liveupdate: truncate getfile name to NAME_MAX
Subsystems (3)
PATCH 1/1 liveupdate: truncate getfile name to NAME_MAX
From: Luca Boccassi <[email protected]>
dynamic_dname() harcodes its limit to NAME_MAX bytes, which includes a NUL terminator and a 'anon_inode:' prefix. If the name passed on goes above the limit it results in userspace getting a ENAMETOOLONG error.
Truncate the name to NAME_MAX - 12 - 1 characters to ensure this never fails (accounting for prefix and NUL).
---
Builds on top of this just sent by Aleksa:
https://lore.kernel.org/linux-fsdevel/[email protected]/
kernel/liveupdate/luo_session.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
kernel/liveupdate/luo_session.c +5 -2
@@ -366,11 +366,14 @@ static const struct file_operations luo_session_fops = {/* Create a "struct file" for session */ static int luo_session_getfile (struct luo_session *session,struct file **filep) { -char name_buf[128 ]; +char name_buf[NAME_MAX ];struct file *file;lockdep_assert_held (&session->mutex); -snprintf (name_buf,sizeof (name_buf),"[luo_session] %s" , session->name); +/* dynamic_dname() rejects names above NAME_MAX bytes, including NUL terminator +* and a 'anon_inode:' prefix. Truncate to NAME_MAX - 12 - 1 to avoid +* ENAMETOOLONG. */ +snprintf (name_buf,NAME_MAX -12 -1 ,"[luo_session] %s" , session->name); file =anon_inode_getfile (name_buf, &luo_session_fops, session,O_RDWR );if (IS_ERR (file))return PTR_ERR (file);
Signature · 1 line
2.47.3
Pasha Tatashin
maintainerQuoted 29 lines · On Tue, Mar 31, 2026 at 1:55 PM <[email protected]> wrote:
From: Luca Boccassi <[email protected]>
dynamic_dname()harcodes its limit toNAME_MAXbytes, which includes a NUL terminator and a 'anon_inode:' prefix. If the name passed on goes above the limit it results in userspace getting a ENAMETOOLONG error.Truncate the name to
NAME_MAX- 12 - 1 characters to ensure this never fails (accounting for prefix and NUL).Fixes 0153094d03df ("liveupdate: luo_session: add sessions support")Signed-off-by Luca Boccassi <[email protected]>---
Builds on top of this just sent by Aleksa:https://lore.kernel.org/linux-fsdevel/[email protected]/
kernel/liveupdate/luo_session.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)kernel/liveupdate/luo_session.c +0
-0@@ -366,11 +366,14 @@ static const struct file_operations luo_session_fops = {/* Create a "struct file" for session */ static int luo_session_getfile (struct luo_session *session,struct file **filep) {
kernel/liveupdate/luo_session.c static const struct file_operations luo_session_fops = { · line 376
Quoted 10 lines
kernel/liveupdate/luo_session.c +5
-2-char name_buf[128 ]; +char name_buf[NAME_MAX ];struct file *file;lockdep_assert_held (&session->mutex); -snprintf (name_buf,sizeof (name_buf),"[luo_session] %s" , session->name); +/* dynamic_dname() rejects names above NAME_MAX bytes, including NUL terminator +* and a 'anon_inode:' prefix. Truncate to NAME_MAX - 12 - 1 to avoid +* ENAMETOOLONG. */ +snprintf (name_buf,NAME_MAX -12 -1 ,"[luo_session] %s" , session->name);
I am confused by this change, the maximum session name length is: 64, as defined in:include/uapi/linux/liveupdate.h
/* The maximum length of session name including null termination */ #define LIVEUPDATE_SESSION_NAME_LENGTH 64
So, 128 should be enough to include:
"[luo_session] %s", session->name
Or am I missing something?
