libbpf: Fix the case of running as non-root with capabilities
PATCH v2 1/1 libbpf: Fix the case of running as non-root with capabilities
From: Jon Doron <[email protected]>
When running rootless with special capabilities like:
FOWNER / DAC_OVERRIDE / DAC_READ_SEARCH
The "access" API will not make the proper check if there is really access to a file or not.
From the access man page:
"
The check is done using the calling process's real UID and GID, rather than the effective IDs as is done when actually attempting an operation (e.g., open(2)) on the file. Similarly, for the root user, the check uses the set of permitted capabilities rather than the set of effective capabilities; ***and for non-root users, the check uses an empty set of capabilities.***
"
What that means is that for non-root user the access API will not do the proper validation if the process really has permission to a file or not.
To resolve this this patch replaces all the access API calls with stat.
--- tools/lib/bpf/btf.c | 3 ++- tools/lib/bpf/libbpf.c | 11 ++++++++--- tools/lib/bpf/usdt.c | 4 +++- 3 files changed, 13 insertions(+), 5 deletions(-)
tools/lib/bpf/btf.c +2 -1
@@ -4663,13 +4663,14 @@ struct btf *btf__load_vmlinux_btf(void)struct utsname buf;struct btf *btf;int i, err; +struct stat sb;uname (&buf);for (i =0 ; i <ARRAY_SIZE (locations); i++) {snprintf (path,PATH_MAX , locations[i].path_fmt, buf.release); -if (access (path,R_OK )) +if (stat (path, &sb))continue ;if (locations[i].raw_btf)
tools/lib/bpf/libbpf.c +8 -3
@@ -875,8 +875,9 @@ __u32 get_kernel_version(void)const char *ubuntu_kver_file ="/proc/version_signature" ;__u32 major, minor, patch;struct utsname info; +struct stat sb; -if (access (ubuntu_kver_file,R_OK ) ==0 ) { +if (stat (ubuntu_kver_file, &sb) ==0 ) {FILE *f; f =fopen (ubuntu_kver_file,"r" );
@@ -9877,9 +9878,10 @@ static int append_to_file(const char *file, const char *fmt, ...)static bool use_debugfs (void ) {static int has_debugfs =-1 ; +struct stat sb;if (has_debugfs <0 ) - has_debugfs =access (DEBUGFS ,F_OK ) ==0 ; + has_debugfs =stat (DEBUGFS , &sb) ==0 ;return has_debugfs ==1 ; }
@@ -10681,6 +10683,7 @@ static int resolve_full_path(const char *file, char *result, size_t result_sz)for (s = search_paths[i]; s !=NULL ; s =strchr (s,':' )) {char *next_path;int seg_len; +struct stat sb;if (s[0 ] ==':' ) s++;
@@ -10690,7 +10693,9 @@ static int resolve_full_path(const char *file, char *result, size_t result_sz)continue ;snprintf (result, result_sz,"%.*s/%s" , seg_len, s, file);/* ensure it is an executable file/link */ -if (access (result,R_OK |X_OK ) <0 ) +if (stat (result, &sb) <0 ) +continue ; +if ((sb.st_mode & (S_IROTH |S_IXOTH )) != (S_IROTH |S_IXOTH ))continue ;pr_debug ("resolved '%s' to '%s'\n" , file, result);return 0 ;
tools/lib/bpf/usdt.c +3 -1
@@ -7,6 +7,7 @@#include <libelf.h> #include <gelf.h> #include <unistd.h> +#include <sys/stat.h> #include <linux/ptrace.h> #include <linux/kernel.h>
@@ -257,6 +258,7 @@ struct usdt_manager *usdt_manager_new(struct bpf_object *obj)static const char *ref_ctr_sysfs_path ="/sys/bus/event_source/devices/uprobe/format/ref_ctr_offset" ;struct usdt_manager *man;struct bpf_map *specs_map, *ip_to_spec_id_map; +struct stat sb; specs_map =bpf_object__find_map_by_name (obj,"__bpf_usdt_specs" ); ip_to_spec_id_map =bpf_object__find_map_by_name (obj,"__bpf_usdt_ip_to_spec_id" );
@@ -282,7 +284,7 @@ struct usdt_manager *usdt_manager_new(struct bpf_object *obj)* If this is not supported, USDTs with semaphores will not be supported. * Added in: a6ca88b241d5 ("trace_uprobe: support reference counter in fd-based uprobe") */ - man->has_sema_refcnt =access (ref_ctr_sysfs_path,F_OK ) ==0 ; + man->has_sema_refcnt =stat (ref_ctr_sysfs_path, &sb) ==0 ;return man; }
Signature · 1 line
2.37.3
Quoted 22 lines · Jon Doron
From: Jon Doron <[email protected]>
When running rootless with special capabilities like:
FOWNER /DAC_OVERRIDE/DAC_READ_SEARCHThe "access" API will not make the proper check if there is really access to a file or not.
