]> git.ipfire.org Git - thirdparty/kernel/stable-queue.git/commitdiff
6.6-stable patches
authorGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 22 Apr 2025 07:55:44 +0000 (09:55 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 22 Apr 2025 07:55:44 +0000 (09:55 +0200)
added patches:
fix-mmu-notifiers-for-range-based-invalidates.patch
misc-pci_endpoint_test-avoid-issue-of-interrupts-remaining-after-request_irq-error.patch
misc-pci_endpoint_test-fix-displaying-irq_type-after-request_irq-error.patch
misc-pci_endpoint_test-fix-irq_type-to-convey-the-correct-type.patch
mptcp-sockopt-fix-getting-freebind-transparent.patch
selftests-mptcp-add-mptcp_lib_wait_local_port_listen.patch

queue-6.6/fix-mmu-notifiers-for-range-based-invalidates.patch [new file with mode: 0644]
queue-6.6/misc-pci_endpoint_test-avoid-issue-of-interrupts-remaining-after-request_irq-error.patch [new file with mode: 0644]
queue-6.6/misc-pci_endpoint_test-fix-displaying-irq_type-after-request_irq-error.patch [new file with mode: 0644]
queue-6.6/misc-pci_endpoint_test-fix-irq_type-to-convey-the-correct-type.patch [new file with mode: 0644]
queue-6.6/mptcp-sockopt-fix-getting-freebind-transparent.patch [new file with mode: 0644]
queue-6.6/selftests-mptcp-add-mptcp_lib_wait_local_port_listen.patch [new file with mode: 0644]
queue-6.6/series

diff --git a/queue-6.6/fix-mmu-notifiers-for-range-based-invalidates.patch b/queue-6.6/fix-mmu-notifiers-for-range-based-invalidates.patch
new file mode 100644 (file)
index 0000000..0eddcf9
--- /dev/null
@@ -0,0 +1,98 @@
+From f7edb07ad7c66eab3dce57384f33b9799d579133 Mon Sep 17 00:00:00 2001
+From: Piotr Jaroszynski <pjaroszynski@nvidia.com>
+Date: Tue, 4 Mar 2025 00:51:27 -0800
+Subject: Fix mmu notifiers for range-based invalidates
+
+From: Piotr Jaroszynski <pjaroszynski@nvidia.com>
+
+commit f7edb07ad7c66eab3dce57384f33b9799d579133 upstream.
+
+Update the __flush_tlb_range_op macro not to modify its parameters as
+these are unexepcted semantics. In practice, this fixes the call to
+mmu_notifier_arch_invalidate_secondary_tlbs() in
+__flush_tlb_range_nosync() to use the correct range instead of an empty
+range with start=end. The empty range was (un)lucky as it results in
+taking the invalidate-all path that doesn't cause correctness issues,
+but can certainly result in suboptimal perf.
+
+This has been broken since commit 6bbd42e2df8f ("mmu_notifiers: call
+invalidate_range() when invalidating TLBs") when the call to the
+notifiers was added to __flush_tlb_range(). It predates the addition of
+the __flush_tlb_range_op() macro from commit 360839027a6e ("arm64: tlb:
+Refactor the core flush algorithm of __flush_tlb_range") that made the
+bug hard to spot.
