the issue by failing the
boot process of a secondary core if vlenb mismatches.
The second part of the series provide a finer grain view of the Vector
extension. Patch 3 give the obsolete ISA parser the ability to expand
ISA extensions for sigle letter extensions. Patch 3, 4 introduces Zve32x,
Zve32f, Zve64x, Zve64f, Zve64d for isa parsing and hwprobe. Patch 5
updates all callsites such that Vector subextensions are maximumly
supported by the kernel.
Two parts of the series are sent together to ease the effort of picking
dependency patches. The first part can be merged independent of the
second one if necessary.
The series is tested on a QEMU and verified that booting, Vector
programs context-switch, signal, ptrace, prctl interfaces works when we
only report partial V from the ISA.
Note that the signal test was performed after applying the commit
c27fa53b858b ("riscv: Fix vector state restore in rt_sigreturn()")
This patch should be able to apply on risc-v for-next branch on top of
the commit ba5ea59f768f ("riscv: Do not save the scratch CSR during suspend")
[1]: https://lore.kernel.org/all/20240228-vicinity-cornstalk-4b8eb5fe5730@spud/T…
Changes in v4:
- Add a patch to trigger prctl test on ZVE32X (9)
- Add a patch to fix integer promotion bug in hwprobe (8)
- Fix a build fail on !CONFIG_RISCV_ISA_V (7)
- Add more comment in the assembly code change (2)
- Link to v3: https://lore.kernel.org/r/20240318-zve-detection-v3-0-e12d42107fa8@sifive.c…
Changelog v3:
- Include correct maintainers and mailing list into CC.
- Cleanup isa string parser code (3)
- Adjust extensions order and name (4, 5)
- Refine commit message (6)
Changelog v2:
- Update comments and commit messages (1, 2, 7)
- Refine isa_exts[] lists for zve extensions (4)
- Add a patch for dt-binding (5)
- Make ZVE* extensions depend on has_vector(ZVE32X) (6, 7)
---
---
Andy Chiu (9):
riscv: vector: add a comment when calling riscv_setup_vsize()
riscv: smp: fail booting up smp if inconsistent vlen is detected
riscv: cpufeature: call match_isa_ext() for single-letter extensions
riscv: cpufeature: add zve32[xf] and zve64[xfd] isa detection
dt-bindings: riscv: add Zve32[xf] Zve64[xfd] ISA extension description
riscv: hwprobe: add zve Vector subextensions into hwprobe interface
riscv: vector: adjust minimum Vector requirement to ZVE32X
hwprobe: fix integer promotion in RISCV_HWPROBE_EXT macro
selftest: run vector prctl test for ZVE32X
Documentation/arch/riscv/hwprobe.rst | 15 ++++++
.../devicetree/bindings/riscv/extensions.yaml | 30 ++++++++++++
arch/riscv/include/asm/hwcap.h | 5 ++
arch/riscv/include/asm/switch_to.h | 2 +-
arch/riscv/include/asm/vector.h | 25 ++++++----
arch/riscv/include/asm/xor.h | 2 +-
arch/riscv/include/uapi/asm/hwprobe.h | 7 ++-
arch/riscv/kernel/cpufeature.c | 56 ++++++++++++++++++----
arch/riscv/kernel/head.S | 19 +++++---
arch/riscv/kernel/kernel_mode_vector.c | 4 +-
arch/riscv/kernel/process.c | 4 +-
arch/riscv/kernel/signal.c | 6 +--
arch/riscv/kernel/smpboot.c | 14 ++++--
arch/riscv/kernel/sys_hwprobe.c | 13 ++++-
arch/riscv/kernel/vector.c | 15 +++---
arch/riscv/lib/uaccess.S | 2 +-
.../testing/selftests/riscv/vector/vstate_prctl.c | 6 +--
17 files changed, 174 insertions(+), 51 deletions(-)
---
base-commit: ba5ea59f768f67d127b319b26ba209ff67e0d9a5
change-id: 20240318-zve-detection-50106d2da527
Best regards,
--
Andy Chiu <andy.chiu(a)sifive.com>
From: Geliang Tang <tanggeliang(a)kylinos.cn>
This patchset adds post_socket_cb pointer together with 'struct
post_socket_opts cb_opts' into struct network_helper_opts to make
start_server_addr() helper more flexible. With these modifications,
many duplicate codes can be dropped.
Patches 1-3 address Martin's comments in the previous series.
Geliang Tang (6):
selftests/bpf: Add post_socket_cb for network_helper_opts
selftests/bpf: Use start_server_addr in sockopt_inherit
selftests/bpf: Use start_server_addr in test_tcp_check_syncookie
selftests/bpf: Use connect_to_fd in sockopt_inherit
selftests/bpf: Use connect_to_fd in test_tcp_check_syncookie
selftests/bpf: Drop get_port in test_tcp_check_syncookie
tools/testing/selftests/bpf/Makefile | 1 +
tools/testing/selftests/bpf/network_helpers.c | 25 ++--
tools/testing/selftests/bpf/network_helpers.h | 4 +
.../bpf/prog_tests/sockopt_inherit.c | 63 ++--------
.../bpf/test_tcp_check_syncookie_user.c | 117 ++++--------------
5 files changed, 61 insertions(+), 149 deletions(-)
--
2.43.0
As it is said in the docs, we should always check the KVM API version
before running the KVM-based applications. Add the function which
queries the current KVM API version through `ioctl` to the `kvm_util.c`
file.
Add a new TEST_REQUIRE statement to the `vm_open` function in order
to verify the version of the KVM API before creating a VM.
