Skip to contentSkip to conversation

fs/fcntl: restore checking against COMPAT_LOFF_T_MAX for F_GETLK64

PATCH Superseded Vitaly Lipatov 3 participants 13 messages Nov 14, 2017 – Nov 15, 2017

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 fs/fcntl: restore checking against COMPAT_LOFF_T_MAX for F_GETLK64

for fcntl64 with F_GETLK64 we need use checking against COMPAT_LOFF_T_MAX.

Fixes 94073ad77fff2 "fs/locks: don't mess with the address limit in compat_fcntl64"
---
 fs/fcntl.c | 21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)
fs/fcntl.c +20 -1
index 30f47d0..fa17f67 100644
@@ -604,6 +604,25 @@ static int fixup_compat_flock(struct flock *flock)
 	return 0;
 }
 
+/*
+ * GETLK64 was successful and we need to return the data, but it needs to fit in
+ * the compat structure.
+ * l_start shouldn't be too big, unless the original start + end is greater than
+ * COMPAT_LOFF_T_MAX, in which case the app was asking for trouble, so we return
+ * -EOVERFLOW in that case.  l_len could be too big, in which case we just
+ * truncate it, and only allow the app to see that part of the conflicting lock
+ * that might make sense to it anyway
+ */
+
+static int fixup_compat_l_flock(struct flock *flock)
+{
+	if (flock->l_start > COMPAT_LOFF_T_MAX)
+		return -EOVERFLOW;
+	if (flock->l_len > COMPAT_LOFF_T_MAX)
+		flock->l_len = COMPAT_LOFF_T_MAX;
+	return 0;
+}
+
 COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		       compat_ulong_t, arg)
 {
@@ -644,7 +663,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_l_flock(&flock);
 		if (err)
 			return err;
 		err = put_compat_flock64(&flock, compat_ptr(arg));
Signature · 1 line
2.10.4
maintainer
#1 lore raw
Quoted 26 lines · Vitaly Lipatov

for fcntl64 with F_GETLK64 we need use checking against COMPAT_LOFF_T_MAX.

Fixes 94073ad77fff2 "fs/locks: don't mess with the address limit in compat_fcntl64"
---
 fs/fcntl.c | 21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)
fs/fcntl.c +9 -0
index 30f47d0..fa17f67 100644
@@ -604,6 +604,25 @@ static int fixup_compat_flock(struct flock *flock)
 	return 0;
 }
 
+/*
+ * GETLK64 was successful and we need to return the data, but it needs to fit in
+ * the compat structure.
+ * l_start shouldn't be too big, unless the original start + end is greater than
+ * COMPAT_LOFF_T_MAX, in which case the app was asking for trouble, so we return
+ * -EOVERFLOW in that case.  l_len could be too big, in which case we just
+ * truncate it, and only allow the app to see that part of the conflicting lock
+ * that might make sense to it anyway
+ */

fs/fcntl.c static int fixup_compat_flock(struct flock *flock) · line 625

Quoted 10 lines
fs/fcntl.c +10 -0
+
+static int fixup_compat_l_flock(struct flock *flock)
+{
+	if (flock->l_start > COMPAT_LOFF_T_MAX)
+		return -EOVERFLOW;
+	if (flock->l_len > COMPAT_LOFF_T_MAX)
+		flock->l_len = COMPAT_LOFF_T_MAX;
+	return 0;
+}
+

(cc'ing Christoph since he wrote the original patch)

This patch looks correct to me, but could we rename it to fixup_compat_flock64 to match the other functions here?

Also, I think this should probably go to stable -- any objections?

Quoted 12 lines
fs/fcntl.c +1 -1
 COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		       compat_ulong_t, arg)
 {
@@ -644,7 +663,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_l_flock(&flock);
 		if (err)
 			return err;
 		err = put_compat_flock64(&flock, compat_ptr(arg));
Signature · 1 line
Jeff Layton <[email protected]>
#2 lore raw

Jeff Layton писал 14.11.17 14:29:

Quoted 4 lines

(cc'ing Christoph since he wrote the original patch)

This patch looks correct to me, but could we rename it to fixup_compat_flock64 to match the other functions here?