+
+Fixes: 6bbd42e2df8f ("mmu_notifiers: call invalidate_range() when invalidating TLBs")
+
+Signed-off-by: Piotr Jaroszynski <pjaroszynski@nvidia.com>
+Cc: Catalin Marinas <catalin.marinas@arm.com>
+Cc: Will Deacon <will@kernel.org>
+Cc: Robin Murphy <robin.murphy@arm.com>
+Cc: Alistair Popple <apopple@nvidia.com>
+Cc: Raghavendra Rao Ananta <rananta@google.com>
+Cc: SeongJae Park <sj@kernel.org>
+Cc: Jason Gunthorpe <jgg@nvidia.com>
+Cc: John Hubbard <jhubbard@nvidia.com>
+Cc: Nicolin Chen <nicolinc@nvidia.com>
+Cc: linux-arm-kernel@lists.infradead.org
+Cc: iommu@lists.linux.dev
+Cc: linux-mm@kvack.org
+Cc: linux-kernel@vger.kernel.org
+Cc: stable@vger.kernel.org
+Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
+Reviewed-by: Alistair Popple <apopple@nvidia.com>
+Link: https://lore.kernel.org/r/20250304085127.2238030-1-pjaroszynski@nvidia.com
+Signed-off-by: Will Deacon <will@kernel.org>
+[will: Resolve conflicts due to lack of LPA2 support]
+Signed-off-by: Will Deacon <will@kernel.org>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ arch/arm64/include/asm/tlbflush.h |   22 ++++++++++++----------
+ 1 file changed, 12 insertions(+), 10 deletions(-)
+
+--- a/arch/arm64/include/asm/tlbflush.h
++++ b/arch/arm64/include/asm/tlbflush.h
+@@ -369,31 +369,33 @@ static inline void arch_tlbbatch_flush(s
+ #define __flush_tlb_range_op(op, start, pages, stride,                        \
+                               asid, tlb_level, tlbi_user)             \
+ do {                                                                  \
++      typeof(start) __flush_start = start;                            \
++      typeof(pages) __flush_pages = pages;                            \
+       int num = 0;                                                    \
+       int scale = 3;                                                  \
+       unsigned long addr;                                             \
+                                                                       \
+-      while (pages > 0) {                                             \
++      while (__flush_pages > 0) {                                     \
+               if (!system_supports_tlb_range() ||                     \
+-                  pages == 1) {                                       \
+-                      addr = __TLBI_VADDR(start, asid);               \
++                  __flush_pages == 1) {                               \
++                      addr = __TLBI_VADDR(__flush_start, asid);       \
+                       __tlbi_level(op, addr, tlb_level);              \
+                       if (tlbi_user)                                  \
+                               __tlbi_user_level(op, addr, tlb_level); \
+-                      start += stride;                                \
+-                      pages -= stride >> PAGE_SHIFT;                  \
++                      __flush_start += stride;                        \
++                      __flush_pages -= stride >> PAGE_SHIFT;          \
+                       continue;                                       \
+               }                                                       \
+                                                                       \
+-              num = __TLBI_RANGE_NUM(pages, scale);                   \
++              num = __TLBI_RANGE_NUM(__flush_pages, scale);           \
+               if (num >= 0) {                                         \
+-                      addr = __TLBI_VADDR_RANGE(start, asid, scale,   \
+-                                                num, tlb_level);      \
++                      addr = __TLBI_VADDR_RANGE(__flush_start, asid,  \
++                                              scale, num, tlb_level); \
+                       __tlbi(r##op, addr);                            \
+                       if (tlbi_user)                                  \
+                               __tlbi_user(r##op, addr);               \
+-                      start += __TLBI_RANGE_PAGES(num, scale) << PAGE_SHIFT; \
+-                      pages -= __TLBI_RANGE_PAGES(num, scale);        \
++                      __flush_start += __TLBI_RANGE_PAGES(num, scale) << PAGE_SHIFT; \
++                      __flush_pages -= __TLBI_RANGE_PAGES(num, scale);\
+               }                                                       \
+               scale--;                                                \
+       }                                                               \
diff --git a/queue-6.6/misc-pci_endpoint_test-avoid-issue-of-interrupts-remaining-after-request_irq-error.patch b/queue-6.6/misc-pci_endpoint_test-avoid-issue-of-interrupts-remaining-after-request_irq-error.patch
new file mode 100644 (file)
index 0000000..8a8aa75
--- /dev/null
@@ -0,0 +1,51 @@
+From f6cb7828c8e17520d4f5afb416515d3fae1af9a9 Mon Sep 17 00:00:00 2001
+From: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+Date: Tue, 25 Feb 2025 20:02:48 +0900
+Subject: misc: pci_endpoint_test: Avoid issue of interrupts remaining after request_irq error
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+From: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+
+commit f6cb7828c8e17520d4f5afb416515d3fae1af9a9 upstream.
+
+After devm_request_irq() fails with error in pci_endpoint_test_request_irq(),
+the pci_endpoint_test_free_irq_vectors() is called assuming that all IRQs
+have been released.
