Skip to contentSkip to conversation

dropbear: Disable DSS correctly

PATCH Superseded [email protected] 3 participants 7 messages Jun 29 – Jul 10

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 dropbear: Disable DSS correctly

From: Mike Crowe <[email protected]>

Take upstream patch that stops sysoptions.h unconditionally turning DROPBEAR_DSS back on even when it has been disabled (which is the default).

---
 ...PBEAR_DSS-is-only-forced-for-fuzzing.patch | 29 +++++++++++++++++++
 .../recipes-core/dropbear/dropbear_2022.83.bb |  1 +
 2 files changed, 30 insertions(+)
 create mode 100644 meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch
meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch new file +29 -0
new file mode 100644
index 0000000000..b9705a2bbb
@@ -0,0 +1,29 @@
+From c7dfaebd5f6a4cde4198f4d2a7baabaa1f632274 Mon Sep 17 00:00:00 2001
+From: Matt Johnston <[email protected]>
+Date: Tue, 6 Dec 2022 22:34:11 +0800
+Subject: [PATCH] Fix so DROPBEAR_DSS is only forced for fuzzing
+
+Regression from 787391ea3b5af2acf5e3c83372510f0c79477ad7,
+was missing fuzzing conditional
+
+Upstream-Status: Backport [https://github.com/mkj/dropbear/commit/c043efb47c3173072fa636ca0da0d19875d4511f]
+---
+ sysoptions.h | 2 ++
+ 1 file changed, 2 insertions(+)
+
+diff --git a/sysoptions.h b/sysoptions.h
+index fb6adc7..12db59c 100644
+--- a/sysoptions.h
++++ b/sysoptions.h
+@@ -383,9 +383,11 @@
+ #endif
+ 
+ /* Fuzzing expects all key types to be enabled */
++#if DROPBEAR_FUZZ
+ #if defined(DROPBEAR_DSS)
+ #undef DROPBEAR_DSS
+ #endif
+ #define DROPBEAR_DSS 1
++#endif
+ 
+ /* no include guard for this file */
meta/recipes-core/dropbear/dropbear_2022.83.bb +1 -0
index 93563aa3b4..d203fee34b 100644
@@ -28,6 +28,7 @@ SRC_URI = "http://matt.ucc.asn.au/dropbear/releases/dropbear-${PV}.tar.bz2 \
            file://0001-Handle-arbitrary-length-paths-and-commands-in-multih.patch \
            file://0001-cli-runopts.c-add-missing-DROPBEAR_CLI_PUBKEY_AUTH.patch \
            file://0001-Avoid-unused-variable-with-DROPBEAR_CLI_PUBKEY_AUTH-.patch \
+           file://0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch \
            file://CVE-2025-47203.patch \
            file://CVE-2019-6111.patch \
            "
Signature · 1 line
2.47.3
#1 lore raw

Hello,

Quoted 19 lines · Mike Crowe via lists.openembedded.org

From: Mike Crowe <[email protected]>

Take upstream patch that stops sysoptions.h unconditionally turning DROPBEAR_DSS back on even when it has been disabled (which is the default).

---
 ...PBEAR_DSS-is-only-forced-for-fuzzing.patch | 29 +++++++++++++++++++
 .../recipes-core/dropbear/dropbear_2022.83.bb |  1 +
 2 files changed, 30 insertions(+)
 create mode 100644 meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch
meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch new file +0 -0
new file mode 100644
index 0000000000..b9705a2bbb

meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch line 9

Quoted 9 lines
meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch new file +9 -0
@@ -0,0 +1,29 @@
+From c7dfaebd5f6a4cde4198f4d2a7baabaa1f632274 Mon Sep 17 00:00:00 2001
+From: Matt Johnston <[email protected]>
+Date: Tue, 6 Dec 2022 22:34:11 +0800
+Subject: [PATCH] Fix so DROPBEAR_DSS is only forced for fuzzing
+
+Regression from 787391ea3b5af2acf5e3c83372510f0c79477ad7,
+was missing fuzzing conditional
+
+Upstream-Status: Backport [https://github.com/mkj/dropbear/commit/c043efb47c3173072fa636ca0da0d19875d4511f]