Signed-off-by: Ivan Orlov <ivan.orlov(a)codethink.co.uk>
---
.../testing/selftests/kvm/include/kvm_util_base.h | 2 ++
tools/testing/selftests/kvm/lib/kvm_util.c | 14 ++++++++++++++
2 files changed, 16 insertions(+)
diff --git a/tools/testing/selftests/kvm/include/kvm_util_base.h b/tools/testing/selftests/kvm/include/kvm_util_base.h
index 3e0db283a46a..d7a83387ae33 100644
--- a/tools/testing/selftests/kvm/include/kvm_util_base.h
+++ b/tools/testing/selftests/kvm/include/kvm_util_base.h
@@ -38,6 +38,7 @@
#define kvm_static_assert(expr, ...) __kvm_static_assert(expr, ##__VA_ARGS__, #expr)
#define KVM_DEV_PATH "/dev/kvm"
+#define KVM_DEFAULT_API_VERSION 12
#define KVM_MAX_VCPUS 512
#define NSEC_PER_SEC 1000000000L
@@ -275,6 +276,7 @@ int get_kvm_intel_param_integer(const char *param);
int get_kvm_amd_param_integer(const char *param);
unsigned int kvm_check_cap(long cap);
+int kvm_get_api_version(void);
static inline bool kvm_has_cap(long cap)
{
diff --git a/tools/testing/selftests/kvm/lib/kvm_util.c b/tools/testing/selftests/kvm/lib/kvm_util.c
index b2262b5fad9e..58a5deccb388 100644
--- a/tools/testing/selftests/kvm/lib/kvm_util.c
+++ b/tools/testing/selftests/kvm/lib/kvm_util.c
@@ -176,6 +176,19 @@ unsigned int kvm_check_cap(long cap)
return (unsigned int)ret;
}
+int kvm_get_api_version(void)
+{
+ int ret;
+ int kvm_fd;
+
+ kvm_fd = open_kvm_dev_path_or_exit();
+ ret = __kvm_ioctl(kvm_fd, KVM_GET_API_VERSION, NULL);
+
+ close(kvm_fd);
+
+ return ret;
+}
+
void vm_enable_dirty_ring(struct kvm_vm *vm, uint32_t ring_size)
{
if (vm_check_cap(vm, KVM_CAP_DIRTY_LOG_RING_ACQ_REL))
@@ -190,6 +203,7 @@ static void vm_open(struct kvm_vm *vm)
vm->kvm_fd = _open_kvm_dev_path_or_exit(O_RDWR);
TEST_REQUIRE(kvm_has_cap(KVM_CAP_IMMEDIATE_EXIT));
+ TEST_REQUIRE(kvm_get_api_version() == KVM_DEFAULT_API_VERSION);
vm->fd = __kvm_ioctl(vm->kvm_fd, KVM_CREATE_VM, (void *)vm->type);
TEST_ASSERT(vm->fd >= 0, KVM_IOCTL_ERROR(KVM_CREATE_VM, vm->fd));
--
2.34.1
The test is inspired by the pmu_event_filter_test which implemented by x86. On
the arm64 platform, there is the same ability to set the pmu_event_filter
through the KVM_ARM_VCPU_PMU_V3_FILTER attribute. So add the test for arm64.
The series first create the helper function which can be used
for the vpmu related tests. Then, it implement the test.
Changelog:
----------
v7->v8:
- Rebased to kvm-arm/next.
- Deleted the GIC layout related staff.
- Fixed the checking logic in the kvm_pmu_support_events.
v6->v7:
- Rebased to v6.9-rc3.
v5->v6:
- Rebased to v6.9-rc1.
- Collect RB.
- Add multiple filter test.
v4->v5:
- Rebased to v6.8-rc6.
- Refactor the helper function, make it fine-grained and easy to be used.
- Namimg improvements.
- Use the kvm_device_attr_set() helper.
- Make the test descriptor array readable and clean.
- Delete the patch which moves the pmu related helper to vpmu.h.
- Remove the kvm_supports_pmu_event_filter() function since nobody will run
this on a old kernel.
v3->v4:
- Rebased to the v6.8-rc2.
v2->v3:
- Check the pmceid in guest code instead of pmu event count since different
hardware may have different event count result, check pmceid makes it stable
on different platform. [Eric]
- Some typo fixed and commit message improved.
v1->v2:
- Improve the commit message. [Eric]
- Fix the bug in [enable|disable]_counter. [Raghavendra & Marc]
- Add the check if kvm has attr KVM_ARM_VCPU_PMU_V3_FILTER.
- Add if host pmu support the test event throught pmceid0.
- Split the test_invalid_filter() to another patch. [Eric]
v1: https://lore.kernel.org/all/20231123063750.2176250-1-shahuang@redhat.com/
v2: https://lore.kernel.org/all/20231129072712.2667337-1-shahuang@redhat.com/
v3: https://lore.kernel.org/all/20240116060129.55473-1-shahuang@redhat.com/
v4: https://lore.kernel.org/all/20240202025659.5065-1-shahuang@redhat.com/
v5: https://lore.kernel.org/all/20240229065625.114207-1-shahuang@redhat.com/
v6: https://lore.kernel.org/all/20240326033706.117189-1-shahuang@redhat.com/
v7: https://lore.kernel.org/all/20240409030320.182591-1-shahuang@redhat.com/
Shaoqin Huang (3):
KVM: selftests: aarch64: Add helper function for the vpmu vcpu
creation
KVM: selftests: aarch64: Introduce pmu_event_filter_test
KVM: selftests: aarch64: Add invalid filter test in
pmu_event_filter_test
tools/testing/selftests/kvm/Makefile | 1 +
.../kvm/aarch64/pmu_event_filter_test.c | 341 ++++++++++++++++++
.../kvm/aarch64/vpmu_counter_access.c | 32 +-
.../selftests/kvm/include/aarch64/vpmu.h | 28 ++
4 files changed, 376 insertions(+), 26 deletions(-)
create mode 100644 tools/testing/selftests/kvm/aarch64/pmu_event_filter_test.c
create mode 100644 tools/testing/selftests/kvm/include/aarch64/vpmu.h
--
2.40.1
The test is inspired by the pmu_event_filter_test which implemented by x86. On
the arm64 platform, there is the same ability to set the pmu_event_filter
through the KVM_ARM_VCPU_PMU_V3_FILTER attribute. So add the test for arm64.
