[MERGED] osmo-pcu[master]: fix PACCH paging: don't return early in case of NULL TBF

[Neels Hofmeyr]

Neels Hofmeyr has submitted this change and it was merged.

Change subject: fix PACCH paging: don't return early in case of NULL TBF
..


fix PACCH paging: don't return early in case of NULL TBF

Commit b78a4a6dfef217c538d45949a6ae725e22a36b05 tried to fix a NULL dereference
error, but apparently was overly eager to return, because it looked like all
code paths would dereference the tbf.

In fact the code path further above, for msg != NULL, has "always" dereferenced
the tbf, but the lower code path, the one effecting the paging, has only
started to dereference tbf since shortly before the overly eager fix: in
da7250ad2c1cd5ddc7d3c6e10435a00b357ef8f7, to "update the dl ctrl msg counter
for ms". It seems that this tbf dereference in the paging path is bogus and the
cause for the segfault that made me write the early exit fix.

Fix that fix:

Do not exit early if tbf == NULL, stay in there to be able to reach the paging
path below.

In case of a message to be sent, assume that tbf is present, and verify: print
an error message and abort if there is a msg but no tbf, so that we will see
the error if I'm wrong there. If a tbf is missing, free the msg.

In case of no message, go on to send pending pagings, but do not attempt to
count ctrl messages for a tbf -- IIUC there will never be a tbf if we're
paging.

This should avoid segfaults while keeping PACCH paging intact.

Tweak a comment for and add a blank line above the paging section.

Related: OS#2176 CID#158969
Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
---
M src/gprs_rlcmac_sched.cpp
1 file changed, 8 insertions(+), 7 deletions(-)

Approvals:
  lynxis lazus: Looks good to me, but someone else must approve
  Neels Hofmeyr: Looks good to me, approved
  Jenkins Builder: Verified



diff --git a/src/gprs_rlcmac_sched.cpp b/src/gprs_rlcmac_sched.cpp
index 3b940f4..97ee53e 100644
--- a/src/gprs_rlcmac_sched.cpp
+++ b/src/gprs_rlcmac_sched.cpp
@@ -178,11 +178,14 @@
}
}
 
-   if (!tbf)
-   return NULL;
-
/* any message */
if (msg) {
+   if (!tbf) {
+   LOGP(DRLCMACSCHED, LOGL_ERROR,
+"Control message to be scheduled, but no TBF 
(TRX=%d, TS=%d)\n", trx, ts);
+   msgb_free(msg);
+   return NULL;
+   }
tbf->rotate_in_list();
LOGP(DRLCMACSCHED, LOGL_DEBUG, "Scheduling control "
"message at RTS for %s (TRX=%d, TS=%d)\n",
@@ -191,14 +194,12 @@
tbf->ms()->update_dl_ctrl_msg();
return msg;
}
-   /* schedule PACKET PAGING REQUEST */
+
+   /* schedule PACKET PAGING REQUEST, if any are pending */
msg = pdch->packet_paging_request();
if (msg) {
LOGP(DRLCMACSCHED, LOGL_DEBUG, "Scheduling paging request "
"message at RTS for (TRX=%d, TS=%d)\n", trx, ts);
-
-   /* Updates the dl ctrl msg counter for ms */
-   tbf->ms()->update_dl_ctrl_msg();
return msg;
}
 

-- 
To view, visit https://gerrit.osmocom.org/2420
To unsubscribe, visit https://gerrit.osmocom.org/settings

Gerrit-MessageType: merged
Gerrit-Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
Gerrit-PatchSet: 4
Gerrit-Project: osmo-pcu
Gerrit-Branch: master
Gerrit-Owner: Neels Hofmeyr 
Gerrit-Reviewer: Harald Welte 
Gerrit-Reviewer: Holger Freyther 
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: Max 
Gerrit-Reviewer: Neels Hofmeyr 
Gerrit-Reviewer: lynxis lazus 
Gerrit-Reviewer: sivasankari 

osmo-pcu[master]: fix PACCH paging: don't return early in case of NULL TBF

[lynxis lazus]

