https://bugs.winehq.org/show_bug.cgi?id=52441
Bug ID: 52441 Summary: Soft-Denchi DRM fails to register UCX.ocx Product: Wine Version: 7.0-rc5 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: advapi32 Assignee: wine-bugs@winehq.org Reporter: felineswift+wine@gmail.com Distribution: ---
This software is required by some games (e.g. Armored warrior iris).
To test it install MFC42 and MSVC 2008 SP1 runtime to 32-bit prefix and install Soft-denchi, then run "regsvr32 UCX.ocx" in C:/Program Files/Soft-Denchi.
I tried to understand why it fails. Maybe someone more knowledgeable can find/understand the root cause:
regsvr32 UCX.ocx
it starts loading UCOpgDlg.dll
... call some crypt functions ...
call crypt32.CryptBinaryToStringA with flag CRYPT_STRING_HEX 03e8:fixme:crypt:CryptBinaryToStringA Unimplemented type 4
call advapi32.CryptExportKey
... call some more crypt functions ...
call advapi32.CryptVerifySignatureA -> returns NTE_BAD_SIGNATURE and EXCEPTION_WINE_CXX_EXCEPTION is thrown
...
03e8:warn:module:process_attach Initialization of L"UCOpgDlg.dll" failed
...
regsvr32: loading UCX.ocx failed (it was localized error message)
https://bugs.winehq.org/show_bug.cgi?id=52441
Harold felineswift+wine@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |download URL| |https://acf.paltio.co.jp/sd | |rt/1252/SdRt.exe
https://bugs.winehq.org/show_bug.cgi?id=52441
Fabian Maurer dark.shadow4@web.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dark.shadow4@web.de Ever confirmed|0 |1 Status|UNCONFIRMED |NEW
--- Comment #1 from Fabian Maurer dark.shadow4@web.de --- Added archive: https://web.archive.org/web/20220122200226/https://acf.paltio.co.jp/sdrt/125...
"winetricks vcrun2008" does not work for some reason, use 7z on the sdrt.exe and use the vcredist_x86.exe from there.
Confirming.
https://bugs.winehq.org/show_bug.cgi?id=52441
Fabian Maurer dark.shadow4@web.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Component|advapi32 |crypt32 Summary|Soft-Denchi DRM fails to |Soft-Denchi DRM fails to |register UCX.ocx |register UCX.ocx (needs | |CryptBinaryToStringA with | |CRYPT_STRING_HEX)
--- Comment #2 from Fabian Maurer dark.shadow4@web.de --- I assume it's because of "crypt32.CryptBinaryToStringA with flag CRYPT_STRING_HEX".
https://bugs.winehq.org/show_bug.cgi?id=52441
Fabian Maurer dark.shadow4@web.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|Soft-Denchi DRM fails to |Soft-Denchi DRM fails to |register UCX.ocx (needs |register UCX.ocx (probably |CryptBinaryToStringA with |needs CryptBinaryToStringA |CRYPT_STRING_HEX) |with CRYPT_STRING_HEX)
https://bugs.winehq.org/show_bug.cgi?id=52441
--- Comment #3 from Harold felineswift+wine@gmail.com --- (In reply to Fabian Maurer from comment #1)
"winetricks vcrun2008" does not work for some reason
I assume vcrun2008 doesn't include service pack 1 and there is no verb that provides it either (?).
https://bugs.winehq.org/show_bug.cgi?id=52441
--- Comment #4 from Fabian Maurer dark.shadow4@web.de --- According to winetricks, vcrun2008sp1 is deprecated and just calls vcrun2008.
https://bugs.winehq.org/show_bug.cgi?id=52441
Sagawa sagawa.aki+winebugs@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |sagawa.aki+winebugs@gmail.c | |om
https://bugs.winehq.org/show_bug.cgi?id=52441
Bernhard Übelacker bernhardu@mailbox.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |bernhardu@mailbox.org
--- Comment #5 from Bernhard Übelacker bernhardu@mailbox.org --- The missing CRYPT_STRING_HEX is also tracked in #48487 and maybe #50990.
The patches submitted in #48487 by Aaro Altonen unfortunately got not committed, but can still be applied, and compile and tests succeed, if the strchrW is replaced by wcschr.
https://source.winehq.org/patches/data/182107 https://source.winehq.org/patches/data/182108
https://bugs.winehq.org/show_bug.cgi?id=52441
--- Comment #6 from Gijs Vermeulen gijsvrm@gmail.com --- (In reply to Bernhard Übelacker from comment #5)
The missing CRYPT_STRING_HEX is also tracked in #48487 and maybe #50990.
