Skip to contentSkip to conversation

Recent changes break our update scripts, because of getenv_yesno()

U-Boot Reinhard Meyer 4 participants 7 messages Dec 31, 2010

lore
Share & tools
lore
#0 lore raw

Hello,

in common/image.c:

int getenv_yesno (char *var)

{
	char *s = getenv (var);
	return (s && (*s == 'n')) ? 0 : 1;
}

Is that supposed to return TRUE when the env variable does NOT exist?

Because each TFTP/USB/whatever load of an image will automatically start it!

(and why is such a generic function in image.c?)

Best Regards,
Reinhard

#1 lore raw
Quoted 11 lines · Reinhard Meyer

Hello,

in common/image.c:

int getenv_yesno (char *var)

{
	char *s = getenv (var);
	return (s&&  (*s == 'n')) ? 0 : 1;
}
Quoted 6 lines

Is that supposed to return TRUE when the env variable does NOT exist?

Because each TFTP/USB/whatever load of an image will automatically start it!

(and why is such a generic function in image.c?)

OK, digging deeper, the culprit actually is:

2010-11-28
Mike Frysinger
boot cmds: convert to getenv_yesno() with autostart

Example:
index c657b03..973fa21 100644 (file)

common/cmd_net.c +1 -1
@@ -211,7 +211,7 @@ netboot_common (proto_t proto, cmd_tbl_t *cmdtp, int argc, char * const argv[])
         flush_cache(load_addr, size);
 
         /* Loading ok, check if we should attempt an auto-start */
-       if (((s = getenv("autostart")) != NULL) && (strcmp(s,"yes") == 0)) {
+       if (getenv_yesno("autostart")) {
                 char *local_args[2];
                 local_args[0] = argv[0];
                 local_args[1] = NULL;

Originally, a missing env var resulted in FALSE, with the helper function it results in TRUE!!

Reinhard

#2 lore raw
Quoted 9 lines · Reinhard Meyer
         /* Loading ok, check if we should attempt an auto-start */
common/cmd_net.c +1 -1
-       if (((s = getenv("autostart")) != NULL) && (strcmp(s,"yes") == 0)) {
+       if (getenv_yesno("autostart")) {
                 char *local_args[2];
                 local_args[0] = argv[0];
                 local_args[1] = NULL;

Originally, a missing env var resulted in FALSE, with the helper function it results in TRUE!!

Thanks, this explains why it happens ! We should check again for a missing env var.

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80  Email: office at denx.de
=====================================================================
#3 lore raw
In reply to Reinhard Meyer

Dear Reinhard Meyer,

Quoted 6 lines · In message <[email protected]> you wrote:

OK, digging deeper, the culprit actually is:

2010-11-28
Mike Frysinger
boot cmds: convert to getenv_yesno() with autostart

There has been agreement to revert that commit.

Best regards,

Wolfgang Denk

Signature · 5 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
"Spock, did you see the looks on their faces?"
"Yes, Captain, a sort of vacant contentment."
#4 lore raw

Dear Wolfgang Denk,

Quoted 11 lines

Dear Reinhard Meyer,

Quoted 6 lines · In message<[email protected]> you wrote:

OK, digging deeper, the culprit actually is:

2010-11-28
Mike Frysinger
boot cmds: convert to getenv_yesno() with autostart

There has been agreement to revert that commit.

OK, I must have missed that agreement ;)

What about adding a second parameter to getenv_yesno(), that gives the default value when the env var does not exist?

And maybe move that function into env related source?

Best Regards,
Reinhard

#5 lore raw
Quoted 7 lines · On Friday, December 31, 2010 06:54:17 Reinhard Meyer wrote:
Quoted 7 lines
Quoted 6 lines

OK, digging deeper, the culprit actually is:

2010-11-28

Mike Frysinger
boot cmds: convert to getenv_yesno() with autostart

Quoted 6 lines
Quoted 1 line

There has been agreement to revert that commit.

OK, I must have missed that agreement ;)

What about adding a second parameter to getenv_yesno(), that gives the default value when the env var does not exist?

that's what i proposed, but Wolfgang would rather revert it seems
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
Url : http://lists.denx.de/pipermail/u-boot/attachments/20101231/d6efaf6e/attachment.pgp

#6 lore raw
In reply to Reinhard Meyer
Quoted 6 lines · Reinhard Meyer

Hello,

in common/image.c:

int getenv_yesno (char *var)

Quoted 8 lines
{
	char *s = getenv (var);
	return (s && (*s == 'n')) ? 0 : 1;
}

Is that supposed to return TRUE when the env variable does NOT exist?

Because each TFTP/USB/whatever load of an image will automatically start it!

You are right. I am stucking with the same issue, and inverting the test makes things working again.

However, I see this behavior only with 2010.12 and the suspicious function is really old, I am wondering why we see this behavior only now.

Quoted 2 lines

(and why is such a generic function in image.c?)

Another good point...

Best regards,
Stefano Babic

Signature · 5 lines
=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80  Email: office at denx.de
=====================================================================