Skip to contentSkip to conversation

env export fixes

PATCH New Pierre Aubert 3 participants 10 messages Nov 14, 2013 – Mar 3, 2014 Library code

lore
Share & tools
lore
#0 lore raw

cover env export fixes

These patches fix two issues in the export of environment variables with CRC calculation.

Pierre Aubert (2):
hashtable: fix the export lenght computation.
env export fix: compute the CRC on the real lenght of the exported
variables.

 common/cmd_nvedit.c |    5 +++--
 lib/hashtable.c     |    2 +-
 2 files changed, 4 insertions(+), 3 deletions(-)
Signature · 1 line
1.7.6.5
#1 lore raw

PATCH 1/2 hashtable: fix the export lenght computation.

The room for the '=' and the sep char was reserved twice.

---
 lib/hashtable.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
lib/hashtable.c +1 -1
index 4356b23..480d207 100644
@@ -623,7 +623,7 @@ ssize_t hexport_r(struct hsearch_data *htab, const char sep, int flag,
 
 			list[n++] = ep;
 
-			totlen += strlen(ep->key) + 2;
+			totlen += strlen(ep->key);
 
 			if (sep == '\0') {
 				totlen += strlen(ep->data);
Signature · 1 line
1.7.6.5
#2 lore raw

PATCH 1/2 hashtable: fix the export lenght computation.

Dear Pierre Aubert,

Quoted 1 line · In message <[email protected]> you wrote:

The room for the '=' and the sep char was reserved twice.

Are you sure? Keep in mind that the termination of the list is a _double_ NUL byte. IIRC this is what we do here.

Best regards,

Wolfgang Denk

Signature · 6 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
When properly administered, vacations do not  diminish  productivity:
for every week you're away and get nothing done, there's another when
your boss is away and you get twice as much done.  -- Daniel B. Luten
#3 lore raw

PATCH 1/2 hashtable: fix the export lenght computation.

Dear Wolfgang Denk,

Le 14/11/2013 18:25, Wolfgang Denk a ?crit :

Quoted 6 lines

Dear Pierre Aubert,

Quoted 1 line · In message <[email protected]> you wrote:

The room for the '=' and the sep char was reserved twice.

Are you sure? Keep in mind that the termination of the list is a _double_ NUL byte. IIRC this is what we do here.

Yes, I'm sure. The termination of the list (last NULL after the last separator) is added at line 666

Quoted 1 line

666 size = totlen + 1;

In the actual code, the lenght is two large of 2 char for each exported variable.

Best regards

Quoted 5 lines

Best regards,

Wolfgang Denk

#4 lore raw

PATCH 2/2 env export fix: compute the CRC on the real lenght of the exported variables.

In reply to Pierre Aubert
---
 common/cmd_nvedit.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)
common/cmd_nvedit.c +3 -2
index 5bcc324..c32a932 100644
@@ -922,14 +922,15 @@ NXTARG:		;
 
 	len = hexport_r(&env_htab, '\0',
 			H_MATCH_KEY | H_MATCH_IDENT,
-			&res, ENV_SIZE, argc, argv);
+			&res, size, argc, argv);
+
 	if (len < 0) {
 		error("Cannot export environment: errno = %d\n", errno);
 		return 1;
 	}
 
 	if (chk) {
-		envp->crc = crc32(0, envp->data, ENV_SIZE);
+		envp->crc = crc32(0, envp->data, len);
 #ifdef CONFIG_ENV_ADDR_REDUND
 		envp->flags = ACTIVE_FLAG;
 #endif
Signature · 1 line
1.7.6.5
#5 lore raw

PATCH 2/2 env export fix: compute the CRC on the real lenght of the exported variables.

Dear Pierre Aubert,

Quoted 14 lines · In message <[email protected]> you wrote:
---
 common/cmd_nvedit.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)
common/cmd_nvedit.c +0 -1
index 5bcc324..c32a932 100644
@@ -922,14 +922,15 @@ NXTARG:		;
 
 	len = hexport_r(&env_htab, '\0',
 			H_MATCH_KEY | H_MATCH_IDENT,
-			&res, ENV_SIZE, argc, argv);

common/cmd_nvedit.c NXTARG: ; · line 933

Quoted 10 lines
common/cmd_nvedit.c +3 -1
+			&res, size, argc, argv);
+
 	if (len < 0) {
 		error("Cannot export environment: errno = %d\n", errno);
 		return 1;
 	}
 
 	if (chk) {
-		envp->crc = crc32(0, envp->data, ENV_SIZE);
+		envp->crc = crc32(0, envp->data, len);

This is not correct. When exporting with CRC, then the CRC computation should be the same as is done with the persistently stored environment, i. e. it should be taken over ENV_SIZE bytes.

Best regards,

Wolfgang Denk

Signature · 4 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Hindsight is an exact science.
#6 lore raw

PATCH 2/2 env export fix: compute the CRC on the real lenght of the exported variables.

