5.15-stable patches

added patches:
	net-sunrpc-remap-eperm-in-case-of-connection-failure-in-xs_tcp_setup_socket.patch
	x86-mm-fix-pti-for-i386-some-more.patch

diff --git a/queue-5.15/net-sunrpc-remap-eperm-in-case-of-connection-failure-in-xs_tcp_setup_socket.patch b/queue-5.15/net-sunrpc-remap-eperm-in-case-of-connection-failure-in-xs_tcp_setup_socket.patch
new file mode 100644 (file)
index 0000000..2c976a7
--- /dev/null
+++ b/queue-5.15/net-sunrpc-remap-eperm-in-case-of-connection-failure-in-xs_tcp_setup_socket.patch
@@ -0,0 +1,59 @@
+From 626dfed5fa3bfb41e0dffd796032b555b69f9cde Mon Sep 17 00:00:00 2001
+From: Daniel Borkmann <daniel@iogearbox.net>
+Date: Thu, 4 Jul 2024 08:41:57 +0200
+Subject: net, sunrpc: Remap EPERM in case of connection failure in xs_tcp_setup_socket
+
+From: Daniel Borkmann <daniel@iogearbox.net>
+
+commit 626dfed5fa3bfb41e0dffd796032b555b69f9cde upstream.
+
+When using a BPF program on kernel_connect(), the call can return -EPERM. This
+causes xs_tcp_setup_socket() to loop forever, filling up the syslog and causing
+the kernel to potentially freeze up.
+
+Neil suggested:
+
+  This will propagate -EPERM up into other layers which might not be ready
+  to handle it. It might be safer to map EPERM to an error we would be more
+  likely to expect from the network system - such as ECONNREFUSED or ENETDOWN.
+
+ECONNREFUSED as error seems reasonable. For programs setting a different error
+can be out of reach (see handling in 4fbac77d2d09) in particular on kernels
+which do not have f10d05966196 ("bpf: Make BPF_PROG_RUN_ARRAY return -err
+instead of allow boolean"), thus given that it is better to simply remap for
+consistent behavior. UDP does handle EPERM in xs_udp_send_request().
+
+Fixes: d74bad4e74ee ("bpf: Hooks for sys_connect")
+Fixes: 4fbac77d2d09 ("bpf: Hooks for sys_bind")
+Co-developed-by: Lex Siegel <usiegl00@gmail.com>
+Signed-off-by: Lex Siegel <usiegl00@gmail.com>
+Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
+Cc: Neil Brown <neilb@suse.de>
+Cc: Trond Myklebust <trondmy@kernel.org>
+Cc: Anna Schumaker <anna@kernel.org>
+Link: https://github.com/cilium/cilium/issues/33395
+Link: https://lore.kernel.org/bpf/171374175513.12877.8993642908082014881@noble.neil.brown.name
+Link: https://patch.msgid.link/9069ec1d59e4b2129fc23433349fd5580ad43921.1720075070.git.daniel@iogearbox.net
+Signed-off-by: Paolo Abeni <pabeni@redhat.com>
+Signed-off-by: Hugo SIMELIERE <hsimeliere.opensource@witekio.com>
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ net/sunrpc/xprtsock.c |    7 +++++++
+ 1 file changed, 7 insertions(+)
+
+--- a/net/sunrpc/xprtsock.c
++++ b/net/sunrpc/xprtsock.c
+@@ -2335,6 +2335,13 @@ static void xs_tcp_setup_socket(struct w
+       case -EALREADY:
+               xprt_unlock_connect(xprt, transport);
+               return;
++      case -EPERM:
++              /* Happens, for instance, if a BPF program is preventing
++               * the connect. Remap the error so upper layers can better
++               * deal with it.
++               */
++              status = -ECONNREFUSED;
++              fallthrough;
+       case -EINVAL:
+               /* Happens, for instance, if the user specified a link
+                * local IPv6 address without a scope-id.