Patch Set 3: Code-Review+1

LGTM

-- 
To view, visit https://gerrit.osmocom.org/2420
To unsubscribe, visit https://gerrit.osmocom.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
Gerrit-PatchSet: 3
Gerrit-Project: osmo-pcu
Gerrit-Branch: master
Gerrit-Owner: Neels Hofmeyr 
Gerrit-Reviewer: Harald Welte 
Gerrit-Reviewer: Holger Freyther 
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: Max 
Gerrit-Reviewer: Neels Hofmeyr 
Gerrit-Reviewer: lynxis lazus 
Gerrit-Reviewer: sivasankari 
Gerrit-HasComments: No

osmo-pcu[master]: fix PACCH paging: don't return early in case of NULL TBF

[Neels Hofmeyr]

Patch Set 2:

So, da7250ad2c1cd5ddc7d3c6e10435a00b357ef8f7 introduced a segfault, and 
b78a4a6dfef217c538d45949a6ae725e22a36b05 as well as this patch fix the 
segfault, while still having problems. To cut this short, we could also revert 
all of the patches. Opinions/better patches welcome.

-- 
To view, visit https://gerrit.osmocom.org/2420
To unsubscribe, visit https://gerrit.osmocom.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
Gerrit-PatchSet: 2
Gerrit-Project: osmo-pcu
Gerrit-Branch: master
Gerrit-Owner: Neels Hofmeyr 
Gerrit-Reviewer: Harald Welte 
Gerrit-Reviewer: Holger Freyther 
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: Max 
Gerrit-Reviewer: Neels Hofmeyr 
Gerrit-HasComments: No

osmo-pcu[master]: fix PACCH paging: don't return early in case of NULL TBF

[Neels Hofmeyr]

Patch Set 2:

(2 comments)

Thanks for the review; the thing is, I don't really know what this code is 
doing and why, this issue jumped at my knees and now I'm trying to throw stones 
at it to not be broken. I would very much welcome a more qualified person to 
take this patch over from me; might be easier than telling me what exactly 
should happen here, probably in numerous iterations of patch review too.

Re how was this found: a live installation of a GSM network saw 80% drop in CS 
service reliability, hunted down to missing pagings for phones that were using 
GPRS. In a pcap, before the mentioned patch you see a lot of PACCH pagings, 
with it you see none. The point being that CS hands down pagings to the PCU in 
case of GPRS being active, and the regression is that we skip all of those 
pagings.

https://gerrit.osmocom.org/#/c/2420/2/src/gprs_rlcmac_sched.cpp
File src/gprs_rlcmac_sched.cpp:

Line 186:   return NULL;
> msg is leaked now?
right...


Line 202:   return msg;
> okay but we might still have a tbf here?
I kind of assumed that we'd either have a TBF and a message or neither... 
right, the code is not checking that...


-- 
To view, visit https://gerrit.osmocom.org/2420
To unsubscribe, visit https://gerrit.osmocom.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
Gerrit-PatchSet: 2
Gerrit-Project: osmo-pcu
Gerrit-Branch: master
Gerrit-Owner: Neels Hofmeyr 
Gerrit-Reviewer: Harald Welte 
Gerrit-Reviewer: Holger Freyther 
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: Max 
Gerrit-Reviewer: Neels Hofmeyr 
Gerrit-HasComments: Yes

osmo-pcu[master]: fix PACCH paging: don't return early in case of NULL TBF

[Max]

Patch Set 2:

Out of curiosity - how this was found? I mean in what way PACCH paging was 
broken as a result of incorrect fix? Is there a way this can be formulated as a 
unit test?

