Messages in this thread Patch in this message | | | From | Paul Moore <> | Subject | Re: v3.19-rc2: crashes during boot (syslog-ng, rpcbind ...) | Date | Tue, 30 Dec 2014 13:46:24 -0500 |
| |
On Tuesday, December 30, 2014 09:11:32 AM Paul Moore wrote: > On Monday, December 29, 2014 09:18:44 PM Toralf Förster wrote: > > On 12/29/2014 08:41 PM, Paul Moore wrote: > > > To help verify that I'm heading down the right path, could you share > > > your audit configuration as well? If that's not possible, can you at > > > least confirm that you using a few audit directory watches? > > > > Well, it is just a victim system for trinity - but I did not configured > > auditd in a special manner - so it is just the plain default configuration > > of Gentoo: > > Okay, thanks for the information; the file related syscall watches are > likely what triggered the problem code. Until I've got the fix sorted out, > removing the syscall watches or just disabling auditd from starting at boot > should workaround the problem.
I still want to go over the below patch a bit more to check a few things, but it solves the problem for me and I believe it should solve the problem you are seeing as well. Can you give it a try and let me know what happens?
diff --git a/kernel/auditsc.c b/kernel/auditsc.c index 287b3d3..d834770 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c @@ -72,6 +72,8 @@ #include <linux/fs_struct.h> #include <linux/compat.h> #include <linux/ctype.h> +#include <linux/string.h> +#include <uapi/linux/limits.h> #include "audit.h" @@ -1862,7 +1864,7 @@ void __audit_inode(struct filename *name, const struct dentry *dentry, list_for_each_entry_reverse(n, &context->names_list, list) { /* does the name pointer match? */ - if (!n->name || n->name->name != name->name) + if (!n->name || strcmp(n->name->name, name->name)) continue; /* match the correct record type */ @@ -1881,14 +1883,39 @@ out_alloc: n = audit_alloc_name(context, AUDIT_TYPE_UNKNOWN); if (!n) return; - if (name) - /* since name is not NULL we know there is already a matching - * name record, see audit_getname(), so there must be a type - * mismatch; reuse the string path since the original name - * record will keep the string valid until we free it in - * audit_free_names() */ - n->name = name; + /* unfortunately, while we may have a path name to record with the + * inode, we can't always rely on the string lasting until the end of + * the syscall so we need to create our own copy, it may fail due to + * memory allocation issues, but we do our best */ + if (name) { + /* we can't use getname_kernel() due to size limits */ + struct filename *new = __getname(); + if (unlikely(!new)) + goto out; + + memset(new, 0, sizeof(*new)); + if ((strlen(name->name) + 1) <= (PATH_MAX - sizeof(*new))) { + char *new_name = (char *)(new) + sizeof(*new); + new->name = new_name; + new->separate = false; + } else { + /* this looks odd, but is due to final_putname() */ + struct filename *new2; + new2 = kzalloc(sizeof(*new2), GFP_KERNEL); + if (unlikely(!new2)) { + __putname(new); + goto out; + } + new2->name = (char *)new; + new = new2; + new->separate = true; + } + strcpy((char *)new->name, name->name); + new->aname = n; + n->name = new; + n->name_put = true; + } out: if (parent) { n->name_len = n->name ? parent_len(n->name->name) : AUDIT_NAME_FULL; -- paul moore www.paul-moore.com
| |