Skip to contentSkip to conversation

pseudo: Update to the latest upstream version

PATCH In review Mark Hatle 4 participants 10 messages Jan 14, 2011 – Jan 18, 2011

lore
Share & tools
lore

Mark Hatle

author
#0 lore raw

cover pseudo: Update to the latest upstream version

Update pseudo to the latest version to resolve a number of optimization defects.

The problems were present on both 64-bit and 32-bit systems, but generally only caused issues on 32-bit hosts.

Note: once you upgrade you need to be sure to clean out pseudo. In order to do this ../bitbake/bin/bitbake -c cleanall pseudo-native

Followed by ../bitbake/bin/bitbake pseudo-native

Otherwise you can end up in a wierd situation where the old pseudo is being used to build the new one.

Pull URL: git://git.pokylinux.org/poky-contrib.git
Branch: mhatle/pseudo
Browse: http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=mhatle/pseudo

Thanks,
Mark Hatle <[email protected]>
---

Mark Hatle (1):
pseudo: Update to the latest upstream version

 .../conf/distro/include/poky-default-revisions.inc |    2 +-
 .../895fa7e359e8026a608fba052c2994e42901e45f.patch |  235 --------------------
 meta/recipes-devtools/pseudo/pseudo_git.bb         |    2 +-
 3 files changed, 2 insertions(+), 237 deletions(-)
 delete mode 100644

meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch

Signature · 1 line
1.7.3.4
author
#1 lore raw

PATCH 1/1 pseudo: Update to the latest upstream version

Update pseudo to the latest version to resolve a number of optimization defects.

The problems were present on both 64-bit and 32-bit systems, but generally only caused issues on 32-bit hosts.

---
 .../conf/distro/include/poky-default-revisions.inc |    2 +-
 .../895fa7e359e8026a608fba052c2994e42901e45f.patch |  235 --------------------
 meta/recipes-devtools/pseudo/pseudo_git.bb         |    2 +-
 3 files changed, 2 insertions(+), 237 deletions(-)
 delete mode 100644

meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch

(quoted diff) +0 -0

b/meta/conf/distro/include/poky-default-revisions.inc index 6b98091..413284c 100644

meta/conf/distro/include/poky-default-revisions.inc +1 -1
@@ -150,7 +150,7 @@ PRELINKSRCREV ??= "909470ee441237563d6236c505cb2d02ddc48704"
 SRCREV_pn-prelink ??= "${PRELINKSRCREV}"
 SRCREV_pn-prelink-native ??= "${PRELINKSRCREV}"
 SRCREV_pn-psplash ??= "424"
-PSEUDOSRCREV ??= "c9792c7acdb1cac90549ff08db12a8bf0c53cdcf"
+PSEUDOSRCREV ??= "a1b25a64d8dd012bc8adcac1e1e09b6291c877ad"
 SRCREV_pn-pseudo ??= "${PSEUDOSRCREV}"
 SRCREV_pn-pseudo-native ??= "${PSEUDOSRCREV}"
 SRCREV_pn-pseudo-nativesdk ??= "${PSEUDOSRCREV}"

diff --git a/meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch b/meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch deleted file mode 100644
index d8c2295..0000000
---
a/meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch +++ /dev/null
@@ -1,235 +0,0 @@
-From 895fa7e359e8026a608fba052c2994e42901e45f Mon Sep 17 00:00:00 2001 -From: Mark Hatle <[email protected]>
-Date: Fri, 17 Dec 2010 01:23:47 +0000
-Subject: Revert "Add a cache of the last object found in pseudo_op"
-
-This reverts commit 49d4d35918d457b0e9206679ecad3b9c84f11e66.
-
-Conflicts:
-
- ChangeLog.txt
----
-diff --git a/ChangeLog.txt b/ChangeLog.txt
-index 1db8c7d..200b32e 100644
---- a/ChangeLog.txt
-+++ b/ChangeLog.txt
-@@ -9,7 +9,6 @@
- * (mhatle) as a result of profiling, optimize inode search
- * (mhatle) rearrange the pseudo_op file data operations to reduce
- the number of selects.
-- * (mhatle) add the ability to cache the last select result
- * (mhatle) change the indexing in pseudo_db.c
-
- 2010-12-14:
-diff --git a/pseudo.c b/pseudo.c
-index 3979fd8..6b965be 100644
---- a/pseudo.c
-+++ b/pseudo.c
-@@ -425,9 +425,6 @@ pseudo_op(pseudo_msg_t *msg, const char *program, const char *tag) {
- int found_path = 0, found_ino = 0;
- int prefer_ino = 0;
-
-- static pseudo_msg_t cache_msg = { .op = 0 };
-- static char * cache_path = NULL;
--
- if (!msg)
- return 1;
-
-@@ -471,59 +468,7 @@ pseudo_op(pseudo_msg_t *msg, const char *program, const char *tag) {
-
- /* Lookup the full path, with inode and dev if available */
- if (msg->pathlen && msg->dev && msg->ino) {
-- if (cache_msg.pathlen == msg->pathlen &&

