kmsg: honor dmesg_restrict sysctl on /dev/kmsg
PATCH 1/1 kmsg: honor dmesg_restrict sysctl on /dev/kmsg
To fix /dev/kmsg, let's compare the existing interfaces and what they allow:
- /proc/kmsg allows:
- open (SYSLOG_ACTION_OPEN) if CAP_SYSLOG since it uses a destructive single-reader interface (SYSLOG_ACTION_READ).
- everything, after an open.
- syslog syscall allows:
- anything, if CAP_SYSLOG.
- SYSLOG_ACTION_READ_ALL and SYSLOG_ACTION_SIZE_BUFFER, if dmesg_restrict==0.
- nothing else (EPERM).
The use-cases were:
- dmesg(1) needs to do non-destructive SYSLOG_ACTION_READ_ALLs.
- sysklog(1) needs to open /proc/kmsg, drop privs, and still issue the destructive SYSLOG_ACTION_READs.
AIUI, dmesg(1) is moving to /dev/kmsg, and systemd-journald doesn't clear the ring buffer.
Based on the comments in devkmsg_llseek, it sounds like actions besides reading aren't going to be supported by /dev/kmsg (i.e. SYSLOG_ACTION_CLEAR), so we have a strict subset of the non-destructive syslog syscall actions.
To this end, move the check as Josh had done, but also rename the constants to reflect their new uses (SYSLOG_FROM_CALL becomes SYSLOG_FROM_READER, and SYSLOG_FROM_FILE becomes SYSLOG_FROM_PROC). SYSLOG_FROM_READER allows non-destructive actions, and SYSLOG_FROM_PROC allows destructive actions after a capabilities-constrained SYSLOG_ACTION_OPEN check.
- /dev/kmsg allows:
- open if CAP_SYSLOG or dmesg_restrict==0
- reading/polling, after open
--- fs/proc/kmsg.c | 10 +++--- include/linux/syslog.h | 4 +-- kernel/printk.c | 91 ++++++++++++++++++++++++++---------------------- 3 files changed, 57 insertions(+), 48 deletions(-)
fs/proc/kmsg.c +5 -5
@@ -21,12 +21,12 @@ extern wait_queue_head_t log_wait;static int kmsg_open (struct inode * inode,struct file * file) { -return do_syslog (SYSLOG_ACTION_OPEN ,NULL ,0 ,SYSLOG_FROM_FILE ); +return do_syslog (SYSLOG_ACTION_OPEN ,NULL ,0 ,SYSLOG_FROM_PROC ); }static int kmsg_release (struct inode * inode,struct file * file) { - (void )do_syslog (SYSLOG_ACTION_CLOSE ,NULL ,0 ,SYSLOG_FROM_FILE ); + (void )do_syslog (SYSLOG_ACTION_CLOSE ,NULL ,0 ,SYSLOG_FROM_PROC );return 0 ; }
@@ -34,15 +34,15 @@ static ssize_t kmsg_read(struct file *file, char __user *buf, size_t count, loff_t *ppos) {if ((file->f_flags &O_NONBLOCK ) && - !do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_FILE )) + !do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_PROC ))return -EAGAIN ; -return do_syslog (SYSLOG_ACTION_READ , buf, count,SYSLOG_FROM_FILE ); +return do_syslog (SYSLOG_ACTION_READ , buf, count,SYSLOG_FROM_PROC ); }static unsigned int kmsg_poll (struct file *file,poll_table *wait) {poll_wait (file, &log_wait, wait); -if (do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_FILE )) +if (do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_PROC ))return POLLIN |POLLRDNORM ;return 0 ; }
include/linux/syslog.h +2 -2
@@ -44,8 +44,8 @@/* Return size of the log buffer */ #define SYSLOG_ACTION_SIZE_BUFFER 10 -#define SYSLOG_FROM_CALL 0 -#define SYSLOG_FROM_FILE 1 +#define SYSLOG_FROM_READER 0 +#define SYSLOG_FROM_PROC 1int do_syslog (int type,char __user *buf,int count,bool from_file);
kernel/printk.c +50 -41
115 lines not on this page · raw message
Signature · 6 lines
1.7.9.5 -- Kees Cook Chrome OS Security
Quoted 34 lines · Kees Cook
To fix
/dev/kmsg, let's compare the existing interfaces and what they allow:-
/proc/kmsgallows:
- open (SYSLOG_ACTION_OPEN) ifCAP_SYSLOGsince it uses a destructive single-reader interface (SYSLOG_ACTION_READ).