The series first create the helper function which can be used
for the vpmu related tests. Then, it implement the test.
Changelog:
----------
v6->v7:
- Rebased to v6.9-rc3.
v5->v6:
- Rebased to v6.9-rc1.
- Collect RB.
- Add multiple filter test.
v4->v5:
- Rebased to v6.8-rc6.
- Refactor the helper function, make it fine-grained and easy to be used.
- Namimg improvements.
- Use the kvm_device_attr_set() helper.
- Make the test descriptor array readable and clean.
- Delete the patch which moves the pmu related helper to vpmu.h.
- Remove the kvm_supports_pmu_event_filter() function since nobody will run
this on a old kernel.
v3->v4:
- Rebased to the v6.8-rc2.
v2->v3:
- Check the pmceid in guest code instead of pmu event count since different
hardware may have different event count result, check pmceid makes it stable
on different platform. [Eric]
- Some typo fixed and commit message improved.
v1->v2:
- Improve the commit message. [Eric]
- Fix the bug in [enable|disable]_counter. [Raghavendra & Marc]
- Add the check if kvm has attr KVM_ARM_VCPU_PMU_V3_FILTER.
- Add if host pmu support the test event throught pmceid0.
- Split the test_invalid_filter() to another patch. [Eric]
v1: https://lore.kernel.org/all/20231123063750.2176250-1-shahuang@redhat.com/
v2: https://lore.kernel.org/all/20231129072712.2667337-1-shahuang@redhat.com/
v3: https://lore.kernel.org/all/20240116060129.55473-1-shahuang@redhat.com/
v4: https://lore.kernel.org/all/20240202025659.5065-1-shahuang@redhat.com/
v5: https://lore.kernel.org/all/20240229065625.114207-1-shahuang@redhat.com/
v6: https://lore.kernel.org/all/20240326033706.117189-1-shahuang@redhat.com/
Shaoqin Huang (3):
KVM: selftests: aarch64: Add helper function for the vpmu vcpu
creation
KVM: selftests: aarch64: Introduce pmu_event_filter_test
KVM: selftests: aarch64: Add invalid filter test in
pmu_event_filter_test
tools/testing/selftests/kvm/Makefile | 1 +
.../kvm/aarch64/pmu_event_filter_test.c | 336 ++++++++++++++++++
.../kvm/aarch64/vpmu_counter_access.c | 33 +-
.../selftests/kvm/include/aarch64/vpmu.h | 28 ++
4 files changed, 372 insertions(+), 26 deletions(-)
create mode 100644 tools/testing/selftests/kvm/aarch64/pmu_event_filter_test.c
create mode 100644 tools/testing/selftests/kvm/include/aarch64/vpmu.h
--
2.40.1
RFC v8:
=======
Major Changes:
--------------
- Fixed build error generated by patch-by-patch build.
- Applied docs suggestions from Randy.
RFC v7:
=======
Major Changes:
--------------
This revision largely rebases on top of net-next and addresses the feedback
RFCv6 received from folks, namely Jakub, Yunsheng, Arnd, David, & Pavel.
The series remains in RFC because the queue-API ndos defined in this
series are not yet implemented. I have a GVE implementation I carry out
of tree for my testing. A upstreamable GVE implementation is in the
works. Aside from that, in my estimation all the patches are ready for
review/merge. Please do take a look.
As usual the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v7/
Detailed changelog:
- Use admin-perm in netlink API.
- Addressed feedback from Jakub with regards to netlink API
implementation.
- Renamed devmem.c functions to something more appropriate for that
file.
- Improve the performance seen through the page_pool benchmark.
- Fix the value definition of all the SO_DEVMEM_* uapi.
- Various fixes to documentation.
Perf - page-pool benchmark:
---------------------------
Improved performance of bench_page_pool_simple.ko tests compared to v6:
https://pastebin.com/raw/v5dYRg8L
net-next base: 8 cycle fast path.
RFC v6: 10 cycle fast path.
RFC v7: 9 cycle fast path.
RFC v7 with CONFIG_DMA_SHARED_BUFFER disabled: 8 cycle fast path,
same as baseline.
Perf - Devmem TCP benchmark:
---------------------
Perf is about the same regardless of the changes in v7, namely the
removal of the static_branch_unlikely to improve the page_pool benchmark
performance:
189/200gbps bi-directional throughput with RX devmem TCP and regular TCP
TX i.e. ~95% line rate.
RFC v6:
=======
Major Changes:
--------------
This revision largely rebases on top of net-next and addresses the little
feedback RFCv5 received.
The series remains in RFC because the queue-API ndos defined in this
series are not yet implemented. I have a GVE implementation I carry out
of tree for my testing. A upstreamable GVE implementation is in the
works. Aside from that, in my estimation all the patches are ready for
review/merge. Please do take a look.
As usual the full devmem TCP changes including the full GVE driver
implementation is here:
https://github.com/mina/linux/commits/tcpdevmem-v6/
This version also comes with some performance data recorded in the cover
letter (see below changelog).
Detailed changelog:
- Rebased on top of the merged netmem_ref changes.
- Converted skb->dmabuf to skb->readable (Pavel). Pavel's original
suggestion was to remove the skb->dmabuf flag entirely, but when I
looked into it closely, I found the issue that if we remove the flag
we have to dereference the shinfo(skb) pointer to obtain the first
frag to tell whether an skb is readable or not. This can cause a
performance regression if it dirties the cache line when the
shinfo(skb) was not really needed. Instead, I converted the skb->dmabuf
flag into a generic skb->readable flag which can be re-used by io_uring
0-copy RX.