Pasha
Quoted 55 lines · On Tue, 31 Mar 2026, 19:21 Pasha Tatashin, <[email protected]> wrote:
Quoted 7 lines · On Tue, Mar 31, 2026 at 1:55 PM <[email protected]> wrote:
From: Luca Boccassi <[email protected]>
dynamic_dname()harcodes its limit toNAME_MAXbytes, which includes a NUL terminator and a 'anon_inode:' prefix. If the name passed on goes above the limit it results in userspace getting a ENAMETOOLONG error.Quoted 10 lines
Truncate the name to
NAME_MAX- 12 - 1 characters to ensure this never fails (accounting for prefix and NUL).Fixes 0153094d03df ("liveupdate: luo_session: add sessions support")Signed-off-by Luca Boccassi <[email protected]>---
Builds on top of this just sent by Aleksa:https://lore.kernel.org/linux-fsdevel/[email protected]/
Review
Quoted 5 lines
kernel/liveupdate/luo_session.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)(quoted diff) +0
-0
b/kernel/liveupdate/luo_session.ckernel/liveupdate/luo_session.c static const struct file_operations
Quoted 4 lines
index 7836772956406..4a40c7fdfb44f 100644
kernel/liveupdate/luo_session.c +0
-0@@ -366,11 +366,14 @@ static const struct file_operations
luo_session_fops= {kernel/liveupdate/luo_session.c static const struct file_operations · line 367
Quoted 2 lines
kernel/liveupdate/luo_session.c +0
-0/* Create a "struct file" for session */ static int luo_session_getfile(struct luo_session *session, struct file**filep)
kernel/liveupdate/luo_session.c static const struct file_operations luo_session_fops = {
Quoted 7 lines
kernel/liveupdate/luo_session.c +1
-2{ -char name_buf[128 ]; +char name_buf[NAME_MAX ];struct file *file;lockdep_assert_held (&session->mutex); - snprintf(name_buf, sizeof(name_buf), "[luo_session] %s",
session->name);kernel/liveupdate/luo_session.c static const struct file_operations · line 373
Quoted 1 line
kernel/liveupdate/luo_session.c +1
-0+/* dynamic_dname() rejects names above NAME_MAX bytes, including NUL terminator
kernel/liveupdate/luo_session.c static const struct file_operations · line 374
Quoted 1 line
kernel/liveupdate/luo_session.c +1
-0+* and a 'anon_inode:' prefix. Truncate to NAME_MAX - 12 - 1 to avoid
kernel/liveupdate/luo_session.c static const struct file_operations · line 376
Quoted 2 lines
kernel/liveupdate/luo_session.c +2
-0+* ENAMETOOLONG. */ + snprintf(name_buf, NAME_MAX - 12 - 1, "[luo_session] %s",
session->name);I am confused by this change, the maximum session name length is: 64, as defined in:
include/uapi/linux/liveupdate.h
/* The maximum length of session name including null termination */ #defineLIVEUPDATE_SESSION_NAME_LENGTH64
Quoted 7 lines
So, 128 should be enough to include:
"[luo_session] %s",session->nameOr am I missing something?
Pasha
Yeah we need to to bump that size, it's too small - see the shared doc. This is anyway needed for safety, as the user space call falls apart completely if it goes over the hard coded limit, so it should always be used as a max value
Quoted 1 line
Pasha Tatashin
maintainerQuoted 7 lines · Luca Boccassi
Quoted 5 lines · On Tue, 31 Mar 2026, 19:21 Pasha Tatashin, <[email protected]> wrote:
Quoted 3 lines · On Tue, Mar 31, 2026 at 1:55 PM <[email protected]> wrote:
From: Luca Boccassi <[email protected]>
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
dynamic_dname()harcodes its limit toNAME_MAXbytes, which includes a NUL terminator and a 'anon_inode:' prefix. If the name passed on goes above the limit it results in userspace getting a ENAMETOOLONG error.Truncate the name to
NAME_MAX- 12 - 1 characters to ensure this never fails (accounting for prefix and NUL).Fixes 0153094d03df ("liveupdate: luo_session: add sessions support")
Let's remove this, this patch is not fixing an existing issue, but a preparation for future uapi changes?