Dear Wolfgang Denk,

Le 14/11/2013 18:24, Wolfgang Denk a ?crit :

Quoted 24 lines

Dear Pierre Aubert,

Quoted 21 lines · In message <[email protected]> you wrote:
---
  common/cmd_nvedit.c |    5 +++--
  1 files changed, 3 insertions(+), 2 deletions(-)
common/cmd_nvedit.c +2 -1
index 5bcc324..c32a932 100644
@@ -922,14 +922,15 @@ NXTARG:		;
 
  	len = hexport_r(&env_htab, '\0',
  			H_MATCH_KEY | H_MATCH_IDENT,
-			&res, ENV_SIZE, argc, argv);
+			&res, size, argc, argv);
+
  	if (len < 0) {
  		error("Cannot export environment: errno = %d\n", errno);
  		return 1;
  	}
 
Quoted 6 lines

common/cmd_nvedit.c NXTARG: ; · line 933

Quoted 3 lines
common/cmd_nvedit.c +1 -1
  	if (chk) {
-		envp->crc = crc32(0, envp->data, ENV_SIZE);
+		envp->crc = crc32(0, envp->data, len);

This is not correct. When exporting with CRC, then the CRC computation should be the same as is done with the persistently stored environment, i. e. it should be taken over ENV_SIZE bytes.

In this case, there's an inconstisency between the export and the import. It isn't possible to export some variables and the reimport them:
U-Boot > env export -c <addr>
U-Boot > env import -c <same addr> fails with a CRC error.
The import computes the CRC on the real lenght of the environment variables. IMO, as the import and the export are done to work together, it seems to me that the export should compute its CRC on the real lenght.

Best regards

Quoted 5 lines

Best regards,

Wolfgang Denk

#7 lore raw

PATCH 2/2 env export fix: compute the CRC on the real lenght of the exported variables.

[ Catching up on some old emails ]

Quoted 31 lines · Pierre AUBERT

Dear Wolfgang Denk,

Le 14/11/2013 18:24, Wolfgang Denk a ?crit :

Quoted 20 lines

Dear Pierre Aubert,

Quoted 17 lines · In message <[email protected]> you wrote:
---
 common/cmd_nvedit.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)
common/cmd_nvedit.c +2 -1
index 5bcc324..c32a932 100644
@@ -922,14 +922,15 @@ NXTARG:		;
 	len = hexport_r(&env_htab, '\0',
 			H_MATCH_KEY | H_MATCH_IDENT,
-			&res, ENV_SIZE, argc, argv);
+			&res, size, argc, argv);
+
 	if (len < 0) {
 		error("Cannot export environment: errno = %d\n", errno);
Quoted 8 lines

common/cmd_nvedit.c NXTARG: ; · line 933

Quoted 5 lines
common/cmd_nvedit.c +1 -1
 		return 1;
 	}
 	if (chk) {
-		envp->crc = crc32(0, envp->data, ENV_SIZE);
+		envp->crc = crc32(0, envp->data, len);

This is not correct. When exporting with CRC, then the CRC computation should be the same as is done with the persistently stored environment, i. e. it should be taken over ENV_SIZE bytes.

Quoted 9 lines

In this case, there's an inconstisency between the export and the import. It isn't possible to export some variables and the reimport them:
U-Boot > env export -c <addr>
U-Boot > env import -c <same addr> fails with a CRC error. The import computes the CRC on the real lenght of the environment variables. IMO, as the import and the export are done to work together, it seems to me that the export should compute its CRC on the real lenght.

But env import -c <same addr> $filesize does work as the export sets filesize.

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20140226/4f4673e6/attachment.pgp>
#8 lore raw

PATCH 2/2 env export fix: compute the CRC on the real lenght of the exported variables.

Dear Tom Rini

Le 26/02/2014 21:02, Tom Rini a ?crit :

Quoted 43 lines

[ Catching up on some old emails ]

Quoted 31 lines · Pierre AUBERT

Dear Wolfgang Denk,

Le 14/11/2013 18:24, Wolfgang Denk a ?crit :

Quoted 20 lines

Dear Pierre Aubert,

Quoted 17 lines · In message <[email protected]> you wrote:
---
  common/cmd_nvedit.c |    5 +++--
  1 files changed, 3 insertions(+), 2 deletions(-)
common/cmd_nvedit.c +2 -1
index 5bcc324..c32a932 100644
@@ -922,14 +922,15 @@ NXTARG:		;
  	len = hexport_r(&env_htab, '\0',
  			H_MATCH_KEY | H_MATCH_IDENT,
-			&res, ENV_SIZE, argc, argv);
+			&res, size, argc, argv);
+
  	if (len < 0) {
  		error("Cannot export environment: errno = %d\n", errno);
Quoted 8 lines

common/cmd_nvedit.c NXTARG: ; · line 933

Quoted 5 lines
common/cmd_nvedit.c +1 -1
  		return 1;
  	}
  	if (chk) {
-		envp->crc = crc32(0, envp->data, ENV_SIZE);
+		envp->crc = crc32(0, envp->data, len);

This is not correct. When exporting with CRC, then the CRC computation should be the same as is done with the persistently stored environment, i. e. it should be taken over ENV_SIZE bytes.