- Squashed a few locking optimizations from Eric Dumazet in the RX path
and the DEVMEM_DONTNEED setsockopt.
- Expanded the tests a bit. Added validation for invalid scenarios and
added some more coverage.
Perf - page-pool benchmark:
---------------------------
bench_page_pool_simple.ko tests with and without these changes:
https://pastebin.com/raw/ncHDwAbn
AFAIK the number that really matters in the perf tests is the
'tasklet_page_pool01_fast_path Per elem'. This one measures at about 8
cycles without the changes but there is some 1 cycle noise in some
results.
With the patches this regresses to 9 cycles with the changes but there
is 1 cycle noise occasionally running this test repeatedly.
Lastly I tried disable the static_branch_unlikely() in
netmem_is_net_iov() check. To my surprise disabling the
static_branch_unlikely() check reduces the fast path back to 8 cycles,
but the 1 cycle noise remains.
Perf - Devmem TCP benchmark:
---------------------
189/200gbps bi-directional throughput with RX devmem TCP and regular TCP
TX i.e. ~95% line rate.
Major changes in RFC v5:
========================
1. Rebased on top of 'Abstract page from net stack' series and used the
new netmem type to refer to LSB set pointers instead of re-using
struct page.
2. Downgraded this series back to RFC and called it RFC v5. This is
because this series is now dependent on 'Abstract page from net
stack'[1] and the queue API. Both are removed from the series to
reduce the patch # and those bits are fairly independent or
pre-requisite work.
3. Reworked the page_pool devmem support to use netmem and for some
more unified handling.
4. Reworked the reference counting of net_iov (renamed from
page_pool_iov) to use pp_ref_count for refcounting.
The full changes including the dependent series and GVE page pool
support is here:
https://github.com/mina/linux/commits/tcpdevmem-rfcv5/
[1] https://patchwork.kernel.org/project/netdevbpf/list/?series=810774
Major changes in v1:
====================
1. Implemented MVP queue API ndos to remove the userspace-visible
driver reset.
2. Fixed issues in the napi_pp_put_page() devmem frag unref path.
3. Removed RFC tag.
Many smaller addressed comments across all the patches (patches have
individual change log).
Full tree including the rest of the GVE driver changes:
https://github.com/mina/linux/commits/tcpdevmem-v1
Changes in RFC v3:
==================
1. Pulled in the memory-provider dependency from Jakub's RFC[1] to make the
series reviewable and mergeable.
2. Implemented multi-rx-queue binding which was a todo in v2.
3. Fix to cmsg handling.
The sticking point in RFC v2[2] was the device reset required to refill
the device rx-queues after the dmabuf bind/unbind. The solution
suggested as I understand is a subset of the per-queue management ops
Jakub suggested or similar:
https://lore.kernel.org/netdev/20230815171638.4c057dcd@kernel.org/
This is not addressed in this revision, because:
1. This point was discussed at netconf & netdev and there is openness to
using the current approach of requiring a device reset.
2. Implementing individual queue resetting seems to be difficult for my
test bed with GVE. My prototype to test this ran into issues with the
rx-queues not coming back up properly if reset individually. At the
moment I'm unsure if it's a mistake in the POC or a genuine issue in
the virtualization stack behind GVE, which currently doesn't test
individual rx-queue restart.
3. Our usecases are not bothered by requiring a device reset to refill
the buffer queues, and we'd like to support NICs that run into this
limitation with resetting individual queues.
My thought is that drivers that have trouble with per-queue configs can
use the support in this series, while drivers that support new netdev
ops to reset individual queues can automatically reset the queue as
part of the dma-buf bind/unbind.
The same approach with device resets is presented again for consideration
with other sticking points addressed.
This proposal includes the rx devmem path only proposed for merge. For a
snapshot of my entire tree which includes the GVE POC page pool support &
device memory support:
https://github.com/torvalds/linux/compare/master...mina:linux:tcpdevmem-v3
[1] https://lore.kernel.org/netdev/f8270765-a27b-6ccf-33ea-cda097168d79@redhat.…
[2] https://lore.kernel.org/netdev/CAHS8izOVJGJH5WF68OsRWFKJid1_huzzUK+hpKbLcL4…
Changes in RFC v2:
==================
The sticking point in RFC v1[1] was the dma-buf pages approach we used to
deliver the device memory to the TCP stack. RFC v2 is a proof-of-concept
that attempts to resolve this by implementing scatterlist support in the
networking stack, such that we can import the dma-buf scatterlist
directly. This is the approach proposed at a high level here[2].
Detailed changes:
1. Replaced dma-buf pages approach with importing scatterlist into the
page pool.
2. Replace the dma-buf pages centric API with a netlink API.
3. Removed the TX path implementation - there is no issue with
implementing the TX path with scatterlist approach, but leaving
out the TX path makes it easier to review.
4. Functionality is tested with this proposal, but I have not conducted
perf testing yet. I'm not sure there are regressions, but I removed
perf claims from the cover letter until they can be re-confirmed.
5. Added Signed-off-by: contributors to the implementation.
6. Fixed some bugs with the RX path since RFC v1.
Any feedback welcome, but specifically the biggest pending questions
needing feedback IMO are:
1. Feedback on the scatterlist-based approach in general.
2. Netlink API (Patch 1 & 2).
3. Approach to handle all the drivers that expect to receive pages from
the page pool (Patch 6).
[1] https://lore.kernel.org/netdev/dfe4bae7-13a0-3c5d-d671-f61b375cb0b4@gmail.c…
[2] https://lore.kernel.org/netdev/CAHS8izPm6XRS54LdCDZVd0C75tA1zHSu6jLVO8nzTLX…
==================
* TL;DR:
Device memory TCP (devmem TCP) is a proposal for transferring data to and/or
from device memory efficiently, without bouncing the data to a host memory
buffer.