Sure, feel free to make any renames and so on.

Signature · 3 lines
С уважением,
Виталий Липатов,
Etersoft
#3 lore raw

PATCH v2 1/1 fs/fcntl: restore checking against COMPAT_LOFF_T_MAX for F_GETLK64

In reply to Vitaly Lipatov

for fcntl64 with F_GETLK64 we need use checking against COMPAT_LOFF_T_MAX.

Fixes 94073ad77fff2 "fs/locks: don't mess with the address limit in compat_fcntl64"
---
 fs/fcntl.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
fs/fcntl.c +7 -7
index 30f47d0..e9443d9 100644
@@ -590,17 +590,17 @@ convert_fcntl_cmd(unsigned int cmd)
  * GETLK was successful and we need to return the data, but it needs to fit in
  * the compat structure.
  * l_start shouldn't be too big, unless the original start + end is greater than
- * COMPAT_OFF_T_MAX, in which case the app was asking for trouble, so we return
+ * COMPAT_OFF_T_MAX/COMPAT_LOFF_T_MAX, in which case the app was asking for trouble, so we return
  * -EOVERFLOW in that case.  l_len could be too big, in which case we just
  * truncate it, and only allow the app to see that part of the conflicting lock
  * that might make sense to it anyway
  */
-static int fixup_compat_flock(struct flock *flock)
+static int fixup_compat_flock(struct flock *flock, loff_t off_t_max)
 {
-	if (flock->l_start > COMPAT_OFF_T_MAX)
+	if (flock->l_start > off_t_max)
 		return -EOVERFLOW;
-	if (flock->l_len > COMPAT_OFF_T_MAX)
-		flock->l_len = COMPAT_OFF_T_MAX;
+	if (flock->l_len > off_t_max)
+		flock->l_len = off_t_max;
 	return 0;
 }
 
@@ -631,7 +631,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_OFF_T_MAX);
 		if (err)
 			return err;
 		err = put_compat_flock(&flock, compat_ptr(arg));
@@ -644,7 +644,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_OFF_T_MAX);
 		if (err)
 			return err;
 		err = put_compat_flock64(&flock, compat_ptr(arg));
Signature · 1 line
2.10.4
maintainer
#4 lore raw
Quoted 46 lines · Vitaly Lipatov

for fcntl64 with F_GETLK64 we need use checking against COMPAT_LOFF_T_MAX.

Fixes 94073ad77fff2 "fs/locks: don't mess with the address limit in compat_fcntl64"
---
 fs/fcntl.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
fs/fcntl.c +6 -6
index 30f47d0..e9443d9 100644
@@ -590,17 +590,17 @@ convert_fcntl_cmd(unsigned int cmd)
  * GETLK was successful and we need to return the data, but it needs to fit in
  * the compat structure.
  * l_start shouldn't be too big, unless the original start + end is greater than
- * COMPAT_OFF_T_MAX, in which case the app was asking for trouble, so we return
+ * COMPAT_OFF_T_MAX/COMPAT_LOFF_T_MAX, in which case the app was asking for trouble, so we return
  * -EOVERFLOW in that case.  l_len could be too big, in which case we just
  * truncate it, and only allow the app to see that part of the conflicting lock
  * that might make sense to it anyway
  */
-static int fixup_compat_flock(struct flock *flock)
+static int fixup_compat_flock(struct flock *flock, loff_t off_t_max)
 {
-	if (flock->l_start > COMPAT_OFF_T_MAX)
+	if (flock->l_start > off_t_max)
 		return -EOVERFLOW;
-	if (flock->l_len > COMPAT_OFF_T_MAX)
-		flock->l_len = COMPAT_OFF_T_MAX;
+	if (flock->l_len > off_t_max)
+		flock->l_len = off_t_max;
 	return 0;
 }
 