+
+However, some requested IRQs remain unreleased, so there are still
+/proc/irq/* entries remaining, and this results in WARN() with the
+following message:
+
+  remove_proc_entry: removing non-empty directory 'irq/30', leaking at least 'pci-endpoint-test.0'
+  WARNING: CPU: 0 PID: 202 at fs/proc/generic.c:719 remove_proc_entry +0x190/0x19c
+
+To solve this issue, set the number of remaining IRQs to test->num_irqs,
+and release IRQs in advance by calling pci_endpoint_test_release_irq().
+
+Cc: stable@vger.kernel.org
+Fixes: e03327122e2c ("pci_endpoint_test: Add 2 ioctl commands")
+Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
+Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+Link: https://lore.kernel.org/r/20250225110252.28866-3-hayashi.kunihiko@socionext.com
+[kwilczynski: commit log]
+Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
+Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ drivers/misc/pci_endpoint_test.c |    3 +++
+ 1 file changed, 3 insertions(+)
+
+--- a/drivers/misc/pci_endpoint_test.c
++++ b/drivers/misc/pci_endpoint_test.c
+@@ -260,6 +260,9 @@ fail:
+               break;
+       }
++      test->num_irqs = i;
++      pci_endpoint_test_release_irq(test);
++
+       return false;
+ }
diff --git a/queue-6.6/misc-pci_endpoint_test-fix-displaying-irq_type-after-request_irq-error.patch b/queue-6.6/misc-pci_endpoint_test-fix-displaying-irq_type-after-request_irq-error.patch
new file mode 100644 (file)
index 0000000..aad6b70
--- /dev/null
@@ -0,0 +1,54 @@
+From 919d14603dab6a9cf03ebbeb2cfa556df48737c8 Mon Sep 17 00:00:00 2001
+From: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+Date: Tue, 25 Feb 2025 20:02:49 +0900
+Subject: misc: pci_endpoint_test: Fix displaying 'irq_type' after 'request_irq' error
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+From: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+
+commit 919d14603dab6a9cf03ebbeb2cfa556df48737c8 upstream.
+
+There are two variables that indicate the interrupt type to be used
+in the next test execution, global "irq_type" and "test->irq_type".
+
+The former is referenced from pci_endpoint_test_get_irq() to preserve
+the current type for ioctl(PCITEST_GET_IRQTYPE).
+
+In the pci_endpoint_test_request_irq(), since this global variable
+is referenced when an error occurs, the unintended error message is
+displayed.
+
+For example, after running "pcitest -i 2", the following message
+shows "MSI 3" even if the current IRQ type becomes "MSI-X":
+
+  pci-endpoint-test 0000:01:00.0: Failed to request IRQ 30 for MSI 3
+  SET IRQ TYPE TO MSI-X:          NOT OKAY
+
+Fix this issue by using "test->irq_type" instead of global "irq_type".
+
+Cc: stable@vger.kernel.org
+Fixes: b2ba9225e031 ("misc: pci_endpoint_test: Avoid using module parameter to determine irqtype")
+Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
+Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+Link: https://lore.kernel.org/r/20250225110252.28866-4-hayashi.kunihiko@socionext.com
+[kwilczynski: commit log]
+Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
+Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ drivers/misc/pci_endpoint_test.c |    2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+--- a/drivers/misc/pci_endpoint_test.c
++++ b/drivers/misc/pci_endpoint_test.c
+@@ -243,7 +243,7 @@ static bool pci_endpoint_test_request_ir
+       return true;
+ fail:
+-      switch (irq_type) {
++      switch (test->irq_type) {
+       case IRQ_TYPE_LEGACY:
+               dev_err(dev, "Failed to request IRQ %d for Legacy\n",
+                       pci_irq_vector(pdev, i));
diff --git a/queue-6.6/misc-pci_endpoint_test-fix-irq_type-to-convey-the-correct-type.patch b/queue-6.6/misc-pci_endpoint_test-fix-irq_type-to-convey-the-correct-type.patch
new file mode 100644 (file)
index 0000000..9d642f2
--- /dev/null
@@ -0,0 +1,65 @@
+From baaef0a274cfb75f9b50eab3ef93205e604f662c Mon Sep 17 00:00:00 2001
+From: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+Date: Tue, 25 Feb 2025 20:02:50 +0900
+Subject: misc: pci_endpoint_test: Fix 'irq_type' to convey the correct type
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+From: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+
+commit baaef0a274cfb75f9b50eab3ef93205e604f662c upstream.