The patches submitted in #48487 by Aaro Altonen unfortunately got not committed, but can still be applied, and compile and tests succeed, if the strchrW is replaced by wcschr.
https://source.winehq.org/patches/data/182107 https://source.winehq.org/patches/data/182108
Do note, that those patches and bug 48487 are about CryptStringToBinary, while this one is about CryptBinaryToString.
https://bugs.winehq.org/show_bug.cgi?id=52441
--- Comment #7 from Bernhard Übelacker bernhardu@mailbox.org --- (In reply to Gijs Vermeulen from comment #6)
Do note, that those patches and bug 48487 are about CryptStringToBinary, while this one is about CryptBinaryToString.
Sorry about that, I remembered about the other bugs, but did just search with CRYPT_STRING_HEX, therefore mixed them up.
https://bugs.winehq.org/show_bug.cgi?id=52441
--- Comment #8 from Harold felineswift+wine@gmail.com --- The MSVC 2008 SP1 needs specific version which may not be available from microsoft site, but it's inside the SdRt.exe (use archiver).
Also can be found from github zipped with name: vs2008_vcredist_x86-9.0.30729.17.exe
https://bugs.winehq.org/show_bug.cgi?id=52441
sizuku2718@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |sizuku2718@gmail.com
--- Comment #9 from sizuku2718@gmail.com --- Created attachment 75291 --> https://bugs.winehq.org/attachment.cgi?id=75291 CryptBinaryToStringATest.cpp, output_friendly.txt, output.txt
I implemented `CryptBinaryToStringA` for `CRYPT_STRING_HEX` and I tested this patch. In the process, I found a slight bug (please look at `How to implement this`). Now the daemon runs without crashing, but it doesn't works correctly. I start to talking, supposing to use Wine 8.18.
First, I want to apologize for me not directly contributing the code. I have reservations about sharing my real name on the Internet due to privacy concern. I worry that someone who see my code might not be allowed to send the implementation to Wine if I attach the patch. Please understand my hesitation, and I hope you can accept my efforts to explain my observations and suggestions. (If I seem to misunderstand, let me know.)
# The function's behavior
I tested CryptBinaryTostringA for some sequences in my Windows 11 machine with the attached script I wrote, and I checked outputs. (please see the attachments: CryptBinaryToStringATest.cpp, output_friendly.txt, output.txt) And please refer to the [Microsoft documentation](https://learn.microsoft.com/en-us/windows/win32/api/wincrypt/nf-wincrypt-cry...).
I'll explain what I found out. Note that `_` means a space (0x20) and `$` CR (0x0D) below.
First, let `str` be the string generated from the original byte sequence by `CryptBinaryToStringA` with ``CRYPT_STRING_HEXRAW`, and `len` the length of `str`. `len` is always even number because if the hexademical number sequence had odd length, the `0` would added to the last.
Second, define `s[len/2] = {s[0], ... ,s[len/2 - 1]}` as the sequence made by separete `str` into the two charcters block.
## Case 1 (with `CRYPT_STRING_NOCRLF` and optionaly `CRYPT_STRING_NOCR`):
1. join `s` with a space (0x20). 2. return it.
### Example
- `48656c6c6f2c20776f726c64212020` to `48_65_6c_6c_6f_2c_20_77_6f_72_6c_64_21_20_20` - `727432666e3025676e2a` to `72_74_32_66_6e_30_25_67_6e_2a`
## Case 2 (with `CRYPT_STRING_NOCR`):
1. join `s[8*k], s[8*k + 1], ... ,s[8*k + 7]` with a space (0x20), and put it `t[k]`. (remark that the sequence might stop below `8*k + 7` if `k` is the maximum.) 2. join `t[2*k]` and `t[2*k + 1]` with two spaces and put it `u[k]`. (if `2*k` is the maximum for a `k`, you should set `u[k] = t[2*k]`.) 3. join `u` with LF (0x0A) and, put it also on the last. 4. return it.
### Example
- `727432666e3025676e2a` to `72_74_32_66_6e_30_25_67__6e_2a` - `3633394b71775c2d37395c646b0a4f676532723378522b3151397d594c32434b` to the following. (the last character is LF.)
