Skip to contentSkip to conversation

package/tftpd: fix static build

PATCH New Fabrice Fontaine 3 participants 6 messages Jul 10, 2021 – Jul 18, 2021

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 package/tftpd: fix static build

Fixes:
- http://autobuild.buildroot.org/results/913e8b75422c8beba60df5ea2e2c9e431364566e

---
 .../0002-__progname-is-provided-by-libc.patch | 65 +++++++++++++++++++
 1 file changed, 65 insertions(+)
 create mode 100644 package/tftpd/0002-__progname-is-provided-by-libc.patch
package/tftpd/0002-__progname-is-provided-by-libc.patch new file +65 -0
new file mode 100644
index 0000000000..220fda0646
@@ -0,0 +1,65 @@
+From 18ac1e26f756dd47fef33f5f706b0ec8fa696216 Mon Sep 17 00:00:00 2001
+From: Thorsten Glaser <[email protected]>
+Date: Thu, 31 Jul 2014 16:29:41 +0930
+Subject: __progname[] is provided by libc
+
+Rename local variable to tftpd_progname to avoid a clash with glibc
+global symbols and work around Debian bug #519006 (Closes: #564052).
+
+[ hpa: specifically, double-underscore symbols in C are reserved for
+  the implementation, i.e. compiler/libc. ]
+
+Signed-off-by: Ron Lee <[email protected]>
+Signed-off-by: H. Peter Anvin <[email protected]>
+[Retrieved from:
+https://git.kernel.org/pub/scm/network/tftp/tftp-hpa.git/commit/?id=18ac1e26f756dd47fef33f5f706b0ec8fa696216]
+Signed-off-by: Fabrice Fontaine <[email protected]>
+---
+ tftpd/tftpd.c | 10 +++++-----
+ 1 file changed, 5 insertions(+), 5 deletions(-)
+
+diff --git a/tftpd/tftpd.c b/tftpd/tftpd.c
+index 88d2812..91f5ae1 100644
+--- a/tftpd/tftpd.c
++++ b/tftpd/tftpd.c
+@@ -76,7 +76,7 @@ static int ai_fam = AF_INET;
+ #define TRIES   6               /* Number of attempts to send each packet */
+ #define TIMEOUT_LIMIT ((1 << TRIES)-1)
+ 
+-const char *__progname;
++const char *tftpd_progname;
+ static int peer;
+ static unsigned long timeout  = TIMEOUT;        /* Current timeout value */
+ static unsigned long rexmtval = TIMEOUT;       /* Basic timeout value */
+@@ -387,9 +387,9 @@ int main(int argc, char **argv)
+     /* basename() is way too much of a pain from a portability standpoint */
+ 
+     p = strrchr(argv[0], '/');
+-    __progname = (p && p[1]) ? p + 1 : argv[0];
++    tftpd_progname = (p && p[1]) ? p + 1 : argv[0];
+ 
+-    openlog(__progname, LOG_PID | LOG_NDELAY, LOG_DAEMON);
++    openlog(tftpd_progname, LOG_PID | LOG_NDELAY, LOG_DAEMON);
+ 
+     srand(time(NULL) ^ getpid());
+ 
+@@ -938,14 +938,14 @@ int main(int argc, char **argv)
+        syslog daemon gets restarted by the time we get here. */
+     if (secure && standalone) {
+         closelog();
+-        openlog(__progname, LOG_PID | LOG_NDELAY, LOG_DAEMON);
++        openlog(tftpd_progname, LOG_PID | LOG_NDELAY, LOG_DAEMON);
+     }
+ 
+ #ifdef HAVE_TCPWRAPPERS
+     /* Verify if this was a legal request for us.  This has to be
+        done before the chroot, while /etc is still accessible. */
+     request_init(&wrap_request,
+-                 RQ_DAEMON, __progname,
++                 RQ_DAEMON, tftpd_progname,
+                  RQ_FILE, fd,
+                  RQ_CLIENT_SIN, &from, RQ_SERVER_SIN, &myaddr, 0);
+     sock_methods(&wrap_request);
+-- 
+cgit 1.2.3-1.el7
+
Signature · 1 line
2.30.2
#1 lore raw