Quoted 38 lines
Quoted 38 lines
Quoted 18 lines
Signed-off-by Luca Boccassi <[email protected]>---
Builds on top of this just sent by Aleksa:https://lore.kernel.org/linux-fsdevel/[email protected]/
kernel/liveupdate/luo_session.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)kernel/liveupdate/luo_session.c +0
-0@@ -366,11 +366,14 @@ static const struct file_operations luo_session_fops = {/* Create a "struct file" for session */ static int luo_session_getfile (struct luo_session *session,struct file **filep) {kernel/liveupdate/luo_session.c static const struct file_operations luo_session_fops = { · line 376
Quoted 10 lines
kernel/liveupdate/luo_session.c +5
-2-char name_buf[128 ]; +char name_buf[NAME_MAX ];struct file *file;lockdep_assert_held (&session->mutex); -snprintf (name_buf,sizeof (name_buf),"[luo_session] %s" , session->name); +/* dynamic_dname() rejects names above NAME_MAX bytes, including NUL terminator +* and a 'anon_inode:' prefix. Truncate to NAME_MAX - 12 - 1 to avoid +* ENAMETOOLONG. */ +snprintf (name_buf,NAME_MAX -12 -1 ,"[luo_session] %s" , session->name);I am confused by this change, the maximum session name length is: 64, as defined in:
include/uapi/linux/liveupdate.h
/* The maximum length of session name including null termination */ #defineLIVEUPDATE_SESSION_NAME_LENGTH64So, 128 should be enough to include:
"[luo_session] %s",session->name
Quoted 6 lines
Quoted 3 lines
Or am I missing something?
Pasha
Yeah we need to to bump that size, it's too small - see the shared doc. This is anyway needed for safety, as the user space call falls apart completely if it goes over the hard coded limit,
If there is a safety issue, it should be fixed (even without a limit bump), could you please explain it.
I looked at the shared doc [https://tinyurl.com/luoddesign under limits tab]:
I see:
TODO: this needs to be bumped, as for the varlink API we will impose a cgroup path prefix, which means PATH_MAX + "/" + client-supplied string.
Why would we do that? Why can't we use a 63-byte unique identifier for sessions and an in-memory database that maps the cgroup path prefix to session_ids, where the database also gets passed from one kernel to the other?
Increasing this limit would require a kernel uAPI change, which requires a very strong rationale, and would bump up the KHO ABI for LUO. It also means that the LUO session data in:
struct luo_session_ser {
char name[LIVEUPDATE_SESSION_NAME_LENGTH];
struct luo_file_set_ser file_set_ser;
} __packed;
would get bigger, and thus the kernel would need to preserve more data for every single session.
Ideally, this extra data should be preserved by the userspace orchestrator via a separate memfd.
Pasha
Quoted 13 lines · Pasha Tatashin
Quoted 11 lines · Luca Boccassi
Quoted 9 lines · On Tue, 31 Mar 2026, 19:21 Pasha Tatashin, <[email protected]> wrote:
Quoted 7 lines · On Tue, Mar 31, 2026 at 1:55 PM <[email protected]> wrote:
From: Luca Boccassi <[email protected]>
dynamic_dname()harcodes its limit toNAME_MAXbytes, which includes a NUL terminator and a 'anon_inode:' prefix. If the name passed on goes above the limit it results in userspace getting a ENAMETOOLONG error.
Quoted 7 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Truncate the name to
NAME_MAX- 12 - 1 characters to ensure this never fails (accounting for prefix and NUL).Fixes 0153094d03df ("liveupdate: luo_session: add sessions support")Let's remove this, this patch is not fixing an existing issue, but a preparation for future uapi changes?
The linked patch that increases the buffer to 255 is tagged with 'fixes' so I don't mind, as that's enough to fix the -ENAMETOOLONG error in currently released kernels. The original version of this set the truncation to the current max buffer size, hence why I had the fixes line.