* Problem:
A large amount of data transfers have device memory as the source and/or
destination. Accelerators drastically increased the volume of such transfers.
Some examples include:
- ML accelerators transferring large amounts of training data from storage into
GPU/TPU memory. In some cases ML training setup time can be as long as 50% of
TPU compute time, improving data transfer throughput & efficiency can help
improving GPU/TPU utilization.
- Distributed training, where ML accelerators, such as GPUs on different hosts,
exchange data among them.
- Distributed raw block storage applications transfer large amounts of data with
remote SSDs, much of this data does not require host processing.
Today, the majority of the Device-to-Device data transfers the network are
implemented as the following low level operations: Device-to-Host copy,
Host-to-Host network transfer, and Host-to-Device copy.
The implementation is suboptimal, especially for bulk data transfers, and can
put significant strains on system resources, such as host memory bandwidth,
PCIe bandwidth, etc. One important reason behind the current state is the
kernel’s lack of semantics to express device to network transfers.
* Proposal:
In this patch series we attempt to optimize this use case by implementing
socket APIs that enable the user to:
1. send device memory across the network directly, and
2. receive incoming network packets directly into device memory.
Packet _payloads_ go directly from the NIC to device memory for receive and from
device memory to NIC for transmit.
Packet _headers_ go to/from host memory and are processed by the TCP/IP stack
normally. The NIC _must_ support header split to achieve this.
Advantages:
- Alleviate host memory bandwidth pressure, compared to existing
network-transfer + device-copy semantics.
- Alleviate PCIe BW pressure, by limiting data transfer to the lowest level
of the PCIe tree, compared to traditional path which sends data through the
root complex.
* Patch overview:
** Part 1: netlink API
Gives user ability to bind dma-buf to an RX queue.
** Part 2: scatterlist support
Currently the standard for device memory sharing is DMABUF, which doesn't
generate struct pages. On the other hand, networking stack (skbs, drivers, and
page pool) operate on pages. We have 2 options:
1. Generate struct pages for dmabuf device memory, or,
2. Modify the networking stack to process scatterlist.
Approach #1 was attempted in RFC v1. RFC v2 implements approach #2.
** part 3: page pool support
We piggy back on page pool memory providers proposal:
https://github.com/kuba-moo/linux/tree/pp-providers
It allows the page pool to define a memory provider that provides the
page allocation and freeing. It helps abstract most of the device memory
TCP changes from the driver.
** part 4: support for unreadable skb frags
Page pool iovs are not accessible by the host; we implement changes
throughput the networking stack to correctly handle skbs with unreadable
frags.
** Part 5: recvmsg() APIs
We define user APIs for the user to send and receive device memory.
Not included with this series is the GVE devmem TCP support, just to
simplify the review. Code available here if desired:
https://github.com/mina/linux/tree/tcpdevmem
This series is built on top of net-next with Jakub's pp-providers changes
cherry-picked.
* NIC dependencies:
1. (strict) Devmem TCP require the NIC to support header split, i.e. the
capability to split incoming packets into a header + payload and to put
each into a separate buffer. Devmem TCP works by using device memory
for the packet payload, and host memory for the packet headers.
2. (optional) Devmem TCP works better with flow steering support & RSS support,
i.e. the NIC's ability to steer flows into certain rx queues. This allows the
sysadmin to enable devmem TCP on a subset of the rx queues, and steer
devmem TCP traffic onto these queues and non devmem TCP elsewhere.
The NIC I have access to with these properties is the GVE with DQO support
running in Google Cloud, but any NIC that supports these features would suffice.
I may be able to help reviewers bring up devmem TCP on their NICs.
* Testing:
The series includes a udmabuf kselftest that show a simple use case of
devmem TCP and validates the entire data path end to end without
a dependency on a specific dmabuf provider.
** Test Setup
Kernel: net-next with this series and memory provider API cherry-picked
locally.
Hardware: Google Cloud A3 VMs.
NIC: GVE with header split & RSS & flow steering support.