PATCH 1/1 package/tftpd: fix static build

Hello Fabrice,

Quoted 8 lines · On Sat, 10 Jul 2021 10:43:13 +0200 Fabrice Fontaine…

Fixes:
- http://autobuild.buildroot.org/results/913e8b75422c8beba60df5ea2e2c9e431364566e

---
 .../0002-__progname-is-provided-by-libc.patch | 65 +++++++++++++++++++
 1 file changed, 65 insertions(+)
 create mode 100644 package/tftpd/0002-__progname-is-provided-by-libc.patch

Thanks for the patch. However as usual, the question is: why is this suddenly happening? There hasn't been any recent update to the tftpd package, so why are we seeing this now? Or this an old issue we had never fixed?

All this investigation is necessary to know if the fix is needed on our LTS branch.

Thanks!

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#2 lore raw

PATCH 1/1 package/tftpd: fix static build

Le ven. 16 juil. 2021 ? 23:21, Thomas Petazzoni
<[email protected]> a ?crit :

Quoted 15 lines

Hello Fabrice,

Quoted 8 lines · On Sat, 10 Jul 2021 10:43:13 +0200 Fabrice Fontaine…

Fixes:
- http://autobuild.buildroot.org/results/913e8b75422c8beba60df5ea2e2c9e431364566e

---
 .../0002-__progname-is-provided-by-libc.patch | 65 +++++++++++++++++++
 1 file changed, 65 insertions(+)
 create mode 100644 package/tftpd/0002-__progname-is-provided-by-libc.patch
Quoted 4 lines

Thanks for the patch. However as usual, the question is: why is this suddenly happening? There hasn't been any recent update to the tftpd package, so why are we seeing this now? Or this an old issue we had never fixed?

I don't know why this is suddenly happening perhaps this is related to a specific uclibc configuration or bug.
I know that uclibc does not always define __progname and had some bugs in the past, for example
https://github.com/wbx-github/uclibc-ng/commit/64a2e37a0edd0b6fadb734b98eb0d828240d485d

However, given that this is an upstream commit that was applied 7 years ago, I think it is pretty safe to apply it to LTS.

Quoted 11 lines

All this investigation is necessary to know if the fix is needed on our LTS branch.

Thanks!

Thomas
--
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

Best Regards,

Fabrice

#3 lore raw

PATCH 1/1 package/tftpd: fix static build

Fabrice, Thomas, All,

On 2021-07-16 23:36 +0200, Fabrice Fontaine spake thusly:

Quoted 15 lines

Le ven. 16 juil. 2021 ? 23:21, Thomas Petazzoni
<[email protected]> a ?crit :

Quoted 9 lines
Quoted 7 lines · On Sat, 10 Jul 2021 10:43:13 +0200 Fabrice Fontaine…

Fixes:
- http://autobuild.buildroot.org/results/913e8b75422c8beba60df5ea2e2c9e431364566e

---
 .../0002-__progname-is-provided-by-libc.patch | 65 +++++++++++++++++++
 1 file changed, 65 insertions(+)
 create mode 100644 package/tftpd/0002-__progname-is-provided-by-libc.patch
Quoted 4 lines

Thanks for the patch. However as usual, the question is: why is this suddenly happening? There hasn't been any recent update to the tftpd package, so why are we seeing this now? Or this an old issue we had never fixed?

Quoted 8 lines

I don't know why this is suddenly happening perhaps this is related to a specific uclibc configuration or bug.
I know that uclibc does not always define __progname and had some bugs in the past, for example
https://github.com/wbx-github/uclibc-ng/commit/64a2e37a0edd0b6fadb734b98eb0d828240d485d

However, given that this is an upstream commit that was applied 7 years ago, I think it is pretty safe to apply it to LTS.