``` 36_33_39_4b_71_77_5c_2d__37_39_5c_64_6b_0a_4f_67 65_32_72_33_78_52_2b_31__51_39_7d_59_4c_32_43_4b ```
- `364c76360a41322d265c34246e6d396d6f0d54313079333171657b6e305c432623403724` to the following (the last character is LF.)
``` 36_4c_76_36_0a_41_32_2d__26_5c_34_24_6e_6d_39_6d 6f_0d_54_31_30_79_33_31__71_65_7b_6e_30_5c_43_26 23_40_37_24 ```
## Case 3 (with no more flags):
It is almost the same as Case 2. Please replace the LF with CR LF.
# Implement Suggstions I guess it's not difficult. Please look at [base64.c](https://gitlab.winehq.org/wine/wine/-/blob/wine-8.18/dlls/crypt32/base64.c?r...)
All you have to do is copy `BinaryToHexRawA()` and change it a little (and also edit the switch block in `CryptBinaryToStringA()`)
## `needed`
`needed` has to be change. If `nbin` == 0, it should be 0, 1, or 2 (depends on the flags). Now I suppose `nbin` != 0.
- if `NOCRLF`, `needed` = 0 + 1 + (3 * nbin - 1). - else if `NOCR`, `needed` = 1 + 1 + ((3 * nbin - 1) + (nbin - 1) / 8). - else, `needed` = 1 + 1 + ((3 * nbin - 1) + (nbin - 1) / 8 + (nbin - 1) / 16).
## main loop
- if `!nbin`, break. - else if `!(count & 15)`, insert " ", "\n", or "\r\n". - else if `!(count & 7)`, insert " " or " ". - else, insert " ".
## The Last Part There's a bug in the implementation of BinaryToHexRawA `BinaryToHexRawA`'s implementation has a bug here. If both `CRYPT_STRING_NOCR` and `CRYPT_STRING_NOCRLF` are set, it shouldn't insert any character at the end, but it currently does. If both are set, the function probably causes overflow because `needed` is smaller.
# Identified Issues
During my testing with `$ wine ./sdrt5080_x64.exe`, I encountered the following error:
Error 1904.Module C:\Program Files (x86)\SoftDenchi\UCX.ocx failed to register. HRESULT -2147220473. Contact your support personnel.
(`sdrt5080_x64.exe` : https://acf.paltio.co.jp/sdrt/sdrt5080_x64.exe)
I also attempted to manually run the daemon with the command:
$ wine Program\ Files\ (x86)/SoftDenchi/UCManSvc.exe
However, an application that relies on SoftDenchi still prompted me to install SoftDenchi.
# In Conclusion
I'm not very experienced with using English, so I had ChatGPT proof-read it. But I guess it is still have some difficult point to understand, so I want to apologize for it.
I hope my findings will be of assistance to the Wine community. Thank you for considering these suggestions.
Best regards,
sizuku2718
https://bugs.winehq.org/show_bug.cgi?id=52441
--- Comment #10 from sizuku2718@gmail.com --- (In reply to sizuku2718 from comment #9)
in the subsection `needed`, here is a correction to be made. The initial statement:
- else, `needed` = 1 + 1 + ((3 * nbin - 1) + (nbin - 1) / 8 + (nbin - 1) / 16).
Needs to be revised as follows:
- else, `needed` = 1 + 2 + ((3 * nbin - 1) + (nbin - 1) / 8 + (nbin - 1) / 16).
That's just my mistake, and I'm sorry. If you found other mistakes, let me know.
Thank you.
sizuku2718
https://bugs.winehq.org/show_bug.cgi?id=52441
Vijay Kamuju infyquest@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |infyquest@gmail.com Fixed by SHA1| |d3a4477bad90f37dab584af89e1 | |baa07a989437e Resolution|--- |FIXED
--- Comment #11 from Vijay Kamuju infyquest@gmail.com --- Please use latest wine version to test, as this functionality has been implemented by commit - d3a4477bad90f37dab584af89e1baa07a989437e
https://source.winehq.org/git/wine.git/commitdiff/d3a4477bad90f37dab584af89e...
https://bugs.winehq.org/show_bug.cgi?id=52441
Vijay Kamuju infyquest@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |UNCONFIRMED Fixed by SHA1|d3a4477bad90f37dab584af89e1 | |baa07a989437e | Resolution|FIXED |--- Ever confirmed|1 |0
--- Comment #12 from Vijay Kamuju infyquest@gmail.com --- oops its for CryptBinaryToStringW
-
WineHQ Bugzilla