@@ -631,7 +631,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_OFF_T_MAX);
 		if (err)
 			return err;
 		err = put_compat_flock(&flock, compat_ptr(arg));

fs/fcntl.c COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd, · line 647

Quoted 5 lines
fs/fcntl.c +1 -1
@@ -644,7 +644,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_OFF_T_MAX);

I think you want COMPAT_LOFF_T_MAX here? In any case, I'm fine with the first version, and just renaming the function. I'll plan to push that one unless you have a reason that we should do it this way.

fs/fcntl.c COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd, · line 650

Quoted 3 lines
fs/fcntl.c +0 -0
 		if (err)
 			return err;
 		err = put_compat_flock64(&flock, compat_ptr(arg));

Thanks,

Signature · 1 line
Jeff Layton <[email protected]>
#5 lore raw

Jeff Layton писал 14.11.17 17:06:
...

Quoted 7 lines

fs/fcntl.c COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd, · line 634

Quoted 3 lines
fs/fcntl.c +1 -1
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_OFF_T_MAX);

I think you want COMPAT_LOFF_T_MAX here? In any case, I'm fine with the first version, and just renaming the function. I'll plan to push that one unless you have a reason that we should do it this way.

I would like send v3 with fix the typo you told me. As for me, it is more clean than two functions. Was I wrong with MessageId last time or it is ok to have a new thread for every patch version?

Signature · 3 lines
С уважением,
Виталий Липатов,
Etersoft
#6 lore raw

PATCH v3 1/1 fs/fcntl: restore checking against COMPAT_LOFF_T_MAX for F_GETLK64

In reply to Vitaly Lipatov

for fcntl64 with F_GETLK64 we need use checking against COMPAT_LOFF_T_MAX.

Fixes 94073ad77fff2 "fs/locks: don't mess with the address limit in compat_fcntl64"
---
 fs/fcntl.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
fs/fcntl.c +7 -7
index 30f47d0..e9443d9 100644
@@ -590,17 +590,17 @@ convert_fcntl_cmd(unsigned int cmd)
  * GETLK was successful and we need to return the data, but it needs to fit in
  * the compat structure.
  * l_start shouldn't be too big, unless the original start + end is greater than
- * COMPAT_OFF_T_MAX, in which case the app was asking for trouble, so we return
+ * off_t_max, in which case the app was asking for trouble, so we return
  * -EOVERFLOW in that case.  l_len could be too big, in which case we just
  * truncate it, and only allow the app to see that part of the conflicting lock
  * that might make sense to it anyway
  */
-static int fixup_compat_flock(struct flock *flock)
+static int fixup_compat_flock(struct flock *flock, loff_t off_t_max)
 {
-	if (flock->l_start > COMPAT_OFF_T_MAX)
+	if (flock->l_start > off_t_max)
 		return -EOVERFLOW;
-	if (flock->l_len > COMPAT_OFF_T_MAX)
-		flock->l_len = COMPAT_OFF_T_MAX;
+	if (flock->l_len > off_t_max)
+		flock->l_len = off_t_max;
 	return 0;
 }
 
@@ -631,7 +631,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_OFF_T_MAX);
 		if (err)
 			return err;
 		err = put_compat_flock(&flock, compat_ptr(arg));
@@ -644,7 +644,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_LOFF_T_MAX);
 		if (err)
 			return err;
 		err = put_compat_flock64(&flock, compat_ptr(arg));
Signature · 1 line
2.10.4
#7 lore raw
Quoted 14 lines · Vitaly Lipatov

for fcntl64 with F_GETLK64 we need use checking against COMPAT_LOFF_T_MAX.