This patch is missing your signed-off-by.

meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch line 13

Quoted 4 lines
meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch new file +4 -0
+---
+ sysoptions.h | 2 ++
+ 1 file changed, 2 insertions(+)
+

Regards,

Signature · 1 line
Yoann Congal
author
#2 lore raw

Hi Yoann,

Quoted 23 lines · Yoann Congal
Quoted 22 lines · Mike Crowe via lists.openembedded.org

From: Mike Crowe <[email protected]>

Take upstream patch that stops sysoptions.h unconditionally turning DROPBEAR_DSS back on even when it has been disabled (which is the default).

---
 ...PBEAR_DSS-is-only-forced-for-fuzzing.patch | 29 +++++++++++++++++++
 .../recipes-core/dropbear/dropbear_2022.83.bb |  1 +
 2 files changed, 30 insertions(+)
 create mode 100644 meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch
meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch new file +3 -0
new file mode 100644
index 0000000000..b9705a2bbb
@@ -0,0 +1,29 @@
+From c7dfaebd5f6a4cde4198f4d2a7baabaa1f632274 Mon Sep 17 00:00:00 2001
+From: Matt Johnston <[email protected]>
+Date: Tue, 6 Dec 2022 22:34:11 +0800
Quoted 8 lines

meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch line 9

Quoted 6 lines
meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch new file +6 -0
+Subject: [PATCH] Fix so DROPBEAR_DSS is only forced for fuzzing
+
+Regression from 787391ea3b5af2acf5e3c83372510f0c79477ad7,
+was missing fuzzing conditional
+
+Upstream-Status: Backport [https://github.com/mkj/dropbear/commit/c043efb47c3173072fa636ca0da0d19875d4511f]

This patch is missing your signed-off-by.

Well, it's not my patch. :)

I wasn't aware that I was expected to add one to the patch itself for a backport.

https://docs.yoctoproject.org/scarthgap/contributor-guide/recipe-style-guide.html
doesn't seem to say that a SoB should be added to the patch. It does say:

| Of course, if another person later takes care of submitting this patch
| upstream, the status should be changed to Submitted [where], and an
| additional Signed-off-by: line should be added to the patch by the person
| claiming responsibility for upstreaming.

which sort of implies that a SoB should be added for original patches that are suitable for upstreaming but this patch is a backport which originated upstream (without a SoB) which isn't the same thing.

Adding a SoB to an unmodified upstream backport patch feels like appropriation on my part, but I can do so if that's the policy.

Thanks.

Mike.

#3 lore raw
Quoted 49 lines · Mike Crowe via lists.openembedded.org

Hi Yoann,

Quoted 23 lines · Yoann Congal
Quoted 22 lines · Mike Crowe via lists.openembedded.org

From: Mike Crowe <[email protected]>

Take upstream patch that stops sysoptions.h unconditionally turning DROPBEAR_DSS back on even when it has been disabled (which is the default).

---
 ...PBEAR_DSS-is-only-forced-for-fuzzing.patch | 29 +++++++++++++++++++
 .../recipes-core/dropbear/dropbear_2022.83.bb |  1 +
 2 files changed, 30 insertions(+)
 create mode 100644 meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch
meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch new file +3 -0
new file mode 100644
index 0000000000..b9705a2bbb
@@ -0,0 +1,29 @@
+From c7dfaebd5f6a4cde4198f4d2a7baabaa1f632274 Mon Sep 17 00:00:00 2001
+From: Matt Johnston <[email protected]>
+Date: Tue, 6 Dec 2022 22:34:11 +0800
Quoted 8 lines

meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch line 9

Quoted 6 lines
meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch new file +6 -0
+Subject: [PATCH] Fix so DROPBEAR_DSS is only forced for fuzzing
+
+Regression from 787391ea3b5af2acf5e3c83372510f0c79477ad7,
+was missing fuzzing conditional
+
+Upstream-Status: Backport [https://github.com/mkj/dropbear/commit/c043efb47c3173072fa636ca0da0d19875d4511f]

This patch is missing your signed-off-by.

