maint(gpg2john.c): add missing prototypes #5603

claudioandre-br · 2024-12-02T20:27:52Z

The clang compiler was complaining with the error message: "Passing arguments to a function without a prototype is deprecated in all versions of C".

It looks correct to me, so I'm sending it.

claudioandre-br · 2024-12-02T20:39:41Z

bots are happy!

magnumripper · 2024-12-02T20:53:41Z

That's great! I'll have to let @solardiz approve it though, because I really don't understand any of it 😆

solardiz

I made review comments about implementation detail of your approach, but I also question the whole approach.

It'd be simpler to cast the two different-prototype function addresses to (funcptr) in the initializer and then cast them to the right function pointer type in the two calls. A drawback is we wouldn't let the compiler check that we use the right type for those functions there - we'd take responsibility that we do.

solardiz · 2024-12-02T21:29:35Z

src/gpg2john.c

@@ -1295,7 +1295,7 @@ TAG[] = {
 #define TAG_NUM (sizeof(TAG) * sizeof(string))

 private void
-(*tag_func[])() = {
+(*tag_func[])(int) = {


Could use the funcptr type here.

solardiz · 2024-12-02T21:30:52Z

src/gpg2john.c

@@ -1629,7 +1697,7 @@ parse_packet(char *hash)
 		if (tag < TAG_NUM && tag_func[tag] != NULL) {
 			if (gpg_dbg)
 				fprintf(stderr, "Packet type %d, len %d at offset %d  (Processing) (pkt-type %s) (Partial %s)\n", tag, len, offset, pkt_type(tag), partial?"yes":"no");
-			(*tag_func[tag])(len, 1, partial, hash);	// first packet (possibly only one if partial is false).
+			(*tag_func1[tag])(len, 1, partial, hash);	// first packet (possibly only one if partial is false).


We're now calling through a tag_func1 element, but the NULL check above remained of tag_func - that's probably wrong now.

solardiz · 2024-12-02T21:31:01Z

src/gpg2john.c

@@ -1651,7 +1719,7 @@ parse_packet(char *hash)
 			if (tag < TAG_NUM && tag_func[tag] != NULL) {
 				if (gpg_dbg)
 					fprintf(stderr, "Packet type %d, len %d at offset %d  (Processing) (pkt-type %s) (Partial %s)\n", tag, len, offset, pkt_type(tag), partial?"yes":"no");
-				(*tag_func[tag])(len, 0, partial, hash);	// subsquent packets.
+				(*tag_func1[tag])(len, 0, partial, hash);	// subsquent packets.


claudioandre-br · 2024-12-02T22:04:13Z

tag_func is always called using the 4 parameters. Therefore, only the new tag_func1 is needed (it seems).

magnumripper · 2024-12-03T09:42:20Z

It would be nice IMHO if you reference #5268 if you force-push again. Better traceability when trying to understand historical changes two years from now.

claudioandre-br · 2024-12-03T14:16:38Z

tag_func is always called using the 4 parameters. Therefore, only the new tag_func1 is needed (it seems).

No. Let's rephrase it: it calls some functions that expect only one parameter, but it passes 4 parameters.

claudioandre-br · 2024-12-03T14:43:20Z

Testing now (from john-samples):

Committed code:

john-the-ripper.gpg2john GPG/test.utf8.asc > 1.hash
john-the-ripper.gpg2john GPG/gpg-70_flavors/*.sec >> 1.hash
cat 1.hash | LANG=C sort > old.hash

PR:

../run/gpg2john GPG/test.utf8.asc > 2.hash
../run/gpg2john GPG/gpg-70_flavors/*.sec >> 2.hash
cat 2.hash | LANG=C sort > pr.hash

diff old.hash pr.hash

claudioandre-br · 2024-12-03T14:53:44Z

Both versions are producing the same result (see the tests above).

claudioandre-br · 2024-12-03T15:05:30Z

src/gpg2john.c

+			if (tag_func[tag] != NULL)
+				(*tag_func[tag])(len);
+			else
+				(*tag_func4[tag])(len, 1, partial, hash);	// first packet (possibly only one if partial is false).


None of the john-samples examples are calling this else part. Anyway, the call to the method is correct.

claudioandre-br · 2024-12-03T15:05:43Z

src/gpg2john.c

+				if (tag_func[tag] != NULL)
+					(*tag_func[tag])(len);
+				else
+					(*tag_func4[tag])(len, 0, partial, hash);	// subsquent packets.


The clang compiler was complaining with the error message: "Passing arguments to a function without a prototype is deprecated in all versions of C". Fix: openwall#5268. Signed-off-by: Claudio André <[email protected]>

claudioandre-br · 2024-12-03T18:23:08Z

I created a file to use the symmetric "stuff":

$ gpg -a --symmetric --cipher-algo AES256 old.hash 

$ ../run/gpg2john old.hash.asc

This first “else” was executed once. The second was called 6 times. The same output was produced and it is crackable.

solardiz

Let's figure out the discrepancy in pointer placement/counts.

solardiz · 2024-12-04T03:11:43Z

src/gpg2john.c

+	NULL,
+	NULL,
+	NULL,
+	Symmetrically_Encrypted_and_MDC_Packet,


The functions removed from the old/first array are separated by 8 other pointers, but here they're separated by only 7 NULLs. A bug?