Fixes 94073ad77fff2 "fs/locks: don't mess with the address limit in compat_fcntl64"
---
 fs/fcntl.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
fs/fcntl.c +0 -0
index 30f47d0..e9443d9 100644

fs/fcntl.c convert_fcntl_cmd(unsigned int cmd) · line 592

Quoted 3 lines
fs/fcntl.c +0 -0
@@ -590,17 +590,17 @@ convert_fcntl_cmd(unsigned int cmd)
  * GETLK was successful and we need to return the data, but it needs to fit in
  * the compat structure.
  * l_start shouldn't be too big, unless the original start + end is greater than

I assume that should be start + end.

fs/fcntl.c convert_fcntl_cmd(unsigned int cmd)

Quoted 3 lines
fs/fcntl.c +1 -1
- * COMPAT_OFF_T_MAX, in which case the app was asking for trouble, so we return
+ * off_t_max, in which case the app was asking for trouble, so we return
  * -EOVERFLOW in that case.

It took me a minute to understand. OK, I get it, the application's not supposed to issue a GETLK with offset+len too large, so of course it shouldn't encounter a conflicting lock out there.

I don't think that's true, though, thanks to the special interpretation of length 0 in the argument; it looks to me like we can find a conflict with a lock that starts beyond COMPAT_OFF_T_MAX in that case.

I guess that's independent of your patch, though.

--b.

Quoted 37 lines

l_len could be too big, in which case we just
* truncate it, and only allow the app to see that part of the conflicting lock
* that might make sense to it anyway
*/

fs/fcntl.c +6 -6
-static int fixup_compat_flock(struct flock *flock)
+static int fixup_compat_flock(struct flock *flock, loff_t off_t_max)
 {
-	if (flock->l_start > COMPAT_OFF_T_MAX)
+	if (flock->l_start > off_t_max)
 		return -EOVERFLOW;
-	if (flock->l_len > COMPAT_OFF_T_MAX)
-		flock->l_len = COMPAT_OFF_T_MAX;
+	if (flock->l_len > off_t_max)
+		flock->l_len = off_t_max;
 	return 0;
 }
@@ -631,7 +631,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_OFF_T_MAX);
 		if (err)
 			return err;
 		err = put_compat_flock(&flock, compat_ptr(arg));
@@ -644,7 +644,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_LOFF_T_MAX);
 		if (err)
 			return err;
 		err = put_compat_flock64(&flock, compat_ptr(arg));
Signature · 1 line
2.10.4
maintainer
#8 lore raw
In reply to Vitaly Lipatov
Quoted 26 lines · Vitaly Lipatov

for fcntl64 with F_GETLK64 we need use checking against COMPAT_LOFF_T_MAX.

Fixes 94073ad77fff2 "fs/locks: don't mess with the address limit in compat_fcntl64"
---
 fs/fcntl.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)
fs/fcntl.c +2 -2
index 30f47d0..e9443d9 100644
@@ -590,17 +590,17 @@ convert_fcntl_cmd(unsigned int cmd)
  * GETLK was successful and we need to return the data, but it needs to fit in
  * the compat structure.
  * l_start shouldn't be too big, unless the original start + end is greater than
- * COMPAT_OFF_T_MAX, in which case the app was asking for trouble, so we return
+ * off_t_max, in which case the app was asking for trouble, so we return
  * -EOVERFLOW in that case.  l_len could be too big, in which case we just
  * truncate it, and only allow the app to see that part of the conflicting lock
  * that might make sense to it anyway
  */
-static int fixup_compat_flock(struct flock *flock)
+static int fixup_compat_flock(struct flock *flock, loff_t off_t_max)
 {

fs/fcntl.c convert_fcntl_cmd(unsigned int cmd) · line 606

Quoted 10 lines
fs/fcntl.c +3 -3
-	if (flock->l_start > COMPAT_OFF_T_MAX)
+	if (flock->l_start > off_t_max)
 		return -EOVERFLOW;
-	if (flock->l_len > COMPAT_OFF_T_MAX)
-		flock->l_len = COMPAT_OFF_T_MAX;
+	if (flock->l_len > off_t_max)
+		flock->l_len = off_t_max;
 	return 0;
 }
 

Wait...