Well, it's not my patch. :)

I wasn't aware that I was expected to add one to the patch itself for a backport.

https://docs.yoctoproject.org/scarthgap/contributor-guide/recipe-style-guide.html
doesn't seem to say that a SoB should be added to the patch. It does say:

| Of course, if another person later takes care of submitting this patch
| upstream, the status should be changed to Submitted [where], and an
| additional Signed-off-by: line should be added to the patch by the person
| claiming responsibility for upstreaming.
Quoted 7 lines

which sort of implies that a SoB should be added for original patches that are suitable for upstreaming but this patch is a backport which originated upstream (without a SoB) which isn't the same thing.

Adding a SoB to an unmodified upstream backport patch feels like appropriation on my part, but I can do so if that's the policy.

You're certifying that the patch you're submitting is appropriately licensed; it says nothing about who authored it.

Signature · 1 line
Alex Kiernan
#4 lore raw
Quoted 55 lines · Alex Kiernan
Quoted 53 lines · Mike Crowe via lists.openembedded.org

Hi Yoann,

Quoted 23 lines · Yoann Congal
Quoted 22 lines · Mike Crowe via lists.openembedded.org

From: Mike Crowe <[email protected]>

Take upstream patch that stops sysoptions.h unconditionally turning DROPBEAR_DSS back on even when it has been disabled (which is the default).

---
 ...PBEAR_DSS-is-only-forced-for-fuzzing.patch | 29 +++++++++++++++++++
 .../recipes-core/dropbear/dropbear_2022.83.bb |  1 +
 2 files changed, 30 insertions(+)
 create mode 100644 meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch
meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch new file +3 -0
new file mode 100644
index 0000000000..b9705a2bbb
@@ -0,0 +1,29 @@
+From c7dfaebd5f6a4cde4198f4d2a7baabaa1f632274 Mon Sep 17 00:00:00 2001
+From: Matt Johnston <[email protected]>
+Date: Tue, 6 Dec 2022 22:34:11 +0800
Quoted 8 lines

meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch line 9

Quoted 6 lines
meta/recipes-core/dropbear/dropbear/0001-Fix-so-DROPBEAR_DSS-is-only-forced-for-fuzzing.patch new file +6 -0
+Subject: [PATCH] Fix so DROPBEAR_DSS is only forced for fuzzing
+
+Regression from 787391ea3b5af2acf5e3c83372510f0c79477ad7,
+was missing fuzzing conditional
+
+Upstream-Status: Backport [https://github.com/mkj/dropbear/commit/c043efb47c3173072fa636ca0da0d19875d4511f]

This patch is missing your signed-off-by.

Well, it's not my patch. :)

I wasn't aware that I was expected to add one to the patch itself for a backport.

https://docs.yoctoproject.org/scarthgap/contributor-guide/recipe-style-guide.html
doesn't seem to say that a SoB should be added to the patch. It does say:

| Of course, if another person later takes care of submitting this patch
| upstream, the status should be changed to Submitted [where], and an
| additional Signed-off-by: line should be added to the patch by the person
| claiming responsibility for upstreaming.

which sort of implies that a SoB should be added for original patches that are suitable for upstreaming but this patch is a backport which originated upstream (without a SoB) which isn't the same thing.

Quoted 6 lines
Quoted 3 lines

Adding a SoB to an unmodified upstream backport patch feels like appropriation on my part, but I can do so if that's the policy.

You're certifying that the patch you're submitting is appropriately licensed; it says nothing about who authored it.

I agree with Alex: As the one requesting the backport, you are the one signing-off on this particular operation. The original author had nothing to do with that.