From the access man page:
"
The check is done using the calling process's real UID and GID, rather than the effective IDs as is done when actually attempting an operation (e.g., open(2)) on the file. Similarly, for the root user, the check uses the set of permitted capabilities rather than the set of effective capabilities; ***and for non-root users, the check uses an empty set of capabilities.***
"What that means is that for non-root user the access API will not do the proper validation if the process really has permission to a file or not.
Quoted 9 lines
To resolve this this patch replaces all the access API calls with stat.
Signed-off-by Jon Doron <[email protected]>--- tools/lib/bpf/btf.c | 3 ++- tools/lib/bpf/libbpf.c | 11 ++++++++--- tools/lib/bpf/usdt.c | 4 +++- 3 files changed, 13 insertions(+), 5 deletions(-)
This patch doesn't apply cleanly onto bpf-next, please rebase. But also see comments below.
Quoted 56 lines
tools/lib/bpf/btf.c +2
-1@@ -4663,13 +4663,14 @@ struct btf *btf__load_vmlinux_btf(void)struct utsname buf;struct btf *btf;int i, err; +struct stat sb;uname (&buf);for (i =0 ; i <ARRAY_SIZE (locations); i++) {snprintf (path,PATH_MAX , locations[i].path_fmt, buf.release); -if (access (path,R_OK )) +if (stat (path, &sb))continue ;if (locations[i].raw_btf)tools/lib/bpf/libbpf.c +5
-2@@ -875,8 +875,9 @@ __u32 get_kernel_version(void)const char *ubuntu_kver_file ="/proc/version_signature" ;__u32 major, minor, patch;struct utsname info; +struct stat sb; -if (access (ubuntu_kver_file,R_OK ) ==0 ) { +if (stat (ubuntu_kver_file, &sb) ==0 ) {FILE *f; f =fopen (ubuntu_kver_file,"r" );@@ -9877,9 +9878,10 @@ static int append_to_file(const char *file, const char *fmt, ...)static bool use_debugfs (void ) {static int has_debugfs =-1 ; +struct stat sb;if (has_debugfs <0 ) - has_debugfs =access (DEBUGFS ,F_OK ) ==0 ; + has_debugfs =stat (DEBUGFS , &sb) ==0 ;return has_debugfs ==1 ; }@@ -10681,6 +10683,7 @@ static int resolve_full_path(const char *file, char *result, size_t result_sz)for (s = search_paths[i]; s !=NULL ; s =strchr (s,':' )) {char *next_path;int seg_len; +struct stat sb;if (s[0 ] ==':' ) s++;
tools/lib/bpf/libbpf.c static int resolve_full_path(const char *file, char *result, size_t result_sz) · line 10698
Quoted 7 lines
tools/lib/bpf/libbpf.c +3
-1@@ -10690,7 +10693,9 @@ static int resolve_full_path(const char *file, char *result, size_t result_sz)continue ;snprintf (result, result_sz,"%.*s/%s" , seg_len, s, file);/* ensure it is an executable file/link */ -if (access (result,R_OK |X_OK ) <0 ) +if (stat (result, &sb) <0 ) +continue ; +if ((sb.st_mode & (S_IROTH |S_IXOTH )) != (S_IROTH |S_IXOTH ))
So this bit look a bit suspicious. Why are we only checking others bits? And in general, it seems pretty unfortunate to check access bits manually.
So two thoughts.
1. Have you tried using faccessat2() with AT_EACCESS flag? It seems like that (apart from all the stuff in BUGS section in man page, sigh...) delegate the actual permission checks using effective user to kernel. This seems like a bit better and safer way?
2. Whichever way we go, I think let's add an internal wrapper that will hide struct stat (which we almost never use) and will simulate access just with proper checks? Like libbpf_faccess() or something along those lines? You can expose it in libbpf_internal.h
Quoted 35 lines
tools/lib/bpf/libbpf.c +0
-0continue ;pr_debug ("resolved '%s' to '%s'\n" , file, result);return 0 ;tools/lib/bpf/usdt.c +3
-2@@ -7,6 +7,7 @@#include <libelf.h> #include <gelf.h> #include <unistd.h> +#include <sys/stat.h> #include <linux/ptrace.h> #include <linux/kernel.h> @@ -257,6 +258,7 @@ struct usdt_manager *usdt_manager_new(struct bpf_object *obj)static const char *ref_ctr_sysfs_path ="/sys/bus/event_source/devices/uprobe/format/ref_ctr_offset" ;struct usdt_manager *man;struct bpf_map *specs_map, *ip_to_spec_id_map; +struct stat sb; specs_map =bpf_object__find_map_by_name (obj,"__bpf_usdt_specs" ); ip_to_spec_id_map =bpf_object__find_map_by_name (obj,"__bpf_usdt_ip_to_spec_id" );@@ -282,7 +284,7 @@ struct usdt_manager *usdt_manager_new(struct bpf_object *obj)* If this is not supported, USDTs with semaphores will not be supported. * Added in: a6ca88b241d5 ("trace_uprobe: support reference counter in fd-based uprobe") */ - man->has_sema_refcnt =access (ref_ctr_sysfs_path,F_OK ) ==0 ; + man->has_sema_refcnt =stat (ref_ctr_sysfs_path, &sb) ==0 ;return man; } --2.37.3