kbuild: rust: keep Rust objects out of Clang LTO with inline helpersgrafted

Under `CONFIG_LTO_CLANG` + `CONFIG_RUST_INLINE_HELPERS`, one may hit
`objtool` errors such as:

    vmlinux.o: error: objtool: _R..._3Gsp4boot+0xd6a:
    can't find jump dest instruction at .text._R..._3Gsp4boot+0x1dfd

The reason is that in such builds, the Clang invocation that compiles
the combined Rust plus helpers bitcode emits LLVM bitcode (again) --
the final code generation happens in the linker's LTO step, which the
`-mllvm` trap options passed to Clang do not reach.

This, in turn, means that unreachable traps are missing, and the
impossible paths do not merely fallthrough to the next symbol, but past
the end of their own section, since LTO builds place each function in
its own section.

Thus filter `CC_FLAGS_LTO` out of the Clang invocation, so that it always
emits machine code directly, with the traps in place.

Assisted-by: LLM
Cc: Gary Guo <gary@garyguo.net>
Cc: Boqun Feng <boqun@kernel.org>
Cc: Alice Ryhl <aliceryhl@google.com>
Cc: Matthew Maurer <mmaurer@google.com>
Cc: Josh Poimboeuf <jpoimboe@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: stable@vger.kernel.org
Fixes: 3a2486cc1da5 ("kbuild: rust: provide an option to inline C helpers into Rust")
Acked-by: Gary Guo <gary@garyguo.net>
Link: https://patch.msgid.link/20260816133233.197500-2-ojeda@kernel.org
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>