- everything, after an open.- syslog syscall allows:
- anything, ifCAP_SYSLOG.
-SYSLOG_ACTION_READ_ALLandSYSLOG_ACTION_SIZE_BUFFER, if dmesg_restrict==0.
- nothing else (EPERM).The use-cases were:
- dmesg(1) needs to do non-destructiveSYSLOG_ACTION_READ_ALLs.
- sysklog(1) needs to open/proc/kmsg, drop privs, and still issue the destructiveSYSLOG_ACTION_READs.AIUI, dmesg(1) is moving to
/dev/kmsg, and systemd-journald doesn't clear the ring buffer.Based on the comments in
devkmsg_llseek, it sounds like actions besides reading aren't going to be supported by/dev/kmsg(i.e.SYSLOG_ACTION_CLEAR), so we have a strict subset of the non-destructive syslog syscall actions.To this end, move the check as Josh had done, but also rename the constants to reflect their new uses (
SYSLOG_FROM_CALLbecomesSYSLOG_FROM_READER, andSYSLOG_FROM_FILEbecomesSYSLOG_FROM_PROC).SYSLOG_FROM_READERallows non-destructive actions, andSYSLOG_FROM_PROCallows destructive actions after a capabilities-constrainedSYSLOG_ACTION_OPENcheck.-
/dev/kmsgallows:
- open ifCAP_SYSLOGor dmesg_restrict==0
- reading/polling, after open
Quoted 5 lines
Signed-off-by Kees Cook <[email protected]>Reported-by Christian Kujau <[email protected]>Cc Josh Boyer <[email protected]>Cc Kay Sievers <[email protected]>
So staring at this for a while, I think it looks correct. It's basically the same thing as the v3 I sent out, with the constant rename and no check in devkmsg_read, right?
I'll try and get it tested here locally tomorrow.
josh
Quoted 189 lines
--- fs/proc/kmsg.c | 10 +++--- include/linux/syslog.h | 4 +-- kernel/printk.c | 91 ++++++++++++++++++++++++++---------------------- 3 files changed, 57 insertions(+), 48 deletions(-)fs/proc/kmsg.c +5
-5@@ -21,12 +21,12 @@ extern wait_queue_head_t log_wait;static int kmsg_open (struct inode * inode,struct file * file) { -return do_syslog (SYSLOG_ACTION_OPEN ,NULL ,0 ,SYSLOG_FROM_FILE ); +return do_syslog (SYSLOG_ACTION_OPEN ,NULL ,0 ,SYSLOG_FROM_PROC ); }static int kmsg_release (struct inode * inode,struct file * file) { - (void )do_syslog (SYSLOG_ACTION_CLOSE ,NULL ,0 ,SYSLOG_FROM_FILE ); + (void )do_syslog (SYSLOG_ACTION_CLOSE ,NULL ,0 ,SYSLOG_FROM_PROC );return 0 ; }@@ -34,15 +34,15 @@ static ssize_t kmsg_read(struct file *file, char __user *buf, size_t count, loff_t *ppos) {if ((file->f_flags &O_NONBLOCK ) && - !do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_FILE )) + !do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_PROC ))return -EAGAIN ; -return do_syslog (SYSLOG_ACTION_READ , buf, count,SYSLOG_FROM_FILE ); +return do_syslog (SYSLOG_ACTION_READ , buf, count,SYSLOG_FROM_PROC ); }static unsigned int kmsg_poll (struct file *file,poll_table *wait) {poll_wait (file, &log_wait, wait); -if (do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_FILE )) +if (do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_PROC ))return POLLIN |POLLRDNORM ;return 0 ; }include/linux/syslog.h +2
-2@@ -44,8 +44,8 @@/* Return size of the log buffer */ #define SYSLOG_ACTION_SIZE_BUFFER 10 -#define SYSLOG_FROM_CALL 0 -#define SYSLOG_FROM_FILE 1 +#define SYSLOG_FROM_READER 0 +#define SYSLOG_FROM_PROC 1int do_syslog (int type,char __user *buf,int count,bool from_file);kernel/printk.c +50