Quoted 40 lines
Quoted 40 lines
Quoted 40 lines
Quoted 18 lines
Signed-off-by Luca Boccassi <[email protected]>---
Builds on top of this just sent by Aleksa:https://lore.kernel.org/linux-fsdevel/[email protected]/
kernel/liveupdate/luo_session.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)kernel/liveupdate/luo_session.c +0
-0@@ -366,11 +366,14 @@ static const struct file_operations luo_session_fops = {/* Create a "struct file" for session */ static int luo_session_getfile (struct luo_session *session,struct file **filep) {kernel/liveupdate/luo_session.c static const struct file_operations luo_session_fops = { · line 376
Quoted 10 lines
kernel/liveupdate/luo_session.c +5
-2-char name_buf[128 ]; +char name_buf[NAME_MAX ];struct file *file;lockdep_assert_held (&session->mutex); -snprintf (name_buf,sizeof (name_buf),"[luo_session] %s" , session->name); +/* dynamic_dname() rejects names above NAME_MAX bytes, including NUL terminator +* and a 'anon_inode:' prefix. Truncate to NAME_MAX - 12 - 1 to avoid +* ENAMETOOLONG. */ +snprintf (name_buf,NAME_MAX -12 -1 ,"[luo_session] %s" , session->name);I am confused by this change, the maximum session name length is: 64, as defined in:
include/uapi/linux/liveupdate.h
/* The maximum length of session name including null termination */ #defineLIVEUPDATE_SESSION_NAME_LENGTH64So, 128 should be enough to include:
"[luo_session] %s",session->nameOr am I missing something?
Quoted 7 lines
Quoted 4 lines
Quoted 1 line
Pasha
Yeah we need to to bump that size, it's too small - see the shared doc. This is anyway needed for safety, as the user space call falls apart completely if it goes over the hard coded limit,
If there is a safety issue, it should be fixed (even without a limit bump), could you please explain it.
Nothing specific, more from the abstract point of view - given that this will fail with a hard error if it goes above the hardcoded limit, it seems better to me to check it, rather than assuming it's going to match, as that's error prone.
Quoted 11 lines
I looked at the shared doc [https://tinyurl.com/luoddesign under limits tab]:
I see:
TODO: this needs to be bumped, as for the varlink API we will impose a cgroup path prefix, which meansPATH_MAX+ "/" + client-supplied string.Why would we do that? Why can't we use a 63-byte unique identifier for sessions and an in-memory database that maps the cgroup path prefix to
session_ids, where the database also gets passed from one kernel to the other?
Because that requires tracking state across all runtime, and that's a no-go. The session itself needs to be identifiable by cgroup, so that its owner is uniquely identified in all cases, in all situations, regardless of any state or lack thereof.
The structs have a size as input, so it should be doable to follow the usual process and keep it backward compatible.
Pasha Tatashin
maintainerQuoted 94 lines · Luca Boccassi
Quoted 13 lines · Pasha Tatashin
Quoted 11 lines · Luca Boccassi
Quoted 9 lines · On Tue, 31 Mar 2026, 19:21 Pasha Tatashin, <[email protected]> wrote:
Quoted 7 lines · On Tue, Mar 31, 2026 at 1:55 PM <[email protected]> wrote:
From: Luca Boccassi <[email protected]>
dynamic_dname()harcodes its limit toNAME_MAXbytes, which includes a NUL terminator and a 'anon_inode:' prefix. If the name passed on goes above the limit it results in userspace getting a ENAMETOOLONG error.Quoted 7 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Truncate the name to
NAME_MAX- 12 - 1 characters to ensure this never fails (accounting for prefix and NUL).Fixes 0153094d03df ("liveupdate: luo_session: add sessions support")Let's remove this, this patch is not fixing an existing issue, but a preparation for future uapi changes?