+
+There are two variables that indicate the interrupt type to be used
+in the next test execution, "irq_type" as global and "test->irq_type".
+
+The global is referenced from pci_endpoint_test_get_irq() to preserve
+the current type for ioctl(PCITEST_GET_IRQTYPE).
+
+The type set in this function isn't reflected in the global "irq_type",
+so ioctl(PCITEST_GET_IRQTYPE) returns the previous type.
+
+As a result, the wrong type is displayed in old version of "pcitest"
+as follows:
+
+  - Result of running "pcitest -i 0"
+
+      SET IRQ TYPE TO LEGACY:         OKAY
+
+  - Result of running "pcitest -I"
+
+      GET IRQ TYPE:           MSI
+
+Whereas running the new version of "pcitest" in kselftest results in an
+error as follows:
+
+  #  RUN           pci_ep_basic.LEGACY_IRQ_TEST ...
+  # pci_endpoint_test.c:104:LEGACY_IRQ_TEST:Expected 0 (0) == ret (1)
+  # pci_endpoint_test.c:104:LEGACY_IRQ_TEST:Can't get Legacy IRQ type
+
+Fix this issue by propagating the current type to the global "irq_type".
+
+Fixes: b2ba9225e031 ("misc: pci_endpoint_test: Avoid using module parameter to determine irqtype")
+Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+[kwilczynski: commit log]
+Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
+Reviewed-by: Niklas Cassel <cassel@kernel.org>
+Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
+Cc: stable@vger.kernel.org
+Link: https://lore.kernel.org/r/20250225110252.28866-5-hayashi.kunihiko@socionext.com
+Signed-off-by: Kunihiko Hayashi <hayashi.kunihiko@socionext.com>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ drivers/misc/pci_endpoint_test.c |    1 +
+ 1 file changed, 1 insertion(+)
+
+--- a/drivers/misc/pci_endpoint_test.c
++++ b/drivers/misc/pci_endpoint_test.c
+@@ -711,6 +711,7 @@ static bool pci_endpoint_test_set_irq(st
+       if (!pci_endpoint_test_request_irq(test))
+               goto err;
++      irq_type = test->irq_type;
+       return true;
+ err:
diff --git a/queue-6.6/mptcp-sockopt-fix-getting-freebind-transparent.patch b/queue-6.6/mptcp-sockopt-fix-getting-freebind-transparent.patch
new file mode 100644 (file)
index 0000000..266d063
--- /dev/null
@@ -0,0 +1,67 @@
+From e2f4ac7bab2205d3c4dd9464e6ffd82502177c51 Mon Sep 17 00:00:00 2001
+From: "Matthieu Baerts (NGI0)" <matttbe@kernel.org>
+Date: Fri, 14 Mar 2025 21:11:33 +0100
+Subject: mptcp: sockopt: fix getting freebind & transparent
+
+From: Matthieu Baerts (NGI0) <matttbe@kernel.org>
+
+commit e2f4ac7bab2205d3c4dd9464e6ffd82502177c51 upstream.
+
+When adding a socket option support in MPTCP, both the get and set parts
+are supposed to be implemented.
+
+IP(V6)_FREEBIND and IP(V6)_TRANSPARENT support for the setsockopt part
+has been added a while ago, but it looks like the get part got
+forgotten. It should have been present as a way to verify a setting has
+been set as expected, and not to act differently from TCP or any other
+socket types.
+
+Everything was in place to expose it, just the last step was missing.
+Only new code is added to cover these specific getsockopt(), that seems
+safe.