-41@@ -368,6 +368,53 @@ static void log_store(int facility, int level, log_next_seq++; } +#ifdef CONFIG_SECURITY_DMESG_RESTRICT +int dmesg_restrict =1 ; +#else +int dmesg_restrict; +#endif + +static int syslog_action_restricted (int type) +{ +if (dmesg_restrict) +return 1 ; +/* +* Unless restricted, we allow "read all" and "get buffer size" +* for everybody. +*/ +return type !=SYSLOG_ACTION_READ_ALL && + type !=SYSLOG_ACTION_SIZE_BUFFER ; +} + +static int check_syslog_permissions (int type,bool from_file) +{ +/* +* If this is from /proc/kmsg and we've already opened it, then we've +* already done the capabilities checks at open time. +*/ +if (from_file && type !=SYSLOG_ACTION_OPEN ) +return 0 ; + +if (syslog_action_restricted (type)) { +if (capable (CAP_SYSLOG )) +return 0 ; +/* +* For historical reasons, accept CAP_SYS_ADMIN too, with +* a warning. +*/ +if (capable (CAP_SYS_ADMIN )) { +printk_once (KERN_WARNING "%s (%d): " +"Attempt to access syslog with CAP_SYS_ADMIN " +"but no CAP_SYSLOG (deprecated).\n" , + current->comm,task_pid_nr (current)); +return 0 ; + } +return -EPERM ; + } +return security_syslog (type); +} + +/* /dev/kmsg - userspace message inject/listen interface */ struct devkmsg_user {u64 seq;@@ -624,7 +671,8 @@ static int devkmsg_open(struct inode *inode, struct file *file)if ((file->f_flags &O_ACCMODE ) ==O_WRONLY )return 0 ; - err =security_syslog (SYSLOG_ACTION_READ_ALL ); + err =check_syslog_permissions (SYSLOG_ACTION_READ_ALL , +SYSLOG_FROM_READER );if (err)return err;@@ -817,45 +865,6 @@ static inline void boot_delay_msec(int level) }#endif -#ifdef CONFIG_SECURITY_DMESG_RESTRICT -int dmesg_restrict =1 ; -#else -int dmesg_restrict; -#endif - -static int syslog_action_restricted (int type) -{ -if (dmesg_restrict) -return 1 ; -/* Unless restricted, we allow "read all" and "get buffer size" for everybody */ -return type !=SYSLOG_ACTION_READ_ALL && type !=SYSLOG_ACTION_SIZE_BUFFER ; -} - -static int check_syslog_permissions (int type,bool from_file) -{ -/* -* If this is from /proc/kmsg and we've already opened it, then we've -* already done the capabilities checks at open time. -*/ -if (from_file && type !=SYSLOG_ACTION_OPEN ) -return 0 ; - -if (syslog_action_restricted (type)) { -if (capable (CAP_SYSLOG )) -return 0 ; -/* For historical reasons, accept CAP_SYS_ADMIN too, with a warning */ -if (capable (CAP_SYS_ADMIN )) { -printk_once (KERN_WARNING "%s (%d): " -"Attempt to access syslog with CAP_SYS_ADMIN " -"but no CAP_SYSLOG (deprecated).\n" , - current->comm,task_pid_nr (current)); -return 0 ; - } -return -EPERM ; - } -return 0 ; -} -#if defined(CONFIG_PRINTK_TIME )static bool printk_time =1 ;#else @@ -1253,7 +1262,7 @@ out: SYSCALL_DEFINE3(syslog, int, type, char __user *, buf, int, len) { -return do_syslog (type, buf, len,SYSLOG_FROM_CALL ); +return do_syslog (type, buf, len,SYSLOG_FROM_READER ); }/* Signature · 6 lines
1.7.9.5 -- Kees Cook Chrome OS Security
Quoted 41 lines · Josh Boyer
Quoted 34 lines · Kees Cook
To fix
/dev/kmsg, let's compare the existing interfaces and what they allow:-
/proc/kmsgallows:
- open (SYSLOG_ACTION_OPEN) ifCAP_SYSLOGsince it uses a destructive single-reader interface (SYSLOG_ACTION_READ).