The linked patch that increases the buffer to 255 is tagged with 'fixes' so I don't mind, as that's enough to fix the
-ENAMETOOLONGerror in currently released kernels. The original version of this set the truncation to the current max buffer size, hence why I had the fixes line.Quoted 40 lines
Quoted 40 lines
Quoted 40 lines
Quoted 18 lines
Signed-off-by Luca Boccassi <[email protected]>---
Builds on top of this just sent by Aleksa:https://lore.kernel.org/linux-fsdevel/[email protected]/
kernel/liveupdate/luo_session.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)kernel/liveupdate/luo_session.c +0
-0@@ -366,11 +366,14 @@ static const struct file_operations luo_session_fops = {/* Create a "struct file" for session */ static int luo_session_getfile (struct luo_session *session,struct file **filep) {kernel/liveupdate/luo_session.c static const struct file_operations luo_session_fops = { · line 376
Quoted 10 lines
kernel/liveupdate/luo_session.c +5
-2-char name_buf[128 ]; +char name_buf[NAME_MAX ];struct file *file;lockdep_assert_held (&session->mutex); -snprintf (name_buf,sizeof (name_buf),"[luo_session] %s" , session->name); +/* dynamic_dname() rejects names above NAME_MAX bytes, including NUL terminator +* and a 'anon_inode:' prefix. Truncate to NAME_MAX - 12 - 1 to avoid +* ENAMETOOLONG. */ +snprintf (name_buf,NAME_MAX -12 -1 ,"[luo_session] %s" , session->name);I am confused by this change, the maximum session name length is: 64, as defined in:
include/uapi/linux/liveupdate.h
/* The maximum length of session name including null termination */ #defineLIVEUPDATE_SESSION_NAME_LENGTH64So, 128 should be enough to include:
"[luo_session] %s",session->nameOr am I missing something?
Quoted 7 lines
Quoted 4 lines
Quoted 1 line
Pasha
Yeah we need to to bump that size, it's too small - see the shared doc. This is anyway needed for safety, as the user space call falls apart completely if it goes over the hard coded limit,
If there is a safety issue, it should be fixed (even without a limit bump), could you please explain it.
Nothing specific, more from the abstract point of view - given that this will fail with a hard error if it goes above the hardcoded limit, it seems better to me to check it, rather than assuming it's going to match, as that's error prone.
Quoted 11 lines
I looked at the shared doc [https://tinyurl.com/luoddesign under limits tab]:
I see:
TODO: this needs to be bumped, as for the varlink API we will impose a cgroup path prefix, which meansPATH_MAX+ "/" + client-supplied string.Why would we do that? Why can't we use a 63-byte unique identifier for sessions and an in-memory database that maps the cgroup path prefix to
session_ids, where the database also gets passed from one kernel to the other?
Quoted 6 lines
Because that requires tracking state across all runtime, and that's a no-go. The session itself needs to be identifiable by cgroup, so that its owner is uniquely identified in all cases, in all situations, regardless of any state or lack thereof.
The structs have a size as input, so it should be doable to follow the usual process and keep it backward compatible.
Overall, I am super happy to see LUO integrated as a first-class citizen with systemd, and I fully support this effort. My suggestion is that we step back and prepare a single patch series that performs all the necessary changes together, instead of sending a few random patches that do not make sense by themselves.
As far as I can see, we should:
1. Increase the session name limit to include the data necessary for systemd fd store restore capabilities.
2. Make the number of stored sessions dynamic rather than having a hardcoded limit (I have a patch for this, which we can include in the series).
3. Add the necessary API to query back the incoming and outgoing session names.
Is there anything else that you are aware that we would need to add?
Thank you,
Pasha
Quoted 103 lines · Pasha Tatashin
Quoted 94 lines · Luca Boccassi
Quoted 13 lines · Pasha Tatashin
Quoted 11 lines · Luca Boccassi
Quoted 9 lines · On Tue, 31 Mar 2026, 19:21 Pasha Tatashin, <[email protected]> wrote:
Quoted 7 lines · On Tue, Mar 31, 2026 at 1:55 PM <[email protected]> wrote:
From: Luca Boccassi <[email protected]>
dynamic_dname()harcodes its limit toNAME_MAXbytes, which includes a NUL terminator and a 'anon_inode:' prefix. If the name passed on goes above the limit it results in userspace getting a ENAMETOOLONG error.Quoted 7 lines
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Truncate the name to
NAME_MAX- 12 - 1 characters to ensure this never fails (accounting for prefix and NUL).Fixes 0153094d03df ("liveupdate: luo_session: add sessions support")Let's remove this, this patch is not fixing an existing issue, but a preparation for future uapi changes?