+
+Fixes: c9406a23c116 ("mptcp: sockopt: add SOL_IP freebind & transparent options")
+Cc: stable@vger.kernel.org
+Reviewed-by: Mat Martineau <martineau@kernel.org>
+Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
+Reviewed-by: Simon Horman <horms@kernel.org>
+Link: https://patch.msgid.link/20250314-net-mptcp-fix-data-stream-corr-sockopt-v1-3-122dbb249db3@kernel.org
+Signed-off-by: Paolo Abeni <pabeni@redhat.com>
+[ Conflict in sockopt.c due to commit e08d0b3d1723 ("inet: implement
+  lockless IP_TOS") not being in this version. The conflict is in the
+  context and the modification can still be applied in
+  mptcp_getsockopt_v4() after the IP_TOS case. ]
+Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ net/mptcp/sockopt.c |   12 ++++++++++++
+ 1 file changed, 12 insertions(+)
+
+--- a/net/mptcp/sockopt.c
++++ b/net/mptcp/sockopt.c
+@@ -1388,6 +1388,12 @@ static int mptcp_getsockopt_v4(struct mp
+       switch (optname) {
+       case IP_TOS:
+               return mptcp_put_int_option(msk, optval, optlen, inet_sk(sk)->tos);
++      case IP_FREEBIND:
++              return mptcp_put_int_option(msk, optval, optlen,
++                              inet_test_bit(FREEBIND, sk));
++      case IP_TRANSPARENT:
++              return mptcp_put_int_option(msk, optval, optlen,
++                              inet_test_bit(TRANSPARENT, sk));
+       }
+       return -EOPNOTSUPP;
+@@ -1402,6 +1408,12 @@ static int mptcp_getsockopt_v6(struct mp
+       case IPV6_V6ONLY:
+               return mptcp_put_int_option(msk, optval, optlen,
+                                           sk->sk_ipv6only);
++      case IPV6_TRANSPARENT:
++              return mptcp_put_int_option(msk, optval, optlen,
++                                          inet_test_bit(TRANSPARENT, sk));
++      case IPV6_FREEBIND:
++              return mptcp_put_int_option(msk, optval, optlen,
++                                          inet_test_bit(FREEBIND, sk));
+       }
+       return -EOPNOTSUPP;
diff --git a/queue-6.6/selftests-mptcp-add-mptcp_lib_wait_local_port_listen.patch b/queue-6.6/selftests-mptcp-add-mptcp_lib_wait_local_port_listen.patch
new file mode 100644 (file)
index 0000000..a0b61a9
--- /dev/null
@@ -0,0 +1,225 @@
+From 9369777c29395730cec967e7d0f48aed872b7110 Mon Sep 17 00:00:00 2001
+From: Geliang Tang <geliang.tang@suse.com>
+Date: Tue, 28 Nov 2023 15:18:59 -0800
+Subject: selftests: mptcp: add mptcp_lib_wait_local_port_listen
+
+From: Geliang Tang <geliang.tang@suse.com>
+
+commit 9369777c29395730cec967e7d0f48aed872b7110 upstream.
+
+To avoid duplicated code in different MPTCP selftests, we can add
+and use helpers defined in mptcp_lib.sh.
+
+wait_local_port_listen() helper is defined in diag.sh, mptcp_connect.sh,
+mptcp_join.sh and simult_flows.sh, export it into mptcp_lib.sh and
+rename it with mptcp_lib_ prefix. Use this new helper in all these
+scripts.
+
+Note: We only have IPv4 connections in this helper, not looking at IPv6
+(tcp6) but that's OK because we only have IPv4 connections here in diag.sh.
+
+Reviewed-by: Matthieu Baerts <matttbe@kernel.org>
+Signed-off-by: Geliang Tang <geliang.tang@suse.com>
+Signed-off-by: Mat Martineau <martineau@kernel.org>
+Link: https://lore.kernel.org/r/20231128-send-net-next-2023107-v4-15-8d6b94150f6b@kernel.org
+Signed-off-by: Jakub Kicinski <kuba@kernel.org>
+Stable-dep-of: 5afca7e996c4 ("selftests: mptcp: join: test for prohibited MPC to port-based endp")
+[ Conflict in diag.sh, because commit 1f24ba67ba49 ("selftests: mptcp:
+  diag: check CURRESTAB counters") that is more recent that the one
+  here, has been backported in this kernel version before, introducing
+  chk_msk_cestab() helper in the same context. wait_local_port_listen()
+  was still the same as in the original, and can then be simply removed
+  from diag.sh. ]
+Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ tools/testing/selftests/net/mptcp/diag.sh          |   23 ++-------------------