--		    cache_msg.dev == msg->dev &&
--		    cache_msg.ino == msg->ino &&
--		    !strcmp(cache_path, msg->path)) {

-- /* Found it in the cache! */
--#ifdef NPROFILE
-- pseudo_diag("cache hit -- by_file_exact\n");
--#endif
-- /* Change the cache to match the msg */
-- cache_msg.type = msg->type;
-- cache_msg.op = msg->op;
-- cache_msg.result = msg->result;
-- cache_msg.access = msg->access;
-- cache_msg.fd = msg->fd;
--#ifdef NVALIDATE
-- if (!pdb_find_file_exact(msg)) {
-- if (cache_msg.pathlen != msg->pathlen ||

--				    cache_msg.dev != msg->dev ||
--				    cache_msg.ino != msg->ino ||
--				    cache_msg.uid != msg->uid ||
--				    cache_msg.gid != msg->gid ||
--				    cache_msg.mode != msg->mode ||
--				    cache_msg.rdev != msg->rdev ||
--				    cache_msg.nlink != msg->nlink ||
--				    cache_msg.deleting != msg->deleting) {
--				pseudo_diag("Cache mismatch [%s]: Database differs from cache\n"
--						"     cache vs result:\n"
--						"     dev (%llu,%llu) ino (%llu,%llu)\n"
--						"     uid (%d,%d) gid (%d,%d) mode (0%o,0%o)\n"
--						"     rdev (%llu,%llu) nlink (%d,%d) deleting (%d,%d)\n",

-- msg->pathlen ? msg->path : "no path",
-- (unsigned long long) cache_msg.dev, (unsigned long long) msg->dev,
-- (unsigned long long) cache_msg.ino, (unsigned long long) msg->ino,
-- (unsigned long long) cache_msg.uid, (unsigned int) msg->uid,
-- (unsigned long long) cache_msg.gid, (unsigned int) msg->gid,
-- cache_msg.mode, msg->mode,
-- (unsigned long long) cache_msg.rdev, (unsigned int) msg->rdev,
-- cache_msg.nlink, msg->nlink,
-- cache_msg.deleting, msg->deleting);
-- }
-- } else {
-- pseudo_diag("Cache mismatch [%s]: Not found in DB\n",
-- msg->pathlen ? msg->path : "no path");
-- }
--#endif
--
-- by_path = cache_msg;
-- by_ino = cache_msg;
-- /* no need to restore msg */
-- found_path = 1;
-- found_ino = 1;
-- path_by_ino = msg->path;
-- } else if (!pdb_find_file_exact(msg)) {
-+ if (!pdb_find_file_exact(msg)) {
- /* restore header contents */
- by_path = *msg;
- by_ino = *msg;
-@@ -539,51 +484,7 @@ pseudo_op(pseudo_msg_t *msg, const char *program, const char *tag) {
- if (msg->pathlen) {
- /* for now, don't canonicalize paths anymore */
- /* used to do it here, but now doing it in client */
-- if (cache_msg.pathlen == msg->pathlen &&
-- !strcmp(cache_path, msg->path)) {
--#ifdef NPROFILE
-- pseudo_diag("cache hit -- by_path\n");
--#endif
-- /* Change the cache to match the msg */
-- cache_msg.type = msg->type;
-- cache_msg.op = msg->op;
-- cache_msg.result = msg->result;
-- cache_msg.access = msg->access;
-- cache_msg.fd = msg->fd;
--#ifdef NVALIDATE
-- if (!pdb_find_file_exact(msg)) {
-- if (cache_msg.pathlen != msg->pathlen ||