- everything, after an open.- syslog syscall allows:
- anything, ifCAP_SYSLOG.
-SYSLOG_ACTION_READ_ALLandSYSLOG_ACTION_SIZE_BUFFER, if dmesg_restrict==0.
- nothing else (EPERM).The use-cases were:
- dmesg(1) needs to do non-destructiveSYSLOG_ACTION_READ_ALLs.
- sysklog(1) needs to open/proc/kmsg, drop privs, and still issue the destructiveSYSLOG_ACTION_READs.AIUI, dmesg(1) is moving to
/dev/kmsg, and systemd-journald doesn't clear the ring buffer.Based on the comments in
devkmsg_llseek, it sounds like actions besides reading aren't going to be supported by/dev/kmsg(i.e.SYSLOG_ACTION_CLEAR), so we have a strict subset of the non-destructive syslog syscall actions.To this end, move the check as Josh had done, but also rename the constants to reflect their new uses (
SYSLOG_FROM_CALLbecomesSYSLOG_FROM_READER, andSYSLOG_FROM_FILEbecomesSYSLOG_FROM_PROC).SYSLOG_FROM_READERallows non-destructive actions, andSYSLOG_FROM_PROCallows destructive actions after a capabilities-constrainedSYSLOG_ACTION_OPENcheck.-
/dev/kmsgallows:
- open ifCAP_SYSLOGor dmesg_restrict==0
- reading/polling, after openQuoted 5 lines
Signed-off-by Kees Cook <[email protected]>Reported-by Christian Kujau <[email protected]>Cc Josh Boyer <[email protected]>Cc Kay Sievers <[email protected]>
Quoted 3 lines
So staring at this for a while, I think it looks correct. It's basically the same thing as the v3 I sent out, with the constant rename and no check in
devkmsg_read, right?
Yeah, I just explicitly clarified the use-cases and reasoning. :)
Quoted 1 line
I'll try and get it tested here locally tomorrow.
Great; thanks!
-Kees
--
Kees Cook
Chrome OS Security
Quoted 34 lines · Kees Cook
To fix
/dev/kmsg, let's compare the existing interfaces and what they allow:-
/proc/kmsgallows:
- open (SYSLOG_ACTION_OPEN) ifCAP_SYSLOGsince it uses a destructive single-reader interface (SYSLOG_ACTION_READ).
- everything, after an open.- syslog syscall allows:
- anything, ifCAP_SYSLOG.
-SYSLOG_ACTION_READ_ALLandSYSLOG_ACTION_SIZE_BUFFER, if dmesg_restrict==0.
- nothing else (EPERM).The use-cases were:
- dmesg(1) needs to do non-destructiveSYSLOG_ACTION_READ_ALLs.
- sysklog(1) needs to open/proc/kmsg, drop privs, and still issue the destructiveSYSLOG_ACTION_READs.AIUI, dmesg(1) is moving to
/dev/kmsg, and systemd-journald doesn't clear the ring buffer.Based on the comments in
devkmsg_llseek, it sounds like actions besides reading aren't going to be supported by/dev/kmsg(i.e.SYSLOG_ACTION_CLEAR), so we have a strict subset of the non-destructive syslog syscall actions.To this end, move the check as Josh had done, but also rename the constants to reflect their new uses (
SYSLOG_FROM_CALLbecomesSYSLOG_FROM_READER, andSYSLOG_FROM_FILEbecomesSYSLOG_FROM_PROC).SYSLOG_FROM_READERallows non-destructive actions, andSYSLOG_FROM_PROCallows destructive actions after a capabilities-constrainedSYSLOG_ACTION_OPENcheck.-
/dev/kmsgallows:
- open ifCAP_SYSLOGor dmesg_restrict==0
- reading/polling, after open
Quoted 5 lines
Signed-off-by Kees Cook <[email protected]>Reported-by Christian Kujau <[email protected]>Cc Josh Boyer <[email protected]>Cc Kay Sievers <[email protected]>
OK, this seems to work as expected. Non-privileged users can open /dev/kmsg for reading and read dmesg if dmesg_restrict is not set. If set, dmesg(1) will try both /dev/kmsg and the syslog call and get EPERM back. Running as root still works.