Separately (but maybe related), I notice that the old array initializer had 65 lines, whereas the tag mask is such that there should probably be only 64 - a bug? Harmless?

The first commit (f4df772) of gpg2john had that non-multiple of 4. Jim added a second typedef in 2fee9ee for #477.

solardiz · 2024-12-04T03:14:37Z

This first “else” was executed once. The second was called 6 times. The same output was produced and it is crackable.

That's cool. No idea how it's possible if I counted the pointers between the two 4-argument functions right. Was the code buggy prior to your changes? Did gpg2john produce the same or different output before this PR's changes?

Anyhow, please add your new test input to the john-samples repo. Thank you!

claudioandre-br · 2024-12-04T17:16:37Z

I'll push (soon) it again to see if anything is missing!

solardiz · 2024-12-05T17:56:28Z

@claudioandre-br I see you've inserted the missing NULL now. However, that this bug went unnoticed means we have no test case for this functionality (the function that was presumably at the wrong index before this NULL addition), right?

claudioandre-br · 2024-12-06T06:40:43Z

My local version has been tested ('íf' and 'else'). But let's keep this waiting until i can test it again.

There weren't supposed to be two different versions.

solardiz requested changes Dec 2, 2024

View reviewed changes

claudioandre-br force-pushed the fix/clang branch from 5e8bffe to 0251a0f Compare December 2, 2024 22:00

claudioandre-br force-pushed the fix/clang branch 2 times, most recently from 4289de9 to 0b01ba0 Compare December 3, 2024 09:52

claudioandre-br force-pushed the fix/clang branch from 0b01ba0 to 6614733 Compare December 3, 2024 14:55

claudioandre-br commented Dec 3, 2024

View reviewed changes

maint(gpg2john.c): add missing prototypes

79e8b91

The clang compiler was complaining with the error message: "Passing arguments to a function without a prototype is deprecated in all versions of C". Fix: openwall#5268. Signed-off-by: Claudio André <[email protected]>

solardiz requested changes Dec 4, 2024

View reviewed changes

claudioandre-br force-pushed the fix/clang branch from 6614733 to 79e8b91 Compare December 5, 2024 10:03

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

maint(gpg2john.c): add missing prototypes #5603

maint(gpg2john.c): add missing prototypes #5603

claudioandre-br commented Dec 2, 2024

claudioandre-br commented Dec 2, 2024

magnumripper commented Dec 2, 2024

solardiz left a comment

solardiz Dec 2, 2024

solardiz Dec 2, 2024

solardiz Dec 2, 2024

claudioandre-br commented Dec 2, 2024

magnumripper commented Dec 3, 2024

claudioandre-br commented Dec 3, 2024

claudioandre-br commented Dec 3, 2024 •

edited

Loading

claudioandre-br commented Dec 3, 2024

claudioandre-br Dec 3, 2024

claudioandre-br Dec 3, 2024

claudioandre-br commented Dec 3, 2024

solardiz left a comment

solardiz Dec 4, 2024

magnumripper Dec 8, 2024

solardiz commented Dec 4, 2024

claudioandre-br commented Dec 4, 2024

solardiz commented Dec 5, 2024

claudioandre-br commented Dec 6, 2024

maint(gpg2john.c): add missing prototypes #5603

Are you sure you want to change the base?

maint(gpg2john.c): add missing prototypes #5603

Conversation

claudioandre-br commented Dec 2, 2024

claudioandre-br commented Dec 2, 2024

magnumripper commented Dec 2, 2024

solardiz left a comment

Choose a reason for hiding this comment

solardiz Dec 2, 2024

Choose a reason for hiding this comment

solardiz Dec 2, 2024

Choose a reason for hiding this comment

solardiz Dec 2, 2024

Choose a reason for hiding this comment

claudioandre-br commented Dec 2, 2024

magnumripper commented Dec 3, 2024

claudioandre-br commented Dec 3, 2024

claudioandre-br commented Dec 3, 2024 • edited Loading

claudioandre-br commented Dec 3, 2024

claudioandre-br Dec 3, 2024

Choose a reason for hiding this comment

claudioandre-br Dec 3, 2024

Choose a reason for hiding this comment

claudioandre-br commented Dec 3, 2024

solardiz left a comment

Choose a reason for hiding this comment

solardiz Dec 4, 2024

Choose a reason for hiding this comment

magnumripper Dec 8, 2024

Choose a reason for hiding this comment

solardiz commented Dec 4, 2024

claudioandre-br commented Dec 4, 2024

solardiz commented Dec 5, 2024

claudioandre-br commented Dec 6, 2024

claudioandre-br commented Dec 3, 2024 •

edited

Loading