--					    cache_msg.dev != msg->dev ||
--					    cache_msg.ino != msg->ino ||
--					    cache_msg.uid != msg->uid ||
--					    cache_msg.gid != msg->gid ||
--					    cache_msg.mode != msg->mode ||
--					    cache_msg.rdev != msg->rdev ||
--					    cache_msg.nlink != msg->nlink ||
--					    cache_msg.deleting != msg->deleting) {
--					pseudo_diag("Cache mismatch [%s]: Database differs from cache\n"
--						"     cache vs result:\n"
--						"     dev (%llu,%llu) ino (%llu,%llu)\n"
--						"     uid (%d,%d) gid (%d,%d) mode (0%o,0%o)\n"
--						"     rdev (%llu,%llu) nlink (%d,%d) deleting (%d,%d)\n",

-- msg->pathlen ? msg->path : "no path",
-- (unsigned long long) cache_msg.dev, (unsigned long long) msg->dev,
-- (unsigned long long) cache_msg.ino, (unsigned long long) msg->ino,
-- (unsigned long long) cache_msg.uid, (unsigned int) msg->uid,
-- (unsigned long long) cache_msg.gid, (unsigned int) msg->gid,
-- cache_msg.mode, msg->mode,
-- (unsigned long long) cache_msg.rdev, (unsigned int) msg->rdev,
-- cache_msg.nlink, msg->nlink,
-- cache_msg.deleting, msg->deleting);
-- }
-- } else {
-- pseudo_diag("Cache mismatch [%s]: Not found in DB\n",
-- msg->pathlen ? msg->path : "no path");
-- }
--#endif
-- by_path = cache_msg;
-- found_path = 1;
-- } else if (!pdb_find_file_path(msg)) {
-+ if (!pdb_find_file_path(msg)) {
- by_path = *msg;
- found_path = 1;
- } else {
-@@ -594,51 +495,7 @@ pseudo_op(pseudo_msg_t *msg, const char *program, const char *tag) {
- }
- /* search on original inode -- in case of mismatch */
- if (msg->dev && msg->ino) {
-- if (cache_msg.dev == msg->dev &&
-- cache_msg.ino == msg->ino) {
--#ifdef NPROFILE
-- pseudo_diag("cache hit -- by_ino\n");
--#endif
-- /* Change the cache to match the msg */
-- cache_msg.type = msg->type;
-- cache_msg.op = msg->op;
-- cache_msg.result = msg->result;
-- cache_msg.access = msg->access;
-- cache_msg.fd = msg->fd;
--#ifdef NVALIDATE
-- if (!pdb_find_file_exact(msg)) {
-- if (cache_msg.pathlen != msg->pathlen ||

--					    cache_msg.dev != msg->dev ||
--					    cache_msg.ino != msg->ino ||
--					    cache_msg.uid != msg->uid ||
--					    cache_msg.gid != msg->gid ||
--					    cache_msg.mode != msg->mode ||
--					    cache_msg.rdev != msg->rdev ||
--					    cache_msg.nlink != msg->nlink ||
--					    cache_msg.deleting != msg->deleting) {
--					pseudo_diag("Cache mismatch [%s]: Database differs from cache\n"
--						"     cache vs result:\n"
--						"     dev (%llu,%llu) ino (%llu,%llu)\n"
--						"     uid (%d,%d) gid (%d,%d) mode (0%o,0%o)\n"
--						"     rdev (%llu,%llu) nlink (%d,%d) deleting (%d,%d)\n",

-- msg->pathlen ? msg->path : "no path",
-- (unsigned long long) cache_msg.dev, (unsigned long long) msg->dev,
-- (unsigned long long) cache_msg.ino, (unsigned long long) msg->ino,
-- (unsigned long long) cache_msg.uid, (unsigned int) msg->uid,
-- (unsigned long long) cache_msg.gid, (unsigned int) msg->gid,
-- cache_msg.mode, msg->mode,
-- (unsigned long long) cache_msg.rdev, (unsigned int) msg->rdev,
-- cache_msg.nlink, msg->nlink,
-- cache_msg.deleting, msg->deleting);
-- }
-- } else {
-- pseudo_diag("Cache mismatch [%s]: Not found in DB\n",
-- msg->pathlen ? msg->path : "no path");
-- }
--#endif
-- found_ino = 1;
-- path_by_ino = cache_path;
-- } else if (!pdb_find_file_dev(&by_ino)) {
-+ if (!pdb_find_file_dev(&by_ino)) {
- found_ino = 1;
- path_by_ino = pdb_get_file_path(&by_ino);
- }
-@@ -1016,28 +873,6 @@ pseudo_op(pseudo_msg_t *msg, const char *program, const char *tag) {
- break;
- }
-
-- /* Cache previous values... */
-- if (msg->op != OP_MAY_UNLINK &&