Cc: Pavel Begunkov <asml.silence(a)gmail.com>
Cc: David Wei <dw(a)davidwei.uk>
Cc: Jason Gunthorpe <jgg(a)ziepe.ca>
Cc: Yunsheng Lin <linyunsheng(a)huawei.com>
Cc: Shailend Chand <shailend(a)google.com>
Cc: Harshitha Ramamurthy <hramamurthy(a)google.com>
Cc: Shakeel Butt <shakeel.butt(a)linux.dev>
Cc: Jeroen de Borst <jeroendb(a)google.com>
Cc: Praveen Kaligineedi <pkaligineedi(a)google.com>
Jakub Kicinski (1):
net: page_pool: create hooks for custom page providers
Mina Almasry (13):
queue_api: define queue api
net: netdev netlink api to bind dma-buf to a net device
netdev: support binding dma-buf to netdevice
netdev: netdevice devmem allocator
page_pool: convert to use netmem
page_pool: devmem support
memory-provider: dmabuf devmem memory provider
net: support non paged skb frags
net: add support for skbs with unreadable frags
tcp: RX path for devmem TCP
net: add SO_DEVMEM_DONTNEED setsockopt to release RX frags
net: add devmem TCP documentation
selftests: add ncdevmem, netcat for devmem TCP
Documentation/netlink/specs/netdev.yaml | 57 +++
Documentation/networking/devmem.rst | 256 +++++++++++
Documentation/networking/index.rst | 1 +
arch/alpha/include/uapi/asm/socket.h | 6 +
arch/mips/include/uapi/asm/socket.h | 6 +
arch/parisc/include/uapi/asm/socket.h | 6 +
arch/sparc/include/uapi/asm/socket.h | 6 +
include/linux/netdevice.h | 3 +
include/linux/skbuff.h | 73 +++-
include/linux/socket.h | 1 +
include/net/devmem.h | 124 ++++++
include/net/netdev_queues.h | 27 ++
include/net/netdev_rx_queue.h | 2 +
include/net/netmem.h | 234 +++++++++-
include/net/page_pool/helpers.h | 155 +++++--
include/net/page_pool/types.h | 33 +-
include/net/sock.h | 2 +
include/net/tcp.h | 5 +-
include/trace/events/page_pool.h | 29 +-
include/uapi/asm-generic/socket.h | 6 +
include/uapi/linux/netdev.h | 19 +
include/uapi/linux/uio.h | 17 +
net/bpf/test_run.c | 5 +-
net/core/Makefile | 2 +-
net/core/datagram.c | 6 +
net/core/dev.c | 6 +-
net/core/devmem.c | 425 ++++++++++++++++++
net/core/gro.c | 8 +-
net/core/netdev-genl-gen.c | 23 +
net/core/netdev-genl-gen.h | 6 +
net/core/netdev-genl.c | 107 +++++
net/core/page_pool.c | 364 +++++++++-------
net/core/skbuff.c | 110 ++++-
net/core/sock.c | 61 +++
net/ipv4/esp4.c | 2 +-
net/ipv4/tcp.c | 254 ++++++++++-
net/ipv4/tcp_input.c | 13 +-
net/ipv4/tcp_ipv4.c | 9 +
net/ipv4/tcp_minisocks.c | 2 +
net/ipv4/tcp_output.c | 5 +-
net/ipv6/esp6.c | 2 +-
net/packet/af_packet.c | 4 +-
tools/include/uapi/linux/netdev.h | 19 +
tools/testing/selftests/net/.gitignore | 1 +
tools/testing/selftests/net/Makefile | 5 +
tools/testing/selftests/net/ncdevmem.c | 546 ++++++++++++++++++++++++
46 files changed, 2776 insertions(+), 277 deletions(-)
create mode 100644 Documentation/networking/devmem.rst
create mode 100644 include/net/devmem.h
create mode 100644 net/core/devmem.c
create mode 100644 tools/testing/selftests/net/ncdevmem.c
--
2.44.0.478.gd926399ef9-goog
When building with clang, via:
make LLVM=1 -C tools/testing/selftests
...two types of warnings occur:
warning: absolute value function 'abs' given an argument of type
'long' but has parameter of type 'int' which may cause truncation of
value
warning: taking the absolute value of unsigned type 'unsigned long'
has no effect
Fix these by:
a) using labs() in place of abs(), when long integers are involved, and
b) Change to use signed integer data types, in places where subtraction
is used (and could end up with negative values).
c) Remove a duplicate abs() call in cmt_test.c.
Cc: Ilpo Järvinen <ilpo.jarvinen(a)linux.intel.com>
Reviewed-by: Reinette Chatre <reinette.chatre(a)intel.com>
Signed-off-by: John Hubbard <jhubbard(a)nvidia.com>
---
tools/testing/selftests/resctrl/cmt_test.c | 4 ++--
tools/testing/selftests/resctrl/mba_test.c | 2 +-
tools/testing/selftests/resctrl/mbm_test.c | 2 +-
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/tools/testing/selftests/resctrl/cmt_test.c b/tools/testing/selftests/resctrl/cmt_test.c
index a81f91222a89..05a241519ae8 100644
--- a/tools/testing/selftests/resctrl/cmt_test.c
+++ b/tools/testing/selftests/resctrl/cmt_test.c
@@ -40,11 +40,11 @@ static int show_results_info(unsigned long sum_llc_val, int no_of_bits,
int ret;
avg_llc_val = sum_llc_val / num_of_runs;
- avg_diff = (long)abs(cache_span - avg_llc_val);
+ avg_diff = (long)(cache_span - avg_llc_val);
diff_percent = ((float)cache_span - avg_llc_val) / cache_span * 100;
ret = platform && abs((int)diff_percent) > max_diff_percent &&
- abs(avg_diff) > max_diff;
+ labs(avg_diff) > max_diff;
ksft_print_msg("%s Check cache miss rate within %lu%%\n",
ret ? "Fail:" : "Pass:", max_diff_percent);
diff --git a/tools/testing/selftests/resctrl/mba_test.c b/tools/testing/selftests/resctrl/mba_test.c
index 7946e32e85c8..8fd16b117092 100644
--- a/tools/testing/selftests/resctrl/mba_test.c
+++ b/tools/testing/selftests/resctrl/mba_test.c
@@ -60,8 +60,8 @@ static bool show_mba_info(unsigned long *bw_imc, unsigned long *bw_resc)
/* Memory bandwidth from 100% down to 10% */
for (allocation = 0; allocation < ALLOCATION_MAX / ALLOCATION_STEP;
allocation++) {
- unsigned long avg_bw_imc, avg_bw_resc;
unsigned long sum_bw_imc = 0, sum_bw_resc = 0;
+ long avg_bw_imc, avg_bw_resc;
int avg_diff_per;
float avg_diff;
diff --git a/tools/testing/selftests/resctrl/mbm_test.c b/tools/testing/selftests/resctrl/mbm_test.c
index d67ffa3ec63a..252c94ff2a3d 100644
--- a/tools/testing/selftests/resctrl/mbm_test.c
+++ b/tools/testing/selftests/resctrl/mbm_test.c
@@ -17,8 +17,8 @@
static int
show_bw_info(unsigned long *bw_imc, unsigned long *bw_resc, size_t span)
{
- unsigned long avg_bw_imc = 0, avg_bw_resc = 0;
unsigned long sum_bw_imc = 0, sum_bw_resc = 0;
+ long avg_bw_imc = 0, avg_bw_resc = 0;
int runs, ret, avg_diff_per;
float avg_diff = 0;
base-commit: 45db3ab70092637967967bfd8e6144017638563c
prerequisite-patch-id: b901ece2a5b78503e2fb5480f20e304d36a0ea27
prerequisite-patch-id: 8d96c4b8c3ed6d9ea2588ef7f594ae0f9f83c279
--
2.45.0
When building with clang, via:
make LLVM=1 -C tools/testing/selftests
...two types of warnings occur:
warning: absolute value function 'abs' given an argument of type
'long' but has parameter of type 'int' which may cause truncation of
value
warning: taking the absolute value of unsigned type 'unsigned long'
has no effect
Fix these by:
a) using labs() in place of abs(), when long integers are involved, and
b) Change to use signed integer data types, in places where subtraction
is used (and could end up with negative values).