I created a bug to request clarification in the doc on this point:
16356 – Clarify trailers usage (Signed-off-by, ...) in imported patches
https://bugzilla.yoctoproject.org/show_bug.cgi?id=16356

Regards,

Signature · 2 lines
Yoann Congal
Smile ECS
author
#5 lore raw
Quoted 21 lines · Yoann Congal
Quoted 19 lines · Alex Kiernan
Quoted 17 lines · Mike Crowe via lists.openembedded.org

I wasn't aware that I was expected to add one to the patch itself for a backport.

https://docs.yoctoproject.org/scarthgap/contributor-guide/recipe-style-guide.html
doesn't seem to say that a SoB should be added to the patch. It does say:

| Of course, if another person later takes care of submitting this patch
| upstream, the status should be changed to Submitted [where], and an
| additional Signed-off-by: line should be added to the patch by the person
| claiming responsibility for upstreaming.

which sort of implies that a SoB should be added for original patches that are suitable for upstreaming but this patch is a backport which originated upstream (without a SoB) which isn't the same thing.

Adding a SoB to an unmodified upstream backport patch feels like appropriation on my part, but I can do so if that's the policy.

Quoted 6 lines
Quoted 2 lines

You're certifying that the patch you're submitting is appropriately licensed; it says nothing about who authored it.

I agree with Alex: As the one requesting the backport, you are the one signing-off on this particular operation. The original author had nothing to do with that.

I thought that I did that with the signed-off-by in the openembedded-core commit message.

Quoted 3 lines

I created a bug to request clarification in the doc on this point:
16356 – Clarify trailers usage (Signed-off-by, ...) in imported patches
https://bugzilla.yoctoproject.org/show_bug.cgi?id=16356

Thanks. It feels like this needs an equivalent of oe.qa.check_upstream_status(). Such a check would currently fail since there are lots of existing patches that lack signed-off-bys and adding them would be significant work because each one would need to be manually checked. :( I suppose the check could also allow "Not-signed-off-by:
<[email protected]>" or similar as a workaround for existing patches with a plan to eventually remove it as the patches are updated.

v2 sent with signed-off-by in the patch.

Thanks.

Mike.

#6 lore raw
Quoted 36 lines · Mike Crowe
Quoted 21 lines · Yoann Congal
Quoted 19 lines · Alex Kiernan
Quoted 17 lines · Mike Crowe via lists.openembedded.org

I wasn't aware that I was expected to add one to the patch itself for a backport.

https://docs.yoctoproject.org/scarthgap/contributor-guide/recipe-style-guide.html
doesn't seem to say that a SoB should be added to the patch. It does say:

| Of course, if another person later takes care of submitting this patch
| upstream, the status should be changed to Submitted [where], and an
| additional Signed-off-by: line should be added to the patch by the person
| claiming responsibility for upstreaming.

which sort of implies that a SoB should be added for original patches that are suitable for upstreaming but this patch is a backport which originated upstream (without a SoB) which isn't the same thing.

Adding a SoB to an unmodified upstream backport patch feels like appropriation on my part, but I can do so if that's the policy.

Quoted 6 lines
Quoted 2 lines

You're certifying that the patch you're submitting is appropriately licensed; it says nothing about who authored it.

I agree with Alex: As the one requesting the backport, you are the one signing-off on this particular operation. The original author had nothing to do with that.

I thought that I did that with the signed-off-by in the openembedded-core commit message.

Quoted 3 lines

I created a bug to request clarification in the doc on this point:
16356 – Clarify trailers usage (Signed-off-by, ...) in imported patches
https://bugzilla.yoctoproject.org/show_bug.cgi?id=16356

Quoted 8 lines

Thanks. It feels like this needs an equivalent of oe.qa.check_upstream_status(). Such a check would currently fail since there are lots of existing patches that lack signed-off-bys and adding them would be significant work because each one would need to be manually checked. :( I suppose the check could also allow "Not-signed-off-by:
<[email protected]>" or similar as a workaround for existing patches with a plan to eventually remove it as the patches are updated.

We do have some kind of test in place, in patchtest:
https://git.openembedded.org/openembedded-core/tree/meta/lib/patchtest/tests/test_patch.py#n142
A proper check would be really hard to do without false-positive.

Quoted 2 lines

v2 sent with signed-off-by in the patch.

Thanks! That's appreciated :)

Quoted 4 lines

Thanks.

Mike.

Signature · 2 lines
Yoann Congal
Smile ECS