I was confused, because both commits (on tftpd and uClibc) are about 7 years old, so I was wondering which you were referring to...

But yes, you were rerferring to the tftpd commit. Which got me wondering why we needed to apply a patch that have been applied upstream 7 years ago. Why can't we just bump the version rather than backport patches?

Ah, tftpd hasn't had any new release for the past 10 years now, so *are* up-to-date...

But then, yes, why can't we just switch to the latest commit in the git tree, then, rather than backport the fix? Given that tftpd hasn't seen any commit since 2015, we would not take too big a risk...

Regards,
Yann E. MORIN.

Quoted 18 lines
Quoted 11 lines

All this investigation is necessary to know if the fix is needed on our LTS branch.

Thanks!

Thomas
--
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

Best Regards,

Fabrice
_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#4 lore raw

PATCH 1/1 package/tftpd: fix static build

Quoted 10 lines · On Sun, 18 Jul 2021 21:01:36 +0200 "Yann E. MORIN" <[email protected]>…
Quoted 2 lines

However, given that this is an upstream commit that was applied 7 years ago, I think it is pretty safe to apply it to LTS.

I was confused, because both commits (on tftpd and uClibc) are about 7 years old, so I was wondering which you were referring to...

But yes, you were rerferring to the tftpd commit. Which got me wondering why we needed to apply a patch that have been applied upstream 7 years ago. Why can't we just bump the version rather than backport patches?

Quoted 6 lines

Ah, tftpd hasn't had any new release for the past 10 years now, so *are* up-to-date...

But then, yes, why can't we just switch to the latest commit in the git tree, then, rather than backport the fix? Given that tftpd hasn't seen any commit since 2015, we would not take too big a risk...

But besides that, I'm still confused as to why the problem happened only once in the autobuilders. It does happen with the powerpc/internal toolchain/static configuration that failed in the autobuilders, but doesn't fail with the pre-built br-arm-full-static toolchain, which also uses a statically-linked uClibc...

Thomas

Signature · 3 lines
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#5 lore raw

PATCH 1/1 package/tftpd: fix static build

In reply to Fabrice Fontaine

Fabrice, All,

On 2021-07-10 10:43 +0200, Fabrice Fontaine spake thusly:

Quoted 2 lines

Fixes:
- http://autobuild.buildroot.org/results/913e8b75422c8beba60df5ea2e2c9e431364566e

So, after the discussion in this thread, Thomas and I did some investigations, and the situation is a bit murky.

I was able to reproduce the error on 2020.08.3, with this defconfig

BR2_powerpc=y
BR2_STATIC_LIBS=y
BR2_GCC_VERSION_10_X=y
BR2_INIT_NONE=y
BR2_SYSTEM_BIN_SH_NONE=y
# BR2_PACKAGE_BUSYBOX is not set
# BR2_PACKAGE_IFUPDOWN_SCRIPTS is not set
BR2_PACKAGE_TFTPD=y
# BR2_TARGET_ROOTFS_TAR is not set

But I was also able to find an autobuild success, that hapenned on 2021.02.x:

    http://autobuild.buildroot.org/results/ecb/ecbd19e8bd26fe5527c9ce993857d202834cb0fe/

So, it breaks on 2020.08.3, succeeds on 2021.02.x, and then breaks again on the upcoming 2021.08... WTF? tftpd does not like the summer, or what? ;-)

But the difference is that failing configs use gcc-10, while successfull ones use gcc <9 :

    http://autobuild.buildroot.org/index.php?static=1&status=OK&symbols%5BBR2_PACKAGE_TFTPD%5D=y&symbols%5BBR2_GCC_VERSION_9_X%5D=y

So, not sure what's going on there... gcc-10 changed something?

Anyway, the fix is trivially correct, still I think we should just switch over to using the git tree's current HEAD, which gets us this fix and others too...