-- 
To view, visit https://gerrit.osmocom.org/2420
To unsubscribe, visit https://gerrit.osmocom.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
Gerrit-PatchSet: 2
Gerrit-Project: osmo-pcu
Gerrit-Branch: master
Gerrit-Owner: Neels Hofmeyr 
Gerrit-Reviewer: Harald Welte 
Gerrit-Reviewer: Jenkins Builder
Gerrit-Reviewer: Max 
Gerrit-HasComments: No

osmo-pcu[master]: fix PACCH paging: don't return early in case of NULL TBF

[Harald Welte]

Patch Set 2: Code-Review+1

-- 
To view, visit https://gerrit.osmocom.org/2420
To unsubscribe, visit https://gerrit.osmocom.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
Gerrit-PatchSet: 2
Gerrit-Project: osmo-pcu
Gerrit-Branch: master
Gerrit-Owner: Neels Hofmeyr 
Gerrit-Reviewer: Harald Welte 
Gerrit-Reviewer: Jenkins Builder
Gerrit-HasComments: No

osmo-pcu[master]: fix PACCH paging: don't return early in case of NULL TBF

[Harald Welte]

Patch Set 2: Code-Review+2

-- 
To view, visit https://gerrit.osmocom.org/2420
To unsubscribe, visit https://gerrit.osmocom.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
Gerrit-PatchSet: 2
Gerrit-Project: osmo-pcu
Gerrit-Branch: master
Gerrit-Owner: Neels Hofmeyr 
Gerrit-Reviewer: Harald Welte 
Gerrit-Reviewer: Jenkins Builder
Gerrit-HasComments: No

[PATCH] osmo-pcu[master]: fix PACCH paging: don't return early in case of NULL TBF

[Neels Hofmeyr]

Hello Jenkins Builder,

I'd like you to reexamine a change.  Please visit

https://gerrit.osmocom.org/2420

to look at the new patch set (#2).

fix PACCH paging: don't return early in case of NULL TBF

Commit b78a4a6dfef217c538d45949a6ae725e22a36b05 tried to fix a NULL dereference
error, but apparently was overly eager to return, because it looked like all
code paths would dereference the tbf.

In fact the code path further above, for msg != NULL, has "always" dereferenced
the tbf, but the lower code path, the one effecting the paging, has only
started to dereference tbf since shortly before the overly eager fix: in
da7250ad2c1cd5ddc7d3c6e10435a00b357ef8f7, to "update the dl ctrl msg counter
for ms". It seems that this tbf dereference in the paging path is bogus and the
cause for the segfault that made me write the early exit fix.

Fix that fix:

Do not exit early if tbf == NULL, stay in there to be able to reach the paging
path below.

In case of a message to be sent, assume that tbf is present, and verify: print
an error message and abort if there is a msg but no tbf, so that we will see
the error if I'm wrong there.

In case of no message, go on to send pending pagings, but do not attempt to
count ctrl messages for a tbf -- IIUC there will never be a tbf if we're
paging.

This should avoid segfaults while keeping PACCH paging intact.

Tweak a comment for and add a blank line above the paging section.

Related: OS#2176 CID#158969
Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
---
M src/gprs_rlcmac_sched.cpp
1 file changed, 7 insertions(+), 7 deletions(-)


  git pull ssh://gerrit.osmocom.org:29418/osmo-pcu refs/changes/20/2420/2

diff --git a/src/gprs_rlcmac_sched.cpp b/src/gprs_rlcmac_sched.cpp
index 3b940f4..b40ff9a 100644
--- a/src/gprs_rlcmac_sched.cpp
+++ b/src/gprs_rlcmac_sched.cpp
@@ -178,11 +178,13 @@
}
}
 
-   if (!tbf)
-   return NULL;
-
/* any message */
if (msg) {
+   if (!tbf) {
+   LOGP(DRLCMACSCHED, LOGL_ERROR,
+"Control message to be scheduled, but no TBF 
(TRX=%d, TS=%d)\n", trx, ts);
+   return NULL;
+   }
tbf->rotate_in_list();
LOGP(DRLCMACSCHED, LOGL_DEBUG, "Scheduling control "
"message at RTS for %s (TRX=%d, TS=%d)\n",
@@ -191,14 +193,12 @@
tbf->ms()->update_dl_ctrl_msg();
return msg;
}
-   /* schedule PACKET PAGING REQUEST */
+
+   /* schedule PACKET PAGING REQUEST, if any are pending */
msg = pdch->packet_paging_request();
if (msg) {
LOGP(DRLCMACSCHED, LOGL_DEBUG, "Scheduling paging request "
"message at RTS for (TRX=%d, TS=%d)\n", trx, ts);
-
-   /* Updates the dl ctrl msg counter for ms */
-   tbf->ms()->update_dl_ctrl_msg();
return msg;
}
 