c) Remove a duplicate abs() call in cmt_test.c.
Cc: Reinette Chatre <reinette.chatre(a)intel.com>
Cc: Ilpo Järvinen <ilpo.jarvinen(a)linux.intel.com>
Signed-off-by: John Hubbard <jhubbard(a)nvidia.com>
---
tools/testing/selftests/resctrl/cmt_test.c | 4 ++--
tools/testing/selftests/resctrl/mba_test.c | 2 +-
tools/testing/selftests/resctrl/mbm_test.c | 2 +-
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/tools/testing/selftests/resctrl/cmt_test.c b/tools/testing/selftests/resctrl/cmt_test.c
index a81f91222a89..05a241519ae8 100644
--- a/tools/testing/selftests/resctrl/cmt_test.c
+++ b/tools/testing/selftests/resctrl/cmt_test.c
@@ -40,11 +40,11 @@ static int show_results_info(unsigned long sum_llc_val, int no_of_bits,
int ret;
avg_llc_val = sum_llc_val / num_of_runs;
- avg_diff = (long)abs(cache_span - avg_llc_val);
+ avg_diff = (long)(cache_span - avg_llc_val);
diff_percent = ((float)cache_span - avg_llc_val) / cache_span * 100;
ret = platform && abs((int)diff_percent) > max_diff_percent &&
- abs(avg_diff) > max_diff;
+ labs(avg_diff) > max_diff;
ksft_print_msg("%s Check cache miss rate within %lu%%\n",
ret ? "Fail:" : "Pass:", max_diff_percent);
diff --git a/tools/testing/selftests/resctrl/mba_test.c b/tools/testing/selftests/resctrl/mba_test.c
index 7946e32e85c8..5fffbc9ff6a4 100644
--- a/tools/testing/selftests/resctrl/mba_test.c
+++ b/tools/testing/selftests/resctrl/mba_test.c
@@ -60,7 +60,7 @@ static bool show_mba_info(unsigned long *bw_imc, unsigned long *bw_resc)
/* Memory bandwidth from 100% down to 10% */
for (allocation = 0; allocation < ALLOCATION_MAX / ALLOCATION_STEP;
allocation++) {
- unsigned long avg_bw_imc, avg_bw_resc;
+ long avg_bw_imc, avg_bw_resc;
unsigned long sum_bw_imc = 0, sum_bw_resc = 0;
int avg_diff_per;
float avg_diff;
diff --git a/tools/testing/selftests/resctrl/mbm_test.c b/tools/testing/selftests/resctrl/mbm_test.c
index d67ffa3ec63a..a4c3ea49b0e8 100644
--- a/tools/testing/selftests/resctrl/mbm_test.c
+++ b/tools/testing/selftests/resctrl/mbm_test.c
@@ -17,7 +17,7 @@
static int
show_bw_info(unsigned long *bw_imc, unsigned long *bw_resc, size_t span)
{
- unsigned long avg_bw_imc = 0, avg_bw_resc = 0;
+ long avg_bw_imc = 0, avg_bw_resc = 0;
unsigned long sum_bw_imc = 0, sum_bw_resc = 0;
int runs, ret, avg_diff_per;
float avg_diff = 0;
base-commit: 45db3ab70092637967967bfd8e6144017638563c
--
2.45.0
If mnl_socket_open() or mnl_socket_bind() fails, it's generally due to
not having the user space parts fully installed and configured
correctly. This was previously ignored and reported as a test PASS, but
what really happened is that the tests were being skipped. This led to
generating inaccurate TAP output (I've omitted the leading '#'
character):
Totals: pass:3 fail:0 xfail:0 xpass:0 skip:0 error:0
Fix this by using kselftest's SKIP() macro. The new output on the same
misconfigured system shows:
Totals: pass:0 fail:0 xfail:0 xpass:0 skip:3 error:0
This was briefly discussed already with Felix Huettner [1].