--	    msg->op != OP_DID_UNLINK &&
--	    msg->op != OP_CANCEL_UNLINK &&
--	    msg->op != OP_UNLINK) {

--
-- cache_msg = *msg;
--
-- free(cache_path);
--
-- if (path_by_ino)
-- cache_path = strdup(path_by_ino);
-- else
-- cache_path = strdup(msg->path);
-- } else {
-- cache_msg.pathlen = 0;
-- cache_msg.dev = 0;
-- cache_msg.ino = 0;
-- free(cache_path);
-- cache_path = NULL;
-- }
--
- /* in the case of an exact match, we just used the pointer
- * rather than allocating space
- */
---
-cgit v0.8.3.3-89-gbf82

(quoted diff) +0 -0

b/meta/recipes-devtools/pseudo/pseudo_git.bb
index f9b10f0..df73696 100644

meta/recipes-devtools/pseudo/pseudo_git.bb +1 -1
@@ -10,7 +10,7 @@ PR = "r16"
 
 SRC_URI = "git://github.com/wrpseudo/pseudo.git;protocol=git \
            file://static_sqlite.patch \
-           file://895fa7e359e8026a608fba052c2994e42901e45f.patch"
+	  "
 
 FILES_${PN} = "${libdir}/libpseudo.so ${bindir}/* ${localstatedir}/pseudo"
 PROVIDES += "virtual/fakeroot"
Signature · 1 line
1.7.3.4
#2 lore raw
In reply to Mark Hatle
Quoted 26 lines · Mark Hatle

Update pseudo to the latest version to resolve a number of optimization defects.

The problems were present on both 64-bit and 32-bit systems, but generally only caused issues on 32-bit hosts.

Note: once you upgrade you need to be sure to clean out pseudo. In order to do this ../bitbake/bin/bitbake -c cleanall pseudo-native

Followed by ../bitbake/bin/bitbake pseudo-native

Otherwise you can end up in a wierd situation where the old pseudo is being used to build the new one.

Pull URL: git://git.pokylinux.org/poky-contrib.git
Branch: mhatle/pseudo
Browse: http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=mhatle/pseudo

Thanks,
Mark Hatle<[email protected]>
---

Mark Hatle (1):
pseudo: Update to the latest upstream version

Quoted 7 lines
  .../conf/distro/include/poky-default-revisions.inc |    2 +-
  .../895fa7e359e8026a608fba052c2994e42901e45f.patch |  235 --------------------
  meta/recipes-devtools/pseudo/pseudo_git.bb         |    2 +-
  3 files changed, 2 insertions(+), 237 deletions(-)

delete mode 100644
meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch

Mark,

I did not pull this because of a failure seen on the autobuilder, the failure was in eglibc-initial trying to compile.

The error was:

arm-poky-linux-gnueabi-gcc: error trying to exec 'cc1': execvp: No such file or directory

I know that this seems strange for pseudo to cause, but I removed the psuedo change and this problem went away. It could be a build timing issue or an install location issue.

Check the autobuilder log at:
http://autobuilder.pokylinux.org:8010/builders/distro-testing/builds/44/steps/shell_1/logs/stdio

Unfortunately because I rebuild distro-testing the individual logs are gone. Let me know if we need to rebuild with the error for further diagnoses.

Thanks
Sau!

author
#3 lore raw
Quoted 36 lines · Saul Wold
Quoted 26 lines · Mark Hatle

Update pseudo to the latest version to resolve a number of optimization defects.

The problems were present on both 64-bit and 32-bit systems, but generally only caused issues on 32-bit hosts.

Note: once you upgrade you need to be sure to clean out pseudo. In order to do this ../bitbake/bin/bitbake -c cleanall pseudo-native

Followed by ../bitbake/bin/bitbake pseudo-native

Otherwise you can end up in a wierd situation where the old pseudo is being used to build the new one.