-- 
To view, visit https://gerrit.osmocom.org/2420
To unsubscribe, visit https://gerrit.osmocom.org/settings

Gerrit-MessageType: newpatchset
Gerrit-Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
Gerrit-PatchSet: 2
Gerrit-Project: osmo-pcu
Gerrit-Branch: master
Gerrit-Owner: Neels Hofmeyr 
Gerrit-Reviewer: Jenkins Builder

[PATCH] osmo-pcu[master]: fix PACCH paging: don't return early in case of NULL TBF

[Neels Hofmeyr]

Review at  https://gerrit.osmocom.org/2420

fix PACCH paging: don't return early in case of NULL TBF

Commit b78a4a6dfef217c538d45949a6ae725e22a36b05 tried to fix a NULL dereference
error, but apparently was overly eager to return, because it looked like all
code paths would dereference the tbf.

In fact the code path further above, for msg != NULL, has "always" dereferenced
the tbf, but the lower code path, the one effecting the paging, has only
started to dereference tbf since shortly before the overly eager fix: in
da7250ad2c1cd5ddc7d3c6e10435a00b357ef8f7, to "update the dl ctrl msg counter
for ms". It seems that this tbf dereference in the paging path is bogus and the
cause for the segfault that made me write the early exit fix.

Fix that fix:

Do not exit early if tbf == NULL, stay in there to be able to reach the paging
path below.

In case of a message to be sent, assume that tbf is present, and verify: print
an error message and abort if there is a msg but no tbf, so that we will see
the error if I'm wrong there.

In case of no message, go on to send pending pagings, but do not attempt to
count ctrl messages for a tbf -- IIUC there will never be a tbf if we're
paging.

This should avoid segfaults while keeping PACCH paging intact.

Tweak a log message for and add a blank line above paging section.

Related: OS#2176 CID#158969
Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
---
M src/gprs_rlcmac_sched.cpp
1 file changed, 7 insertions(+), 7 deletions(-)


  git pull ssh://gerrit.osmocom.org:29418/osmo-pcu refs/changes/20/2420/1

diff --git a/src/gprs_rlcmac_sched.cpp b/src/gprs_rlcmac_sched.cpp
index 3b940f4..b40ff9a 100644
--- a/src/gprs_rlcmac_sched.cpp
+++ b/src/gprs_rlcmac_sched.cpp
@@ -178,11 +178,13 @@
}
}
 
-   if (!tbf)
-   return NULL;
-
/* any message */
if (msg) {
+   if (!tbf) {
+   LOGP(DRLCMACSCHED, LOGL_ERROR,
+"Control message to be scheduled, but no TBF 
(TRX=%d, TS=%d)\n", trx, ts);
+   return NULL;
+   }
tbf->rotate_in_list();
LOGP(DRLCMACSCHED, LOGL_DEBUG, "Scheduling control "
"message at RTS for %s (TRX=%d, TS=%d)\n",
@@ -191,14 +193,12 @@
tbf->ms()->update_dl_ctrl_msg();
return msg;
}
-   /* schedule PACKET PAGING REQUEST */
+
+   /* schedule PACKET PAGING REQUEST, if any are pending */
msg = pdch->packet_paging_request();
if (msg) {
LOGP(DRLCMACSCHED, LOGL_DEBUG, "Scheduling paging request "
"message at RTS for (TRX=%d, TS=%d)\n", trx, ts);
-
-   /* Updates the dl ctrl msg counter for ms */
-   tbf->ms()->update_dl_ctrl_msg();
return msg;
}
 

-- 
To view, visit https://gerrit.osmocom.org/2420
To unsubscribe, visit https://gerrit.osmocom.org/settings

Gerrit-MessageType: newchange
Gerrit-Change-Id: Ib79f4a945e211a13ac7d1e511cc37b0940ac6202
Gerrit-PatchSet: 1
Gerrit-Project: osmo-pcu
Gerrit-Branch: master
Gerrit-Owner: Neels Hofmeyr