[1] https://lore.kernel.org/all/edca87be-d8a7-4c62-b9c1-f9b3f5752595@nvidia.com/
Cc: Felix Huettner <felix.huettner(a)mail.schwarz>
Signed-off-by: John Hubbard <jhubbard(a)nvidia.com>
---
tools/testing/selftests/netfilter/conntrack_dump_flush.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tools/testing/selftests/netfilter/conntrack_dump_flush.c b/tools/testing/selftests/netfilter/conntrack_dump_flush.c
index e9df4ae14e16..4a73afad4de4 100644
--- a/tools/testing/selftests/netfilter/conntrack_dump_flush.c
+++ b/tools/testing/selftests/netfilter/conntrack_dump_flush.c
@@ -317,12 +317,12 @@ FIXTURE_SETUP(conntrack_dump_flush)
self->sock = mnl_socket_open(NETLINK_NETFILTER);
if (!self->sock) {
perror("mnl_socket_open");
- exit(EXIT_FAILURE);
+ SKIP(exit(EXIT_FAILURE), "mnl_socket_open() failed");
}
if (mnl_socket_bind(self->sock, 0, MNL_SOCKET_AUTOPID) < 0) {
perror("mnl_socket_bind");
- exit(EXIT_FAILURE);
+ SKIP(exit(EXIT_FAILURE), "mnl_socket_bind() failed");
}
ret = conntracK_count_zone(self->sock, TEST_ZONE_ID);
base-commit: 45db3ab70092637967967bfd8e6144017638563c
prerequisite-patch-id: b901ece2a5b78503e2fb5480f20e304d36a0ea27
prerequisite-patch-id: 9db2d20be98dc44731d8605a3da64ff118d2546d
prerequisite-patch-id: 72413b9b47d66666f20967a664470199892fe282
--
2.45.0
First of all, in order to build with clang at all, one must first apply
Valentin Obst's build fix for LLVM [1]. Furthermore, for this particular
resctrl directory, my pending fix [2] must also be applied. Once those
fixes are in place, then when building with clang, via:
make LLVM=1 -C tools/testing/selftests
...two types of warnings occur:
warning: absolute value function 'abs' given an argument of type
'long' but has parameter of type 'int' which may cause truncation of
value
warning: taking the absolute value of unsigned type 'unsigned long'
has no effect
Fix these by:
a) using labs() in place of abs(), when long integers are involved, and
b) Change to use signed integer data types, in places where subtraction
is used (and could end up with negative values).
[1] https://lore.kernel.org/all/20240329-selftests-libmk-llvm-rfc-v1-1-2f9ed7d1…
[2] https://lore.kernel.org/all/20240503021712.78601-1-jhubbard@nvidia.com/
Cc: Reinette Chatre <reinette.chatre(a)intel.com>
Signed-off-by: John Hubbard <jhubbard(a)nvidia.com>
---
Hi Reinette,
This v2 includes a fix for the bugs that you pointed out (thanks!) in v1.
I kept the changes to signed integers minimal: only what is required in
order to get a clean clang build.
thanks,
John Hubbard
tools/testing/selftests/resctrl/cmt_test.c | 12 ++++++------
tools/testing/selftests/resctrl/mba_test.c | 4 ++--
tools/testing/selftests/resctrl/mbm_test.c | 4 ++--
3 files changed, 10 insertions(+), 10 deletions(-)
diff --git a/tools/testing/selftests/resctrl/cmt_test.c b/tools/testing/selftests/resctrl/cmt_test.c
index a81f91222a89..af33abd1cca7 100644
--- a/tools/testing/selftests/resctrl/cmt_test.c
+++ b/tools/testing/selftests/resctrl/cmt_test.c
@@ -29,22 +29,22 @@ static int cmt_setup(const struct resctrl_test *test,
return 0;
}
-static int show_results_info(unsigned long sum_llc_val, int no_of_bits,
- unsigned long cache_span, unsigned long max_diff,
- unsigned long max_diff_percent, unsigned long num_of_runs,
+static int show_results_info(long sum_llc_val, int no_of_bits,
+ long cache_span, long max_diff,
+ long max_diff_percent, long num_of_runs,
bool platform)
{
- unsigned long avg_llc_val = 0;
+ long avg_llc_val = 0;
float diff_percent;
long avg_diff = 0;
int ret;
avg_llc_val = sum_llc_val / num_of_runs;
- avg_diff = (long)abs(cache_span - avg_llc_val);
+ avg_diff = labs(cache_span - avg_llc_val);
diff_percent = ((float)cache_span - avg_llc_val) / cache_span * 100;
ret = platform && abs((int)diff_percent) > max_diff_percent &&
- abs(avg_diff) > max_diff;
+ labs(avg_diff) > max_diff;
ksft_print_msg("%s Check cache miss rate within %lu%%\n",
ret ? "Fail:" : "Pass:", max_diff_percent);
diff --git a/tools/testing/selftests/resctrl/mba_test.c b/tools/testing/selftests/resctrl/mba_test.c
index 7946e32e85c8..707b07687249 100644
--- a/tools/testing/selftests/resctrl/mba_test.c
+++ b/tools/testing/selftests/resctrl/mba_test.c
@@ -60,8 +60,8 @@ static bool show_mba_info(unsigned long *bw_imc, unsigned long *bw_resc)
/* Memory bandwidth from 100% down to 10% */
for (allocation = 0; allocation < ALLOCATION_MAX / ALLOCATION_STEP;
allocation++) {
- unsigned long avg_bw_imc, avg_bw_resc;
- unsigned long sum_bw_imc = 0, sum_bw_resc = 0;
+ long avg_bw_imc, avg_bw_resc;
+ long sum_bw_imc = 0, sum_bw_resc = 0;
int avg_diff_per;
float avg_diff;
diff --git a/tools/testing/selftests/resctrl/mbm_test.c b/tools/testing/selftests/resctrl/mbm_test.c
index d67ffa3ec63a..30af15020731 100644
--- a/tools/testing/selftests/resctrl/mbm_test.c
+++ b/tools/testing/selftests/resctrl/mbm_test.c
@@ -17,8 +17,8 @@
static int
show_bw_info(unsigned long *bw_imc, unsigned long *bw_resc, size_t span)
{
- unsigned long avg_bw_imc = 0, avg_bw_resc = 0;
- unsigned long sum_bw_imc = 0, sum_bw_resc = 0;
+ long avg_bw_imc = 0, avg_bw_resc = 0;
+ long sum_bw_imc = 0, sum_bw_resc = 0;
int runs, ret, avg_diff_per;
float avg_diff = 0;
base-commit: ddb4c3f25b7b95df3d6932db0b379d768a6ebdf7
prerequisite-patch-id: b901ece2a5b78503e2fb5480f20e304d36a0ea27
prerequisite-patch-id: 8d96c4b8c3ed6d9ea2588ef7f594ae0f9f83c279
--
2.45.0