Pull URL: git://git.pokylinux.org/poky-contrib.git
Branch: mhatle/pseudo
Browse: http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=mhatle/pseudo

Thanks,
Mark Hatle<[email protected]>
---

Mark Hatle (1):
pseudo: Update to the latest upstream version

Quoted 7 lines
  .../conf/distro/include/poky-default-revisions.inc |    2 +-
  .../895fa7e359e8026a608fba052c2994e42901e45f.patch |  235 --------------------
  meta/recipes-devtools/pseudo/pseudo_git.bb         |    2 +-
  3 files changed, 2 insertions(+), 237 deletions(-)

delete mode 100644
meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch

Mark,

Quoted 7 lines

I did not pull this because of a failure seen on the autobuilder, the failure was in eglibc-initial trying to compile.

The error was:

arm-poky-linux-gnueabi-gcc: error trying to exec 'cc1': execvp: No such file or directory

pseudo does capture execvp. So it's possible something is going wrong there. I'm investigating...

--Mark

Quoted 13 lines

I know that this seems strange for pseudo to cause, but I removed the psuedo change and this problem went away. It could be a build timing issue or an install location issue.

Check the autobuilder log at:
http://autobuilder.pokylinux.org:8010/builders/distro-testing/builds/44/steps/shell_1/logs/stdio

Unfortunately because I rebuild distro-testing the individual logs are gone. Let me know if we need to rebuild with the error for further diagnoses.

Thanks
Sau!

author
#4 lore raw

A potential solution to this problem has been pushed to the poky-contrib mhatle/pseudo tree.

If anyone is interested in trying this, I'd appreciate it. I've so far not been able to reproduce the original problem locally, but based on the symptoms I believe we've got a likely fix in place.

--Mark

Quoted 68 lines · Mark Hatle
Quoted 36 lines · Saul Wold
Quoted 26 lines · Mark Hatle

Update pseudo to the latest version to resolve a number of optimization defects.

The problems were present on both 64-bit and 32-bit systems, but generally only caused issues on 32-bit hosts.

Note: once you upgrade you need to be sure to clean out pseudo. In order to do this ../bitbake/bin/bitbake -c cleanall pseudo-native

Followed by ../bitbake/bin/bitbake pseudo-native

Otherwise you can end up in a wierd situation where the old pseudo is being used to build the new one.

Pull URL: git://git.pokylinux.org/poky-contrib.git
Branch: mhatle/pseudo
Browse: http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=mhatle/pseudo

Thanks,
Mark Hatle<[email protected]>
---

Mark Hatle (1):
pseudo: Update to the latest upstream version

Quoted 7 lines
  .../conf/distro/include/poky-default-revisions.inc |    2 +-
  .../895fa7e359e8026a608fba052c2994e42901e45f.patch |  235 --------------------
  meta/recipes-devtools/pseudo/pseudo_git.bb         |    2 +-
  3 files changed, 2 insertions(+), 237 deletions(-)

delete mode 100644
meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch

Mark,

Quoted 7 lines

I did not pull this because of a failure seen on the autobuilder, the failure was in eglibc-initial trying to compile.

The error was:

arm-poky-linux-gnueabi-gcc: error trying to exec 'cc1': execvp: No such file or directory

pseudo does capture execvp. So it's possible something is going wrong there. I'm investigating...

--Mark

Quoted 7 lines

I know that this seems strange for pseudo to cause, but I removed the psuedo change and this problem went away. It could be a build timing issue or an install location issue.

Check the autobuilder log at:
http://autobuilder.pokylinux.org:8010/builders/distro-testing/builds/44/steps/shell_1/logs/stdio

Quoted 6 lines

Unfortunately because I rebuild distro-testing the individual logs are gone. Let me know if we need to rebuild with the error for further diagnoses.

Thanks
Sau!

_______________________________________________
poky mailing list
[email protected]
https://lists.yoctoproject.org/listinfo/poky

#5 lore raw
Quoted 7 lines · Mark Hatle

A potential solution to this problem has been pushed to the poky-contrib mhatle/pseudo tree.

If anyone is interested in trying this, I'd appreciate it. I've so far not been able to reproduce the original problem locally, but based on the symptoms I believe we've got a likely fix in place.

Mark,

I have this patch in a distro/master right now, it fixes the original problem, but I am seeing another problem with the poky-image-sdk target where I am getting a load of permission denied during the tar of the do_populate_sdk task.

Maybe this need to be added to the pseudo.

Sau!

Quoted 77 lines

--Mark

Quoted 61 lines · Mark Hatle
Quoted 36 lines · Saul Wold
Quoted 26 lines · Mark Hatle

Update pseudo to the latest version to resolve a number of optimization defects.

The problems were present on both 64-bit and 32-bit systems, but generally only caused issues on 32-bit hosts.