Does this do anything at all in the case where you pass in COMPAT_LOFF_T_MAX? l_start and l_len are either off_t or loff_t (depending on arch).

Either one will fit in the F_GETLK64/F_OFD_GETLK struct, so I don't see a need to check here.

Quoted 11 lines
fs/fcntl.c +1 -1
@@ -631,7 +631,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_OFF_T_MAX);
 		if (err)
 			return err;
 		err = put_compat_flock(&flock, compat_ptr(arg));

fs/fcntl.c COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd, · line 650

Quoted 8 lines
fs/fcntl.c +1 -1
@@ -644,7 +644,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_LOFF_T_MAX);
 		if (err)
 			return err;
 		err = put_compat_flock64(&flock, compat_ptr(arg));

Maybe a simpler fix would be to just remove the fixup_compat_flock call above?

PS: if you send any more patches, please cc Christoph. He did the earlier work on cleaning up the compat syscall code, and I'd like him to be kept in the loop on this as well.

Thanks,

Signature · 1 line
Jeff Layton <[email protected]>
#9 lore raw

Jeff Layton писал 14.11.17 22:12:
...

Quoted 8 lines

Wait...

Does this do anything at all in the case where you pass in COMPAT_LOFF_T_MAX? l_start and l_len are either off_t or loff_t (depending on arch).

Either one will fit in the F_GETLK64/F_OFD_GETLK struct, so I don't see a need to check here.

I am not sure, can off_t be bigger than loff_t ?
If not, we have just skip checking against COMPAT_LOFF_T_MAX.

...

Quoted 14 lines
Quoted 10 lines
fs/fcntl.c +1 -1
@@ -644,7 +644,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, 
 unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_LOFF_T_MAX);
 		if (err)
 			return err;

err = put_compat_flock64(&flock, compat_ptr(arg));

Maybe a simpler fix would be to just remove the fixup_compat_flock call above?

Quoted 1 line

PS: if you send any more patches, please cc Christoph. He did the

Ok.

Signature · 3 lines
С уважением,
Виталий Липатов,
Etersoft
maintainer
#10 lore raw
Quoted 4 lines · Vitaly Lipatov

Jeff Layton писал 14.11.17 22:12:
...

Quoted 2 lines

Wait...

Quoted 8 lines
Quoted 6 lines

Does this do anything at all in the case where you pass in COMPAT_LOFF_T_MAX? l_start and l_len are either off_t or loff_t (depending on arch).

Either one will fit in the F_GETLK64/F_OFD_GETLK struct, so I don't see a need to check here.

I am not sure, can off_t be bigger than loff_t ?

I don't think so, at least not in any possible situation we care about here.

Quoted 17 lines

If not, we have just skip checking against COMPAT_LOFF_T_MAX.

...

Quoted 14 lines
Quoted 10 lines
fs/fcntl.c +1 -1
@@ -644,7 +644,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd, 
 unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_LOFF_T_MAX);
 		if (err)
 			return err;

err = put_compat_flock64(&flock, compat_ptr(arg));

Maybe a simpler fix would be to just remove the fixup_compat_flock call above?

Ok. If you have a test for this, mind testing and sending a patch?

Thanks,

Signature · 1 line
Jeff Layton <[email protected]>
#11 lore raw

Jeff Layton писал 14.11.17 23:19:

Quoted 9 lines · Vitaly Lipatov
Quoted 8 lines

Jeff Layton писал 14.11.17 22:12:
...

Quoted 6 lines

Wait...

Does this do anything at all in the case where you pass in COMPAT_LOFF_T_MAX? l_start and l_len are either off_t or loff_t (depending on arch).

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

Either one will fit in the F_GETLK64/F_OFD_GETLK struct, so I don't see a need to check here.

I am not sure, can off_t be bigger than loff_t ?

I don't think so, at least not in any possible situation we care about here.