Quoted 189 lines
--- fs/proc/kmsg.c | 10 +++--- include/linux/syslog.h | 4 +-- kernel/printk.c | 91 ++++++++++++++++++++++++++---------------------- 3 files changed, 57 insertions(+), 48 deletions(-)fs/proc/kmsg.c +5
-5@@ -21,12 +21,12 @@ extern wait_queue_head_t log_wait;static int kmsg_open (struct inode * inode,struct file * file) { -return do_syslog (SYSLOG_ACTION_OPEN ,NULL ,0 ,SYSLOG_FROM_FILE ); +return do_syslog (SYSLOG_ACTION_OPEN ,NULL ,0 ,SYSLOG_FROM_PROC ); }static int kmsg_release (struct inode * inode,struct file * file) { - (void )do_syslog (SYSLOG_ACTION_CLOSE ,NULL ,0 ,SYSLOG_FROM_FILE ); + (void )do_syslog (SYSLOG_ACTION_CLOSE ,NULL ,0 ,SYSLOG_FROM_PROC );return 0 ; }@@ -34,15 +34,15 @@ static ssize_t kmsg_read(struct file *file, char __user *buf, size_t count, loff_t *ppos) {if ((file->f_flags &O_NONBLOCK ) && - !do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_FILE )) + !do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_PROC ))return -EAGAIN ; -return do_syslog (SYSLOG_ACTION_READ , buf, count,SYSLOG_FROM_FILE ); +return do_syslog (SYSLOG_ACTION_READ , buf, count,SYSLOG_FROM_PROC ); }static unsigned int kmsg_poll (struct file *file,poll_table *wait) {poll_wait (file, &log_wait, wait); -if (do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_FILE )) +if (do_syslog (SYSLOG_ACTION_SIZE_UNREAD ,NULL ,0 ,SYSLOG_FROM_PROC ))return POLLIN |POLLRDNORM ;return 0 ; }include/linux/syslog.h +2
-2@@ -44,8 +44,8 @@/* Return size of the log buffer */ #define SYSLOG_ACTION_SIZE_BUFFER 10 -#define SYSLOG_FROM_CALL 0 -#define SYSLOG_FROM_FILE 1 +#define SYSLOG_FROM_READER 0 +#define SYSLOG_FROM_PROC 1int do_syslog (int type,char __user *buf,int count,bool from_file);kernel/printk.c +50
-41@@ -368,6 +368,53 @@ static void log_store(int facility, int level, log_next_seq++; } +#ifdef CONFIG_SECURITY_DMESG_RESTRICT +int dmesg_restrict =1 ; +#else +int dmesg_restrict; +#endif + +static int syslog_action_restricted (int type) +{ +if (dmesg_restrict) +return 1 ; +/* +* Unless restricted, we allow "read all" and "get buffer size" +* for everybody. +*/ +return type !=SYSLOG_ACTION_READ_ALL && + type !=SYSLOG_ACTION_SIZE_BUFFER ; +} + +static int check_syslog_permissions (int type,bool from_file) +{ +/* +* If this is from /proc/kmsg and we've already opened it, then we've +* already done the capabilities checks at open time. +*/ +if (from_file && type !=SYSLOG_ACTION_OPEN ) +return 0 ; + +if (syslog_action_restricted (type)) { +if (capable (CAP_SYSLOG )) +return 0 ; +/* +* For historical reasons, accept CAP_SYS_ADMIN too, with +* a warning. +*/ +if (capable (CAP_SYS_ADMIN )) { +printk_once (KERN_WARNING "%s (%d): " +"Attempt to access syslog with CAP_SYS_ADMIN " +"but no CAP_SYSLOG (deprecated).