Note: once you upgrade you need to be sure to clean out pseudo. In order to do this ../bitbake/bin/bitbake -c cleanall pseudo-native

Followed by ../bitbake/bin/bitbake pseudo-native

Otherwise you can end up in a wierd situation where the old pseudo is being used to build the new one.

Pull URL: git://git.pokylinux.org/poky-contrib.git
Branch: mhatle/pseudo
Browse: http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=mhatle/pseudo

Thanks,
Mark Hatle<[email protected]>
---

Mark Hatle (1):
pseudo: Update to the latest upstream version

Quoted 7 lines
   .../conf/distro/include/poky-default-revisions.inc |    2 +-
   .../895fa7e359e8026a608fba052c2994e42901e45f.patch |  235 --------------------
   meta/recipes-devtools/pseudo/pseudo_git.bb         |    2 +-
   3 files changed, 2 insertions(+), 237 deletions(-)

delete mode 100644
meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch

Mark,

Quoted 7 lines

I did not pull this because of a failure seen on the autobuilder, the failure was in eglibc-initial trying to compile.

The error was:

arm-poky-linux-gnueabi-gcc: error trying to exec 'cc1': execvp: No such file or directory

pseudo does capture execvp. So it's possible something is going wrong there. I'm investigating...

--Mark

Quoted 11 lines

I know that this seems strange for pseudo to cause, but I removed the psuedo change and this problem went away. It could be a build timing issue or an install location issue.

Check the autobuilder log at:
http://autobuilder.pokylinux.org:8010/builders/distro-testing/builds/44/steps/shell_1/logs/stdio

Unfortunately because I rebuild distro-testing the individual logs are gone. Let me know if we need to rebuild with the error for further diagnoses.

Quoted 7 lines
Quoted 2 lines

Thanks
Sau!

_______________________________________________
poky mailing list
[email protected]
https://lists.yoctoproject.org/listinfo/poky

_______________________________________________
poky mailing list
[email protected]
https://lists.yoctoproject.org/listinfo/poky

#6 lore raw
Quoted 8 lines · Saul Wold
Quoted 7 lines · Mark Hatle

A potential solution to this problem has been pushed to the poky-contrib mhatle/pseudo tree.

If anyone is interested in trying this, I'd appreciate it. I've so far not been able to reproduce the original problem locally, but based on the symptoms I believe we've got a likely fix in place.

Quoted 6 lines

Mark,

I have this patch in a distro/master right now, it fixes the original problem, but I am seeing another problem with the poky-image-sdk target where I am getting a load of permission denied during the tar of the do_populate_sdk task.

You mean meta-toolchain-sdk?

Cheers,

Richard

author
#7 lore raw
In reply to Saul Wold
Quoted 8 lines · Saul Wold
Quoted 7 lines · Mark Hatle

A potential solution to this problem has been pushed to the poky-contrib mhatle/pseudo tree.

If anyone is interested in trying this, I'd appreciate it. I've so far not been able to reproduce the original problem locally, but based on the symptoms I believe we've got a likely fix in place.

Quoted 6 lines

Mark,

I have this patch in a distro/master right now, it fixes the original problem, but I am seeing another problem with the poky-image-sdk target where I am getting a load of permission denied during the tar of the do_populate_sdk task.

If it's taring "target" files, then yes, it looks like it's not being properly run under pseudo control and that will need to be updated. [or there is another defect we're unaware of].

What are the steps to reproduce the issue and I'll see what I can do here.

I'll pass along that the exec problem seems to be correct to the pseudo maintainer.

--Mark

Quoted 82 lines

Maybe this need to be added to the pseudo.

Sau!

Quoted 77 lines

--Mark

Quoted 61 lines · Mark Hatle
Quoted 36 lines · Saul Wold
Quoted 26 lines · Mark Hatle

Update pseudo to the latest version to resolve a number of optimization defects.

The problems were present on both 64-bit and 32-bit systems, but generally only caused issues on 32-bit hosts.

Note: once you upgrade you need to be sure to clean out pseudo. In order to do this ../bitbake/bin/bitbake -c cleanall pseudo-native

Followed by ../bitbake/bin/bitbake pseudo-native

Otherwise you can end up in a wierd situation where the old pseudo is being used to build the new one.