+ tools/testing/selftests/net/mptcp/mptcp_connect.sh |   19 -----------------
+ tools/testing/selftests/net/mptcp/mptcp_join.sh    |   20 ------------------
+ tools/testing/selftests/net/mptcp/mptcp_lib.sh     |   18 ++++++++++++++++
+ tools/testing/selftests/net/mptcp/simult_flows.sh  |   19 -----------------
+ 5 files changed, 24 insertions(+), 75 deletions(-)
+
+--- a/tools/testing/selftests/net/mptcp/diag.sh
++++ b/tools/testing/selftests/net/mptcp/diag.sh
+@@ -186,23 +186,6 @@ chk_msk_inuse()
+       __chk_nr get_msk_inuse $expected "${msg}" 0
+ }
+-# $1: ns, $2: port
+-wait_local_port_listen()
+-{
+-      local listener_ns="${1}"
+-      local port="${2}"
+-
+-      local port_hex i
+-
+-      port_hex="$(printf "%04X" "${port}")"
+-      for i in $(seq 10); do
+-              ip netns exec "${listener_ns}" cat /proc/net/tcp | \
+-                      awk "BEGIN {rc=1} {if (\$2 ~ /:${port_hex}\$/ && \$4 ~ /0A/) {rc=0; exit}} END {exit rc}" &&
+-                      break
+-              sleep 0.1
+-      done
+-}
+-
+ # $1: cestab nr
+ chk_msk_cestab()
+ {
+@@ -240,7 +223,7 @@ echo "a" | \
+               ip netns exec $ns \
+                       ./mptcp_connect -p 10000 -l -t ${timeout_poll} -w 20 \
+                               0.0.0.0 >/dev/null &
+-wait_local_port_listen $ns 10000
++mptcp_lib_wait_local_port_listen $ns 10000
+ chk_msk_nr 0 "no msk on netns creation"
+ chk_msk_listen 10000
+@@ -265,7 +248,7 @@ echo "a" | \
+               ip netns exec $ns \
+                       ./mptcp_connect -p 10001 -l -s TCP -t ${timeout_poll} -w 20 \
+                               0.0.0.0 >/dev/null &
+-wait_local_port_listen $ns 10001
++mptcp_lib_wait_local_port_listen $ns 10001
+ echo "b" | \
+       timeout ${timeout_test} \
+               ip netns exec $ns \
+@@ -288,7 +271,7 @@ for I in `seq 1 $NR_CLIENTS`; do
+                               ./mptcp_connect -p $((I+10001)) -l -w 20 \
+                                       -t ${timeout_poll} 0.0.0.0 >/dev/null &
+ done
+-wait_local_port_listen $ns $((NR_CLIENTS + 10001))
++mptcp_lib_wait_local_port_listen $ns $((NR_CLIENTS + 10001))
+ for I in `seq 1 $NR_CLIENTS`; do
+       echo "b" | \
+--- a/tools/testing/selftests/net/mptcp/mptcp_connect.sh
++++ b/tools/testing/selftests/net/mptcp/mptcp_connect.sh
+@@ -342,23 +342,6 @@ do_ping()
+       return 0
+ }
+-# $1: ns, $2: port
+-wait_local_port_listen()
+-{
+-      local listener_ns="${1}"
+-      local port="${2}"
+-
+-      local port_hex i
+-
+-      port_hex="$(printf "%04X" "${port}")"
+-      for i in $(seq 10); do
+-              ip netns exec "${listener_ns}" cat /proc/net/tcp* | \
+-                      awk "BEGIN {rc=1} {if (\$2 ~ /:${port_hex}\$/ && \$4 ~ /0A/) {rc=0; exit}} END {exit rc}" &&
+-                      break
+-              sleep 0.1
+-      done
+-}
+-
+ do_transfer()
+ {
+       local listener_ns="$1"
+@@ -448,7 +431,7 @@ do_transfer()
+                               $extra_args $local_addr < "$sin" > "$sout" &
+       local spid=$!
+-      wait_local_port_listen "${listener_ns}" "${port}"
++      mptcp_lib_wait_local_port_listen "${listener_ns}" "${port}"
+       local start
+       start=$(date +%s%3N)
+--- a/tools/testing/selftests/net/mptcp/mptcp_join.sh
++++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh
+@@ -598,24 +598,6 @@ link_failure()
+       done
+ }
+-# $1: ns, $2: port
+-wait_local_port_listen()
+-{
+-      local listener_ns="${1}"
+-      local port="${2}"
+-
+-      local port_hex
+-      port_hex="$(printf "%04X" "${port}")"
+-
+-      local i
+-      for i in $(seq 10); do
+-              ip netns exec "${listener_ns}" cat /proc/net/tcp* | \
+-                      awk "BEGIN {rc=1} {if (\$2 ~ /:${port_hex}\$/ && \$4 ~ /0A/) {rc=0; exit}} END {exit rc}" &&
+-                      break
+-              sleep 0.1
+-      done
+-}
+-
+ rm_addr_count()
+ {
+       mptcp_lib_get_counter "${1}" "MPTcpExtRmAddr"
+@@ -1117,7 +1099,7 @@ do_transfer()
+       fi
+       local spid=$!