1 files changed, 206 insertions, 0 deletions

diff --git a/security/apparmor/resource.c b/security/apparmor/resource.c
new file mode 100644
index 000000000..64212b39b
--- /dev/null
+++ b/security/apparmor/resource.c
@@ -0,0 +1,206 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * AppArmor security module
+ *
+ * This file contains AppArmor resource mediation and attachment
+ *
+ * Copyright (C) 1998-2008 Novell/SUSE
+ * Copyright 2009-2010 Canonical Ltd.
+ */
+
+#include <linux/audit.h>
+#include <linux/security.h>
+
+#include "include/audit.h"
+#include "include/cred.h"
+#include "include/resource.h"
+#include "include/policy.h"
+
+/*
+ * Table of rlimit names: we generate it from resource.h.
+ */
+#include "rlim_names.h"
+
+struct aa_sfs_entry aa_sfs_entry_rlimit[] = {
+	AA_SFS_FILE_STRING("mask", AA_SFS_RLIMIT_MASK),
+	{ }
+};
+
+/* audit callback for resource specific fields */
+static void audit_cb(struct audit_buffer *ab, void *va)
+{
+	struct common_audit_data *sa = va;
+	struct apparmor_audit_data *ad = aad(sa);
+
+	audit_log_format(ab, " rlimit=%s value=%lu",
+			 rlim_names[ad->rlim.rlim], ad->rlim.max);
+	if (ad->peer) {
+		audit_log_format(ab, " peer=");
+		aa_label_xaudit(ab, labels_ns(ad->subj_label), ad->peer,
+				FLAGS_NONE, GFP_ATOMIC);
+	}
+}
+
+/**
+ * audit_resource - audit setting resource limit
+ * @subj_cred: cred setting the resource
+ * @profile: profile being enforced  (NOT NULL)
+ * @resource: rlimit being auditing
+ * @value: value being set
+ * @peer: aa_albel of the task being set
+ * @info: info being auditing
+ * @error: error value
+ *
+ * Returns: 0 or ad->error else other error code on failure
+ */
+static int audit_resource(const struct cred *subj_cred,
+			  struct aa_profile *profile, unsigned int resource,
+			  unsigned long value, struct aa_label *peer,
+			  const char *info, int error)
+{
+	DEFINE_AUDIT_DATA(ad, LSM_AUDIT_DATA_NONE, AA_CLASS_RLIMITS,
+			  OP_SETRLIMIT);
+
+	ad.subj_cred = subj_cred;
+	ad.rlim.rlim = resource;
+	ad.rlim.max = value;
+	ad.peer = peer;
+	ad.info = info;
+	ad.error = error;
+
+	return aa_audit(AUDIT_APPARMOR_AUTO, profile, &ad, audit_cb);
+}
+
+/**
+ * aa_map_resource - map compiled policy resource to internal #
+ * @resource: flattened policy resource number
+ *
+ * Returns: resource # for the current architecture.
+ *
+ * rlimit resource can vary based on architecture, map the compiled policy
+ * resource # to the internal representation for the architecture.
+ */
+int aa_map_resource(int resource)
+{
+	return rlim_map[resource];
+}
+
+static int profile_setrlimit(const struct cred *subj_cred,
+			     struct aa_profile *profile, unsigned int resource,
+			     struct rlimit *new_rlim)
+{
+	struct aa_ruleset *rules = profile->label.rules[0];
+	int e = 0;
+
+	if (rules->rlimits.mask & (1 << resource) && new_rlim->rlim_max >
+	    rules->rlimits.limits[resource].rlim_max)
+		e = -EACCES;
+	return audit_resource(subj_cred, profile, resource, new_rlim->rlim_max,
+			      NULL, NULL, e);
+}
+
+/**
+ * aa_task_setrlimit - test permission to set an rlimit
+ * @subj_cred: cred setting the limit
+ * @label: label confining the task  (NOT NULL)
+ * @task: task the resource is being set on
+ * @resource: the resource being set
+ * @new_rlim: the new resource limit  (NOT NULL)
+ *
+ * Control raising the processes hard limit.
+ *
+ * Returns: 0 or error code if setting resource failed
+ */
+int aa_task_setrlimit(const struct cred *subj_cred, struct aa_label *label,
+		      struct task_struct *task,
+		      unsigned int resource, struct rlimit *new_rlim)
+{
+	struct aa_profile *profile;
+	struct aa_label *peer;
+	int error = 0;
+
+	rcu_read_lock();
+	peer = aa_get_newest_cred_label(__task_cred(task));
+	rcu_read_unlock();
+
+	/* TODO: extend resource control to handle other (non current)
+	 * profiles.  AppArmor rules currently have the implicit assumption
+	 * that the task is setting the resource of a task confined with
+	 * the same profile or that the task setting the resource of another
+	 * task has CAP_SYS_RESOURCE.
+	 */
+
+	if (label != peer &&
+	    aa_capable(subj_cred, label, CAP_SYS_RESOURCE, CAP_OPT_NOAUDIT) != 0)
+		error = fn_for_each(label, profile,
+				audit_resource(subj_cred, profile, resource,
+					       new_rlim->rlim_max, peer,
+					       "cap_sys_resource", -EACCES));
+	else
+		error = fn_for_each_confined(label, profile,
+				profile_setrlimit(subj_cred, profile, resource,
+						  new_rlim));
+	aa_put_label(peer);
+
+	return error;
+}
+
+/**
+ * __aa_transition_rlimits - apply new profile rlimits
+ * @old_l: old label on task  (NOT NULL)
+ * @new_l: new label with rlimits to apply  (NOT NULL)
+ */
+void __aa_transition_rlimits(struct aa_label *old_l, struct aa_label *new_l)
+{
+	unsigned int mask = 0;
+	struct rlimit *rlim, *initrlim;
+	struct aa_profile *old, *new;
+	struct label_it i;
+
+	old = labels_profile(old_l);
+	new = labels_profile(new_l);
+
+	/* for any rlimits the profile controlled, reset the soft limit
+	 * to the lesser of the tasks hard limit and the init tasks soft limit
+	 */
+	label_for_each_confined(i, old_l, old) {
+		struct aa_ruleset *rules = old->label.rules[0];
+		if (rules->rlimits.mask) {
+			int j;
+
+			for (j = 0, mask = 1; j < RLIM_NLIMITS; j++,
+				     mask <<= 1) {
+				if (rules->rlimits.mask & mask) {
+					rlim = current->signal->rlim + j;
+					initrlim = init_task.signal->rlim + j;
+					rlim->rlim_cur = min(rlim->rlim_max,
+							    initrlim->rlim_cur);
+				}
+			}
+		}
+	}
+
+	/* set any new hard limits as dictated by the new profile */
+	label_for_each_confined(i, new_l, new) {
+		struct aa_ruleset *rules = new->label.rules[0];
+		int j;
+
+		if (!rules->rlimits.mask)
+			continue;
+		for (j = 0, mask = 1; j < RLIM_NLIMITS; j++, mask <<= 1) {
+			if (!(rules->rlimits.mask & mask))
+				continue;
+
+			rlim = current->signal->rlim + j;
+			rlim->rlim_max = min(rlim->rlim_max,
+					     rules->rlimits.limits[j].rlim_max);
+			/* soft limit should not exceed hard limit */
+			rlim->rlim_cur = min(rlim->rlim_cur, rlim->rlim_max);
+			if (j == RLIMIT_CPU &&
+			    rlim->rlim_cur != RLIM_INFINITY &&
+			    IS_ENABLED(CONFIG_POSIX_TIMERS))
+				(void) update_rlimit_cpu(current->group_leader,
+							 rlim->rlim_cur);
+		}
+	}
+}