diff --git a/queue-5.15/series b/queue-5.15/series
index 236d6ffcb405e02e1baae182b0f146a49df0cf2a..35a86c809c745b2efe071139d19655f880aa5bec 100644 (file)
--- a/queue-5.15/series
+++ b/queue-5.15/series
@@ -208,3 +208,5 @@ net-more-strict-virtio_net_hdr_gso_udp_l4-validation.patch
 net-change-maximum-number-of-udp-segments-to-128.patch
 gso-fix-dodgy-bit-handling-for-gso_udp_l4.patch
 net-drop-bad-gso-csum_start-and-offset-in-virtio_net_hdr.patch
+x86-mm-fix-pti-for-i386-some-more.patch
+net-sunrpc-remap-eperm-in-case-of-connection-failure-in-xs_tcp_setup_socket.patch

diff --git a/queue-5.15/x86-mm-fix-pti-for-i386-some-more.patch b/queue-5.15/x86-mm-fix-pti-for-i386-some-more.patch
new file mode 100644 (file)
index 0000000..8f14ca2
--- /dev/null
+++ b/queue-5.15/x86-mm-fix-pti-for-i386-some-more.patch
@@ -0,0 +1,160 @@
+From c48b5a4cf3125adb679e28ef093f66ff81368d05 Mon Sep 17 00:00:00 2001
+From: Thomas Gleixner <tglx@linutronix.de>
+Date: Tue, 6 Aug 2024 20:48:43 +0200
+Subject: x86/mm: Fix PTI for i386 some more
+
+From: Thomas Gleixner <tglx@linutronix.de>
+
+commit c48b5a4cf3125adb679e28ef093f66ff81368d05 upstream.
+
+So it turns out that we have to do two passes of
+pti_clone_entry_text(), once before initcalls, such that device and
+late initcalls can use user-mode-helper / modprobe and once after
+free_initmem() / mark_readonly().
+
+Now obviously mark_readonly() can cause PMD splits, and
+pti_clone_pgtable() doesn't like that much.
+
+Allow the late clone to split PMDs so that pagetables stay in sync.
+
+[peterz: Changelog and comments]
+Reported-by: Guenter Roeck <linux@roeck-us.net>
+Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
+Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
+Tested-by: Guenter Roeck <linux@roeck-us.net>
+Link: https://lkml.kernel.org/r/20240806184843.GX37996@noisy.programming.kicks-ass.net
+Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
+---
+ arch/x86/mm/pti.c |   45 +++++++++++++++++++++++++++++----------------
+ 1 file changed, 29 insertions(+), 16 deletions(-)
+
+--- a/arch/x86/mm/pti.c
++++ b/arch/x86/mm/pti.c
+@@ -241,7 +241,7 @@ static pmd_t *pti_user_pagetable_walk_pm
+  *
+  * Returns a pointer to a PTE on success, or NULL on failure.
+  */
+-static pte_t *pti_user_pagetable_walk_pte(unsigned long address)
++static pte_t *pti_user_pagetable_walk_pte(unsigned long address, bool late_text)
+ {
+       gfp_t gfp = (GFP_KERNEL | __GFP_NOTRACK | __GFP_ZERO);
+       pmd_t *pmd;
+@@ -251,10 +251,15 @@ static pte_t *pti_user_pagetable_walk_pt
+       if (!pmd)
+               return NULL;
+ 
+-      /* We can't do anything sensible if we hit a large mapping. */
++      /* Large PMD mapping found */
+       if (pmd_large(*pmd)) {
+-              WARN_ON(1);
+-              return NULL;
++              /* Clear the PMD if we hit a large mapping from the first round */
++              if (late_text) {
++                      set_pmd(pmd, __pmd(0));
++              } else {
++                      WARN_ON_ONCE(1);
++                      return NULL;
++              }
+       }
+ 
+       if (pmd_none(*pmd)) {
+@@ -283,7 +288,7 @@ static void __init pti_setup_vsyscall(vo
+       if (!pte || WARN_ON(level != PG_LEVEL_4K) || pte_none(*pte))
+               return;
+ 
+-      target_pte = pti_user_pagetable_walk_pte(VSYSCALL_ADDR);
++      target_pte = pti_user_pagetable_walk_pte(VSYSCALL_ADDR, false);
+       if (WARN_ON(!target_pte))
+               return;