Could you look into that, please?

Regards,
Yann E. MORIN.

Quoted 84 lines
---
 .../0002-__progname-is-provided-by-libc.patch | 65 +++++++++++++++++++
 1 file changed, 65 insertions(+)
 create mode 100644 package/tftpd/0002-__progname-is-provided-by-libc.patch
package/tftpd/0002-__progname-is-provided-by-libc.patch new file +65 -0
new file mode 100644
index 0000000000..220fda0646
@@ -0,0 +1,65 @@
+From 18ac1e26f756dd47fef33f5f706b0ec8fa696216 Mon Sep 17 00:00:00 2001
+From: Thorsten Glaser <[email protected]>
+Date: Thu, 31 Jul 2014 16:29:41 +0930
+Subject: __progname[] is provided by libc
+
+Rename local variable to tftpd_progname to avoid a clash with glibc
+global symbols and work around Debian bug #519006 (Closes: #564052).
+
+[ hpa: specifically, double-underscore symbols in C are reserved for
+  the implementation, i.e. compiler/libc. ]
+
+Signed-off-by: Ron Lee <[email protected]>
+Signed-off-by: H. Peter Anvin <[email protected]>
+[Retrieved from:
+https://git.kernel.org/pub/scm/network/tftp/tftp-hpa.git/commit/?id=18ac1e26f756dd47fef33f5f706b0ec8fa696216]
+Signed-off-by: Fabrice Fontaine <[email protected]>
+---
+ tftpd/tftpd.c | 10 +++++-----
+ 1 file changed, 5 insertions(+), 5 deletions(-)
+
+diff --git a/tftpd/tftpd.c b/tftpd/tftpd.c
+index 88d2812..91f5ae1 100644
+--- a/tftpd/tftpd.c
++++ b/tftpd/tftpd.c
+@@ -76,7 +76,7 @@ static int ai_fam = AF_INET;
+ #define TRIES   6               /* Number of attempts to send each packet */
+ #define TIMEOUT_LIMIT ((1 << TRIES)-1)
+ 
+-const char *__progname;
++const char *tftpd_progname;
+ static int peer;
+ static unsigned long timeout  = TIMEOUT;        /* Current timeout value */
+ static unsigned long rexmtval = TIMEOUT;       /* Basic timeout value */
+@@ -387,9 +387,9 @@ int main(int argc, char **argv)
+     /* basename() is way too much of a pain from a portability standpoint */
+ 
+     p = strrchr(argv[0], '/');
+-    __progname = (p && p[1]) ? p + 1 : argv[0];
++    tftpd_progname = (p && p[1]) ? p + 1 : argv[0];
+ 
+-    openlog(__progname, LOG_PID | LOG_NDELAY, LOG_DAEMON);
++    openlog(tftpd_progname, LOG_PID | LOG_NDELAY, LOG_DAEMON);
+ 
+     srand(time(NULL) ^ getpid());
+ 
+@@ -938,14 +938,14 @@ int main(int argc, char **argv)
+        syslog daemon gets restarted by the time we get here. */
+     if (secure && standalone) {
+         closelog();
+-        openlog(__progname, LOG_PID | LOG_NDELAY, LOG_DAEMON);
++        openlog(tftpd_progname, LOG_PID | LOG_NDELAY, LOG_DAEMON);
+     }
+ 
+ #ifdef HAVE_TCPWRAPPERS
+     /* Verify if this was a legal request for us.  This has to be
+        done before the chroot, while /etc is still accessible. */
+     request_init(&wrap_request,
+-                 RQ_DAEMON, __progname,
++                 RQ_DAEMON, tftpd_progname,
+                  RQ_FILE, fd,
+                  RQ_CLIENT_SIN, &from, RQ_SERVER_SIN, &myaddr, 0);
+     sock_methods(&wrap_request);
+-- 
+cgit 1.2.3-1.el7
+
Signature · 6 lines
2.30.2

_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot
Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'