+-      wait_local_port_listen "${listener_ns}" "${port}"
++      mptcp_lib_wait_local_port_listen "${listener_ns}" "${port}"
+       extra_cl_args="$extra_args $extra_cl_args"
+       if [ "$test_linkfail" -eq 0 ];then
+--- a/tools/testing/selftests/net/mptcp/mptcp_lib.sh
++++ b/tools/testing/selftests/net/mptcp/mptcp_lib.sh
+@@ -274,3 +274,21 @@ mptcp_lib_events() {
+       ip netns exec "${ns}" ./pm_nl_ctl events >> "${evts}" 2>&1 &
+       pid=$!
+ }
++
++# $1: ns, $2: port
++mptcp_lib_wait_local_port_listen() {
++      local listener_ns="${1}"
++      local port="${2}"
++
++      local port_hex
++      port_hex="$(printf "%04X" "${port}")"
++
++      local _
++      for _ in $(seq 10); do
++              ip netns exec "${listener_ns}" cat /proc/net/tcp* | \
++                      awk "BEGIN {rc=1} {if (\$2 ~ /:${port_hex}\$/ && \$4 ~ /0A/) \
++                           {rc=0; exit}} END {exit rc}" &&
++                      break
++              sleep 0.1
++      done
++}
+--- a/tools/testing/selftests/net/mptcp/simult_flows.sh
++++ b/tools/testing/selftests/net/mptcp/simult_flows.sh
+@@ -123,23 +123,6 @@ setup()
+       grep -q ' kmemleak_init$\| lockdep_init$\| kasan_init$\| prove_locking$' /proc/kallsyms && slack=$((slack+550))
+ }
+-# $1: ns, $2: port
+-wait_local_port_listen()
+-{
+-      local listener_ns="${1}"
+-      local port="${2}"
+-
+-      local port_hex i
+-
+-      port_hex="$(printf "%04X" "${port}")"
+-      for i in $(seq 10); do
+-              ip netns exec "${listener_ns}" cat /proc/net/tcp* | \
+-                      awk "BEGIN {rc=1} {if (\$2 ~ /:${port_hex}\$/ && \$4 ~ /0A/) {rc=0; exit}} END {exit rc}" &&
+-                      break
+-              sleep 0.1
+-      done
+-}
+-
+ do_transfer()
+ {
+       local cin=$1
+@@ -179,7 +162,7 @@ do_transfer()
+                               0.0.0.0 < "$sin" > "$sout" &
+       local spid=$!
+-      wait_local_port_listen "${ns3}" "${port}"
++      mptcp_lib_wait_local_port_listen "${ns3}" "${port}"
+       timeout ${timeout_test} \
+               ip netns exec ${ns1} \
index bc415622f4bb381519e6040fda290f9f416c0047..78ff0b10538246618a3e59af1f85cb2df38b0ed6 100644 (file)
@@ -360,3 +360,9 @@ io_uring-net-fix-accept-multishot-handling.patch
 cpufreq-reference-count-policy-in-cpufreq_update_limits.patch
 kbuild-add-fno-builtin-wcslen.patch
 md-fix-mddev-uaf-while-iterating-all_mddevs-list.patch
+mptcp-sockopt-fix-getting-freebind-transparent.patch
+selftests-mptcp-add-mptcp_lib_wait_local_port_listen.patch
+misc-pci_endpoint_test-avoid-issue-of-interrupts-remaining-after-request_irq-error.patch
+misc-pci_endpoint_test-fix-displaying-irq_type-after-request_irq-error.patch
+misc-pci_endpoint_test-fix-irq_type-to-convey-the-correct-type.patch
+fix-mmu-notifiers-for-range-based-invalidates.patch