Quoted 9 lines

In this case, there's an inconstisency between the export and the import. It isn't possible to export some variables and the reimport them:
U-Boot > env export -c <addr>
U-Boot > env import -c <same addr> fails with a CRC error. The import computes the CRC on the real lenght of the environment variables. IMO, as the import and the export are done to work together, it seems to me that the export should compute its CRC on the real lenght.

Quoted 3 lines

But env import -c <same addr> $filesize does work as the export sets filesize.

You're right, but my purpose wasn't to export and reimport immediatly the same environment.
I thought it was more logical to have the same behaviour between "sister" functions.

Best regards

#9 lore raw

PATCH 2/2 env export fix: compute the CRC on the real lenght of the exported variables.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 47 lines · Pierre AUBERT

Dear Tom Rini

Le 26/02/2014 21:02, Tom Rini a ?crit :

Quoted 44 lines

[ Catching up on some old emails ]

Quoted 41 lines · Pierre AUBERT

Dear Wolfgang Denk,

Le 14/11/2013 18:24, Wolfgang Denk a ?crit :

Quoted 21 lines

Dear Pierre Aubert,

In message <[email protected]>

Quoted 17 lines · you
---
  common/cmd_nvedit.c |    5 +++--
  1 files changed, 3 insertions(+), 2 deletions(-)
common/cmd_nvedit.c +2 -1
index 5bcc324..c32a932 100644
@@ -922,14 +922,15 @@ NXTARG:        ;
      len = hexport_r(&env_htab, '\0',
              H_MATCH_KEY | H_MATCH_IDENT,
-            &res, ENV_SIZE, argc, argv);
+            &res, size, argc, argv);
+
      if (len < 0) {
          error("Cannot export environment: errno = %d\n", errno);
Quoted 8 lines

common/cmd_nvedit.c NXTARG: ; · line 933

Quoted 5 lines
common/cmd_nvedit.c +1 -1
          return 1;
      }
      if (chk) {
-        envp->crc = crc32(0, envp->data, ENV_SIZE);
+        envp->crc = crc32(0, envp->data, len);

This is not correct. When exporting with CRC, then the CRC computation should be the same as is done with the persistently stored environment, i. e. it should be taken over ENV_SIZE bytes.

In this case, there's an inconstisency between the export and the import. It isn't possible to export some variables and the reimport them:
U-Boot > env export -c <addr>
U-Boot > env import -c <same addr> fails with a CRC error. The import computes the CRC on the real lenght of the environment variables. IMO, as the import and the export are done to work together, it seems to me that the export should compute its CRC on the real lenght.

Quoted 8 lines
Quoted 3 lines

But env import -c <same addr> $filesize does work as the export sets filesize.

You're right, but my purpose wasn't to export and reimport immediatly the same environment.
I thought it was more logical to have the same behaviour between "sister" functions.

I see what you mean, yes. I think that we need to make the size parameter to import non-optional as assuming a whole ENV_SIZE chunk would have been imported doesn't make sense.

- --
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQIcBAEBAgAGBQJTFLzwAAoJENk4IS6UOR1WM9sP/RriAefiFFrD5GcVURLfD6Ek GDmPWuRg+NFEWHH/Z4eCyvxXaEurLnJcsVF8ZQblWT7zFYu+2weYpLPNYC/zKxzr VDENmOHndPiDRzr5tfkXGiNXYWE46cSVysx5VYznRDH6sDsE+URJxXqZlefBZgw/ 1l0ct5iQUWLmXkRJGQPsj7E4hpedUrBCNavoMh10adLZifWrtty03jN3sOerXhvE qsAbm6ghrvtnC0b9Yj/FcyYFPgw0iG6VhLfCB10U7kwtVK4/Pl+QQNZaaDAPAgkN OkdqiktIa5SyX4l94TBxb2Ixnxe7emZcub1WKrIa3zWNvy744e0bVPmtIVX1vYyT GoaBTG+RcB0fnELt2Uj+odRqQS08KJrSJDOBzIiNzw0VlfSRwaXgK+wTl+3UY0Zt JFiAPeOFGuYndWQJzRf/GKwhaElRV1NjLS1vFN1DHlNLjkv9xF5XJTiED+j7jGYQ UCA26m75KtpAJXBjfE7Y7kyGENP6e9RxwSjwbnJcf0smbpIXi1KyU8bOXyriIK4e oDtprRIxwfL0EdGAlbBNB+RfNxXjB+/lVsVyZyXjmDSeMogYqPIElhRsTMRJzsPx upHUDvSeHqcTW/3nTtkBlnKTRRc+kze+fP3jwSsY06MWwQQxucvLdoqVzPKbQn/E hBWjNm6FMSV0DYbaf1k8
=GS4O
-----END PGP SIGNATURE-----