+ 
+@@ -301,7 +306,7 @@ enum pti_clone_level {
+ 
+ static void
+ pti_clone_pgtable(unsigned long start, unsigned long end,
+-                enum pti_clone_level level)
++                enum pti_clone_level level, bool late_text)
+ {
+       unsigned long addr;
+ 
+@@ -390,7 +395,7 @@ pti_clone_pgtable(unsigned long start, u
+                               return;
+ 
+                       /* Allocate PTE in the user page-table */
+-                      target_pte = pti_user_pagetable_walk_pte(addr);
++                      target_pte = pti_user_pagetable_walk_pte(addr, late_text);
+                       if (WARN_ON(!target_pte))
+                               return;
+ 
+@@ -452,7 +457,7 @@ static void __init pti_clone_user_shared
+               phys_addr_t pa = per_cpu_ptr_to_phys((void *)va);
+               pte_t *target_pte;
+ 
+-              target_pte = pti_user_pagetable_walk_pte(va);
++              target_pte = pti_user_pagetable_walk_pte(va, false);
+               if (WARN_ON(!target_pte))
+                       return;
+ 
+@@ -475,7 +480,7 @@ static void __init pti_clone_user_shared
+       start = CPU_ENTRY_AREA_BASE;
+       end   = start + (PAGE_SIZE * CPU_ENTRY_AREA_PAGES);
+ 
+-      pti_clone_pgtable(start, end, PTI_CLONE_PMD);
++      pti_clone_pgtable(start, end, PTI_CLONE_PMD, false);
+ }
+ #endif /* CONFIG_X86_64 */
+ 
+@@ -492,11 +497,11 @@ static void __init pti_setup_espfix64(vo
+ /*
+  * Clone the populated PMDs of the entry text and force it RO.
+  */
+-static void pti_clone_entry_text(void)
++static void pti_clone_entry_text(bool late)
+ {
+       pti_clone_pgtable((unsigned long) __entry_text_start,
+                         (unsigned long) __entry_text_end,
+-                        PTI_LEVEL_KERNEL_IMAGE);
++                        PTI_LEVEL_KERNEL_IMAGE, late);
+ }
+ 
+ /*
+@@ -571,7 +576,7 @@ static void pti_clone_kernel_text(void)
+        * pti_set_kernel_image_nonglobal() did to clear the
+        * global bit.
+        */
+-      pti_clone_pgtable(start, end_clone, PTI_LEVEL_KERNEL_IMAGE);
++      pti_clone_pgtable(start, end_clone, PTI_LEVEL_KERNEL_IMAGE, false);
+ 
+       /*
+        * pti_clone_pgtable() will set the global bit in any PMDs
+@@ -638,8 +643,15 @@ void __init pti_init(void)
+ 
+       /* Undo all global bits from the init pagetables in head_64.S: */
+       pti_set_kernel_image_nonglobal();
++
+       /* Replace some of the global bits just for shared entry text: */
+-      pti_clone_entry_text();
++      /*
++       * This is very early in boot. Device and Late initcalls can do
++       * modprobe before free_initmem() and mark_readonly(). This
++       * pti_clone_entry_text() allows those user-mode-helpers to function,
++       * but notably the text is still RW.
++       */
++      pti_clone_entry_text(false);
+       pti_setup_espfix64();
+       pti_setup_vsyscall();
+ }
+@@ -656,10 +668,11 @@ void pti_finalize(void)
+       if (!boot_cpu_has(X86_FEATURE_PTI))
+               return;
+       /*
+-       * We need to clone everything (again) that maps parts of the
+-       * kernel image.
++       * This is after free_initmem() (all initcalls are done) and we've done
++       * mark_readonly(). Text is now NX which might've split some PMDs
++       * relative to the early clone.
+        */
+-      pti_clone_entry_text();
++      pti_clone_entry_text(true);
+       pti_clone_kernel_text();
+ 
+       debug_checkwx_user();