We have this checking for ages:
if (f.l_start > COMPAT_LOFF_T_MAX)
ret = -EOVERFLOW;
http://debian.securedservers.com/kernel/pub/linux/kernel/people/akpm/patches/2.6/2.6.15-rc5/2.6.15-rc5-mm1/broken-out/fix-overflow-tests-for-compat_sys_fcntl64-locking.patch

Quoted 19 lines
Quoted 18 lines

If not, we have just skip checking against COMPAT_LOFF_T_MAX.

...

Quoted 14 lines
Quoted 10 lines
fs/fcntl.c +1 -1
@@ -644,7 +644,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd,
 unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_LOFF_T_MAX);
 		if (err)
 			return err;

err = put_compat_flock64(&flock, compat_ptr(arg));

Maybe a simpler fix would be to just remove the fixup_compat_flock call above?

Quoted 1 line

Ok. If you have a test for this, mind testing and sending a patch?

I think if COMPAT_LOFF_T_MAX is exists, that value can be smaller than can fit in off_t. Checking against COMPAT_LOFF_T_MAX keep old logic works for me last 10 years.

I have some tests around wine project I worked on. May be later I will do additional tests.

Signature · 3 lines
С уважением,
Виталий Липатов,
Etersoft
maintainer
#12 lore raw
Quoted 15 lines · Vitaly Lipatov

Jeff Layton писал 14.11.17 23:19:

Quoted 14 lines · Vitaly Lipatov
Quoted 12 lines

Jeff Layton писал 14.11.17 22:12:
...

Quoted 8 lines

Wait...

Does this do anything at all in the case where you pass in COMPAT_LOFF_T_MAX? l_start and l_len are either off_t or loff_t (depending on arch).

Either one will fit in the F_GETLK64/F_OFD_GETLK struct, so I don't see a need to check here.

I am not sure, can off_t be bigger than loff_t ?

Quoted 8 lines
Quoted 2 lines

I don't think so, at least not in any possible situation we care about here.

We have this checking for ages:
if (f.l_start > COMPAT_LOFF_T_MAX)
ret = -EOVERFLOW;
http://debian.securedservers.com/kernel/pub/linux/kernel/people/akpm/patches/2.6/2.6.15-rc5/2.6.15-rc5-mm1/broken-out/fix-overflow-tests-for-compat_sys_fcntl64-locking.patch

I'm not convinced that those checks ever did anything, tbh.

Quoted 21 lines
Quoted 19 lines
Quoted 18 lines

If not, we have just skip checking against COMPAT_LOFF_T_MAX.

...

Quoted 14 lines
Quoted 10 lines
fs/fcntl.c +1 -1
@@ -644,7 +644,7 @@ COMPAT_SYSCALL_DEFINE3(fcntl64, unsigned int, fd,
 unsigned int, cmd,
 		err = fcntl_getlk(f.file, convert_fcntl_cmd(cmd), &flock);
 		if (err)
 			break;
-		err = fixup_compat_flock(&flock);
+		err = fixup_compat_flock(&flock, COMPAT_LOFF_T_MAX);
 		if (err)
 			return err;

err = put_compat_flock64(&flock, compat_ptr(arg));

Maybe a simpler fix would be to just remove the fixup_compat_flock call above?

Quoted 1 line

Ok. If you have a test for this, mind testing and sending a patch?

Quoted 8 lines

I think if COMPAT_LOFF_T_MAX is exists, that value can be smaller than can fit in off_t.
Checking against COMPAT_LOFF_T_MAX keep old logic works for me last 10 years.

I have some tests around wine project I worked on. May be later I will do additional tests.

I am making an assumption here that l_start and l_end can never be larger than a signed 64-bit value. I don't see how it ever could be, given that it's defined as a long long, but I suppose we could add some exotic arch later that does something weird.

Maybe we can just add a BUILD_BUG_ON for that? I'll send along an alternate patch in a few mins.

Signature · 1 line
Jeff Layton <[email protected]>