Pull URL: git://git.pokylinux.org/poky-contrib.git
Branch: mhatle/pseudo
Browse: http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=mhatle/pseudo

Thanks,
Mark Hatle<[email protected]>
---

Mark Hatle (1):
pseudo: Update to the latest upstream version

Quoted 7 lines
   .../conf/distro/include/poky-default-revisions.inc |    2 +-
   .../895fa7e359e8026a608fba052c2994e42901e45f.patch |  235 --------------------
   meta/recipes-devtools/pseudo/pseudo_git.bb         |    2 +-
   3 files changed, 2 insertions(+), 237 deletions(-)

delete mode 100644
meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch

Mark,

Quoted 7 lines

I did not pull this because of a failure seen on the autobuilder, the failure was in eglibc-initial trying to compile.

The error was:

arm-poky-linux-gnueabi-gcc: error trying to exec 'cc1': execvp: No such file or directory

pseudo does capture execvp. So it's possible something is going wrong there. I'm investigating...

--Mark

Quoted 11 lines

I know that this seems strange for pseudo to cause, but I removed the psuedo change and this problem went away. It could be a build timing issue or an install location issue.

Check the autobuilder log at:
http://autobuilder.pokylinux.org:8010/builders/distro-testing/builds/44/steps/shell_1/logs/stdio

Unfortunately because I rebuild distro-testing the individual logs are gone. Let me know if we need to rebuild with the error for further diagnoses.

Quoted 7 lines
Quoted 2 lines

Thanks
Sau!

_______________________________________________
poky mailing list
[email protected]
https://lists.yoctoproject.org/listinfo/poky

_______________________________________________
poky mailing list
[email protected]
https://lists.yoctoproject.org/listinfo/poky

#8 lore raw
In reply to Mark Hatle

Hi Mark,

The issues you mentioned include the build performance? Recently we saw the bitbake performance is worse than before.

We haven't investigated it yet, and just want to get a quick hint. Thanks!

For some detail performance data, you can refer to the link https://wiki.yoctoproject.org/wiki/Performance

Thanks,
Dongxiao

Quoted 34 lines · Mark Hatle

Update pseudo to the latest version to resolve a number of optimization defects.

The problems were present on both 64-bit and 32-bit systems, but generally only caused issues on 32-bit hosts.

Note: once you upgrade you need to be sure to clean out pseudo. In order to do this ../bitbake/bin/bitbake -c cleanall pseudo-native

Followed by ../bitbake/bin/bitbake pseudo-native

Otherwise you can end up in a wierd situation where the old pseudo is being used to build the new one.

Pull URL: git://git.pokylinux.org/poky-contrib.git
Branch: mhatle/pseudo
Browse:
http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=mhatle/pseudo

Thanks,
Mark Hatle <[email protected]>
---

Mark Hatle (1):
pseudo: Update to the latest upstream version

 .../conf/distro/include/poky-default-revisions.inc |    2 +-
 .../895fa7e359e8026a608fba052c2994e42901e45f.patch |  235

-------------------- meta/recipes-devtools/pseudo/pseudo_git.bb
| 2 +- 3 files changed, 2 insertions(+), 237 deletions(-) delete
mode 100644 meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch

#9 lore raw
In reply to Mark Hatle
Quoted 26 lines · Mark Hatle

Update pseudo to the latest version to resolve a number of optimization defects.

The problems were present on both 64-bit and 32-bit systems, but generally only caused issues on 32-bit hosts.

Note: once you upgrade you need to be sure to clean out pseudo. In order to do this ../bitbake/bin/bitbake -c cleanall pseudo-native

Followed by ../bitbake/bin/bitbake pseudo-native

Otherwise you can end up in a wierd situation where the old pseudo is being used to build the new one.

Pull URL: git://git.pokylinux.org/poky-contrib.git
Branch: mhatle/pseudo
Browse: http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=mhatle/pseudo

Thanks,
Mark Hatle<[email protected]>
---

Mark Hatle (1):
pseudo: Update to the latest upstream version

Quoted 7 lines
  .../conf/distro/include/poky-default-revisions.inc |    2 +-
  .../895fa7e359e8026a608fba052c2994e42901e45f.patch |  235 --------------------
  meta/recipes-devtools/pseudo/pseudo_git.bb         |    2 +-
  3 files changed, 2 insertions(+), 237 deletions(-)

delete mode 100644
meta/recipes-devtools/pseudo/pseudo/895fa7e359e8026a608fba052c2994e42901e45f.patch

Pulled into Master, with the patch

Thanks
Sau!