\n" , + current->comm,task_pid_nr (current)); +return 0 ; + } +return -EPERM ; + } +return security_syslog (type); +} + +/* /dev/kmsg - userspace message inject/listen interface */ struct devkmsg_user {u64 seq;@@ -624,7 +671,8 @@ static int devkmsg_open(struct inode *inode, struct file *file)if ((file->f_flags &O_ACCMODE ) ==O_WRONLY )return 0 ; - err =security_syslog (SYSLOG_ACTION_READ_ALL ); + err =check_syslog_permissions (SYSLOG_ACTION_READ_ALL , +SYSLOG_FROM_READER );if (err)return err;@@ -817,45 +865,6 @@ static inline void boot_delay_msec(int level) }#endif -#ifdef CONFIG_SECURITY_DMESG_RESTRICT -int dmesg_restrict =1 ; -#else -int dmesg_restrict; -#endif - -static int syslog_action_restricted (int type) -{ -if (dmesg_restrict) -return 1 ; -/* Unless restricted, we allow "read all" and "get buffer size" for everybody */ -return type !=SYSLOG_ACTION_READ_ALL && type !=SYSLOG_ACTION_SIZE_BUFFER ; -} - -static int check_syslog_permissions (int type,bool from_file) -{ -/* -* If this is from /proc/kmsg and we've already opened it, then we've -* already done the capabilities checks at open time. -*/ -if (from_file && type !=SYSLOG_ACTION_OPEN ) -return 0 ; - -if (syslog_action_restricted (type)) { -if (capable (CAP_SYSLOG )) -return 0 ; -/* For historical reasons, accept CAP_SYS_ADMIN too, with a warning */ -if (capable (CAP_SYS_ADMIN )) { -printk_once (KERN_WARNING "%s (%d): " -"Attempt to access syslog with CAP_SYS_ADMIN " -"but no CAP_SYSLOG (deprecated).\n" , - current->comm,task_pid_nr (current)); -return 0 ; - } -return -EPERM ; - } -return 0 ; -} -#if defined(CONFIG_PRINTK_TIME )static bool printk_time =1 ;#else @@ -1253,7 +1262,7 @@ out: SYSCALL_DEFINE3(syslog, int, type, char __user *, buf, int, len) { -return do_syslog (type, buf, len,SYSLOG_FROM_CALL ); +return do_syslog (type, buf, len,SYSLOG_FROM_READER ); }/* Signature · 6 lines
1.7.9.5 -- Kees Cook Chrome OS Security
Quoted 30 lines · On Tue, 30 Apr 2013 10:25:41 -0700 Kees Cook <[email protected]> wrote:
To fix
/dev/kmsg, let's compare the existing interfaces and what they allow:-
/proc/kmsgallows:
- open (SYSLOG_ACTION_OPEN) ifCAP_SYSLOGsince it uses a destructive single-reader interface (SYSLOG_ACTION_READ).
- everything, after an open.- syslog syscall allows:
- anything, ifCAP_SYSLOG.
-SYSLOG_ACTION_READ_ALLandSYSLOG_ACTION_SIZE_BUFFER, if dmesg_restrict==0.
- nothing else (EPERM).The use-cases were:
- dmesg(1) needs to do non-destructiveSYSLOG_ACTION_READ_ALLs.
- sysklog(1) needs to open/proc/kmsg, drop privs, and still issue the destructiveSYSLOG_ACTION_READs.AIUI, dmesg(1) is moving to
/dev/kmsg, and systemd-journald doesn't clear the ring buffer.Based on the comments in
devkmsg_llseek, it sounds like actions besides reading aren't going to be supported by/dev/kmsg(i.e.SYSLOG_ACTION_CLEAR), so we have a strict subset of the non-destructive syslog syscall actions.To this end, move the check as Josh had done, but also rename the constants to reflect their new uses (
SYSLOG_FROM_CALLbecomesSYSLOG_FROM_READER, andSYSLOG_FROM_FILEbecomesSYSLOG_FROM_PROC).SYSLOG_FROM_READERallows non-destructive actions, andSYSLOG_FROM_PROCallows destructive actions after a capabilities-constrainedSYSLOG_ACTION_OPENcheck.