The linked patch that increases the buffer to 255 is tagged with 'fixes' so I don't mind, as that's enough to fix the
-ENAMETOOLONGerror in currently released kernels. The original version of this set the truncation to the current max buffer size, hence why I had the fixes line.Quoted 40 lines
Quoted 40 lines
Quoted 40 lines
Quoted 18 lines
Signed-off-by Luca Boccassi <[email protected]>---
Builds on top of this just sent by Aleksa:https://lore.kernel.org/linux-fsdevel/[email protected]/
kernel/liveupdate/luo_session.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)kernel/liveupdate/luo_session.c +0
-0@@ -366,11 +366,14 @@ static const struct file_operations luo_session_fops = {/* Create a "struct file" for session */ static int luo_session_getfile (struct luo_session *session,struct file **filep) {kernel/liveupdate/luo_session.c static const struct file_operations luo_session_fops = { · line 376
Quoted 10 lines
kernel/liveupdate/luo_session.c +5
-2-char name_buf[128 ]; +char name_buf[NAME_MAX ];struct file *file;lockdep_assert_held (&session->mutex); -snprintf (name_buf,sizeof (name_buf),"[luo_session] %s" , session->name); +/* dynamic_dname() rejects names above NAME_MAX bytes, including NUL terminator +* and a 'anon_inode:' prefix. Truncate to NAME_MAX - 12 - 1 to avoid +* ENAMETOOLONG. */ +snprintf (name_buf,NAME_MAX -12 -1 ,"[luo_session] %s" , session->name);I am confused by this change, the maximum session name length is: 64, as defined in:
include/uapi/linux/liveupdate.h
/* The maximum length of session name including null termination */ #defineLIVEUPDATE_SESSION_NAME_LENGTH64So, 128 should be enough to include:
"[luo_session] %s",session->nameOr am I missing something?
Quoted 7 lines
Quoted 4 lines
Quoted 1 line
Pasha
Yeah we need to to bump that size, it's too small - see the shared doc. This is anyway needed for safety, as the user space call falls apart completely if it goes over the hard coded limit,
If there is a safety issue, it should be fixed (even without a limit bump), could you please explain it.
Nothing specific, more from the abstract point of view - given that this will fail with a hard error if it goes above the hardcoded limit, it seems better to me to check it, rather than assuming it's going to match, as that's error prone.
Quoted 11 lines
I looked at the shared doc [https://tinyurl.com/luoddesign under limits tab]:
I see:
TODO: this needs to be bumped, as for the varlink API we will impose a cgroup path prefix, which meansPATH_MAX+ "/" + client-supplied string.Why would we do that? Why can't we use a 63-byte unique identifier for sessions and an in-memory database that maps the cgroup path prefix to
session_ids, where the database also gets passed from one kernel to the other?Quoted 6 lines
Because that requires tracking state across all runtime, and that's a no-go. The session itself needs to be identifiable by cgroup, so that its owner is uniquely identified in all cases, in all situations, regardless of any state or lack thereof.
The structs have a size as input, so it should be doable to follow the usual process and keep it backward compatible.
Quoted 5 lines
Overall, I am super happy to see LUO integrated as a first-class citizen with systemd, and I fully support this effort. My suggestion is that we step back and prepare a single patch series that performs all the necessary changes together, instead of sending a few random patches that do not make sense by themselves.
Yeah I don't mind, I noticed these two issues and fixed them as I was going, so just thought of sending them as I verified they work. Thankfully Aleksa's patch is enough to fix the ENAMETOOLONG issue and should get into the released kernels too, so this can wait.
Quoted 9 lines
As far as I can see, we should:
1. Increase the session name limit to include the data necessary for systemd fd store restore capabilities.
2. Make the number of stored sessions dynamic rather than having a hardcoded limit (I have a patch for this, which we can include in the series).
3. Add the necessary API to query back the incoming and outgoing session names.Is there anything else that you are aware that we would need to add?
Yeah also mentioned in the doc - we should bump the limit of FDs per session, as for the fdstore I am using a single session for the whole system, as we talked about two weeks ago. Apart from this I haven't come across anything, but I just got the initial proof of concept to work end to end this morning so haven't had a lot of time. Also the per-client session is still in progress. If I find more requirements I'll definitely bring them up ASAP.