[PATCH RFC] coredump: virtualize core dump path configuration

From: Konstantin Khlebnikov
Date: Sat Feb 25 2017 - 07:56:38 EST

This patch adds per-mount-namespace core dump pattern.

Kernel writes coredump in chroot/container where application is
executed or starts pipe helper in the same chroot according to
pattern set by sysctl "kernel.core_pattern". This configuration is
global and this sysctl couldn't be extended without breaking anything.

This patch adds second sysctl "kernel.core_pattern_ns" which overrides
global configuration for tasks in current mount namespace.
Resetting it to empty string reverts core dumps back to global pattern.

New namespace gets a copy of this configuration from parent.

Signed-off-by: Konstantin Khlebnikov <khlebnikov@xxxxxxxxxxxxxx>
---
Documentation/sysctl/kernel.txt | 9 +++++++++
fs/coredump.c | 27 ++++++++++++++++++++++++++-
fs/mount.h | 1 +
fs/namespace.c | 12 ++++++++++++
kernel/sysctl.c | 26 +++++++++++++++++++++++++-
5 files changed, 73 insertions(+), 2 deletions(-)

diff --git a/Documentation/sysctl/kernel.txt b/Documentation/sysctl/kernel.txt
index a32b4b748644..769aa00df898 100644
--- a/Documentation/sysctl/kernel.txt
+++ b/Documentation/sysctl/kernel.txt
@@ -26,6 +26,7 @@ show up in /proc/sys/kernel:
- callhome [ S390 only ]
- cap_last_cap
- core_pattern
+- core_pattern_ns
- core_pipe_limit
- core_uses_pid
- ctrl-alt-del
@@ -219,6 +220,14 @@ core_pattern is used to specify a core dumpfile pattern name.

==============================================================

+core_pattern_ns:
+
+This sysctl has the same format as core_pattern. Any non-empty string
+set here overrides core_pattern for tasks in current mount namespace.
+New mount namespace gets a copy of this configuration from parent.
+
+==============================================================
+
core_pipe_limit:

This sysctl is only applicable when core_pattern is configured to pipe
diff --git a/fs/coredump.c b/fs/coredump.c
index ae6b05629ca1..9c789496d4c6 100644
--- a/fs/coredump.c
+++ b/fs/coredump.c
@@ -45,6 +45,7 @@

#include <trace/events/task.h>
#include "internal.h"
+#include "mount.h"

#include <trace/events/sched.h>

@@ -180,6 +181,30 @@ static int cn_print_exe_file(struct core_name *cn)
return ret;
}

+char *namespace_core_pattern(bool alloc)
+{
+ struct mnt_namespace *ns = current->nsproxy->mnt_ns;
+
+ if (!ns->core_pattern && alloc) {
+ char *new = kzalloc(CORENAME_MAX_SIZE, GFP_KERNEL);
+
+ if (new && cmpxchg(&ns->core_pattern, NULL, new))
+ kfree(new);
+ }
+
+ return ns->core_pattern;
+}
+
+static char *current_core_pattern(void)
+{
+ struct mnt_namespace *ns = current->nsproxy->mnt_ns;
+
+ if (ns->core_pattern && ns->core_pattern[0])
+ return ns->core_pattern;
+
+ return core_pattern;
+}
+
/* format_corename will inspect the pattern parameter, and output a
* name into corename, which must have space for at least
* CORENAME_MAX_SIZE bytes plus one byte for the zero terminator.
@@ -187,7 +212,7 @@ static int cn_print_exe_file(struct core_name *cn)
static int format_corename(struct core_name *cn, struct coredump_params *cprm)
{
const struct cred *cred = current_cred();
- const char *pat_ptr = core_pattern;
+ const char *pat_ptr = current_core_pattern();
int ispipe = (*pat_ptr == '|');
int pid_in_pattern = 0;
int err = 0;
diff --git a/fs/mount.h b/fs/mount.h
index 2c856fc47ae3..894bca887104 100644
--- a/fs/mount.h
+++ b/fs/mount.h
@@ -16,6 +16,7 @@ struct mnt_namespace {
u64 event;
unsigned int mounts; /* # of mounts in the namespace */
unsigned int pending_mounts;
+ char *core_pattern;
};

struct mnt_pcp {
diff --git a/fs/namespace.c b/fs/namespace.c
index 487ba30bb5c6..a8dd58eb10da 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -24,6 +24,7 @@
#include <linux/magic.h>
#include <linux/bootmem.h>
#include <linux/task_work.h>
+#include <linux/binfmts.h> /* CORENAME_MAX_SIZE */
#include "pnode.h"
#include "internal.h"

@@ -2828,6 +2829,7 @@ static void free_mnt_ns(struct mnt_namespace *ns)
ns_free_inum(&ns->ns);
dec_mnt_namespaces(ns->ucounts);
put_user_ns(ns->user_ns);
+ kfree(ns->core_pattern);
kfree(ns);
}

@@ -2872,6 +2874,7 @@ static struct mnt_namespace *alloc_mnt_ns(struct user_namespace *user_ns)
new_ns->ucounts = ucounts;
new_ns->mounts = 0;
new_ns->pending_mounts = 0;
+ new_ns->core_pattern = NULL;
return new_ns;
}

@@ -2899,6 +2902,15 @@ struct mnt_namespace *copy_mnt_ns(unsigned long flags, struct mnt_namespace *ns,
if (IS_ERR(new_ns))
return new_ns;

+ if (ns->core_pattern) {
+ new_ns->core_pattern = kmemdup(ns->core_pattern,
+ CORENAME_MAX_SIZE, GFP_KERNEL);
+ if (!new_ns->core_pattern) {
+ free_mnt_ns(new_ns);
+ return ERR_PTR(-ENOMEM);
+ }
+ }
+
namespace_lock();
/* First pass: copy the tree topology */
copy_flags = CL_COPY_UNBINDABLE | CL_EXPIRE;
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index 1aea594a54db..9e66daf1e236 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -483,6 +483,13 @@ static struct ctl_table kern_table[] = {
.proc_handler = proc_dostring_coredump,
},
{
+ .procname = "core_pattern_ns",
+ .data = NULL,
+ .maxlen = CORENAME_MAX_SIZE,
+ .mode = 0644,
+ .proc_handler = proc_dostring_coredump,
+ },
+ {
.procname = "core_pipe_limit",
.data = &core_pipe_limit,
.maxlen = sizeof(unsigned int),
@@ -2408,10 +2415,27 @@ static int proc_dointvec_minmax_coredump(struct ctl_table *table, int write,
}

#ifdef CONFIG_COREDUMP
+extern char *namespace_core_pattern(bool alloc);
+
static int proc_dostring_coredump(struct ctl_table *table, int write,
void __user *buffer, size_t *lenp, loff_t *ppos)
{
- int error = proc_dostring(table, write, buffer, lenp, ppos);
+ struct ctl_table tmp_table;
+ char empty[] = "";
+ int error;
+
+ if (!table->data) {
+ tmp_table = *table;
+ table = &tmp_table;
+ table->data = namespace_core_pattern(write);
+ if (!table->data) {
+ if (write)
+ return -ENOMEM;
+ table->data = empty;
+ }
+ }
+
+ error = proc_dostring(table, write, buffer, lenp, ppos);
if (!error)
validate_coredump_safety();
return error;