Quoted 3 lines
-
/dev/kmsgallows:
- open ifCAP_SYSLOGor dmesg_restrict==0
- reading/polling, after open
hm, that changelog is waaaay down in the weeds and anyone who hasn't been following this with a microscope won't have a clue.
I went into an earlier patch and dug out this:
: The dmesg_restrict sysctl currently covers the syslog method for access : dmesg, however /dev/kmsg isn't covered by the same protections. Most : people haven't noticed because util-linux dmesg(1) defaults to using the : syslog method for access in older versions. With util-linux dmesg(1) : defaults to reading directly from /dev/kmsg.
:
: Addresses https://bugzilla.redhat.com/show_bug.cgi?id=903192
Which is still accurate and relevant, yes?
Quoted 49 lines · Andrew Morton
Quoted 30 lines · On Tue, 30 Apr 2013 10:25:41 -0700 Kees Cook <[email protected]> wrote:
To fix
/dev/kmsg, let's compare the existing interfaces and what they allow:-
/proc/kmsgallows:
- open (SYSLOG_ACTION_OPEN) ifCAP_SYSLOGsince it uses a destructive single-reader interface (SYSLOG_ACTION_READ).
- everything, after an open.- syslog syscall allows:
- anything, ifCAP_SYSLOG.
-SYSLOG_ACTION_READ_ALLandSYSLOG_ACTION_SIZE_BUFFER, if dmesg_restrict==0.
- nothing else (EPERM).The use-cases were:
- dmesg(1) needs to do non-destructiveSYSLOG_ACTION_READ_ALLs.
- sysklog(1) needs to open/proc/kmsg, drop privs, and still issue the destructiveSYSLOG_ACTION_READs.AIUI, dmesg(1) is moving to
/dev/kmsg, and systemd-journald doesn't clear the ring buffer.Based on the comments in
devkmsg_llseek, it sounds like actions besides reading aren't going to be supported by/dev/kmsg(i.e.SYSLOG_ACTION_CLEAR), so we have a strict subset of the non-destructive syslog syscall actions.To this end, move the check as Josh had done, but also rename the constants to reflect their new uses (
SYSLOG_FROM_CALLbecomesSYSLOG_FROM_READER, andSYSLOG_FROM_FILEbecomesSYSLOG_FROM_PROC).SYSLOG_FROM_READERallows non-destructive actions, andSYSLOG_FROM_PROCallows destructive actions after a capabilities-constrainedSYSLOG_ACTION_OPENcheck.Quoted 3 lines
-
/dev/kmsgallows:
- open ifCAP_SYSLOGor dmesg_restrict==0
- reading/polling, after openhm, that changelog is waaaay down in the weeds and anyone who hasn't been following this with a microscope won't have a clue.
I went into an earlier patch and dug out this:
: The
dmesg_restrictsysctl currently covers the syslog method for access : dmesg, however/dev/kmsgisn't covered by the same protections. Most : people haven't noticed because util-linux dmesg(1) defaults to using the : syslog method for access in older versions. With util-linux dmesg(1) : defaults to reading directly from/dev/kmsg.
:
: Addresses https://bugzilla.redhat.com/show_bug.cgi?id=903192
Quoted 1 line
Which is still accurate and relevant, yes?
Yeah, that'll be fine. I kind of like having the longer rationale in the commit message for future reference (i.e. destructive vs non-destructive, etc), but I'd rather see the code fixed. :)
-Kees
--
Kees Cook
Chrome OS Security
Quoted 3 lines · Kees Cook
Yeah, that'll be fine. I kind of like having the longer rationale in the commit message for future reference (i.e. destructive vs non-destructive, etc), but I'd rather see the code fixed. :)
There's no reason not to have both, is there?
-- Steve