Created attachment 45091 [details]
stdout

Created attachment 45092 [details]
stderr

Created attachment 45093 [details]
plain blizz-provided log

Created attachment 45185 [details]
Adds simple support for HCCE_LOCAL_MACHINE in the crypt32 dll CertGetCertificateChain function.

The Battle.net desktop app makes a call to CertGetCertificateChain passing in HCCE_LOCAL_MACHINE for the hChainEngine parameter. Lack of support for this causes the crash. This patch hacks in simple support for HCCE_LOCAL_MACHINE.

Comment on attachment 45185 [details]
Adds simple support for HCCE_LOCAL_MACHINE in the crypt32 dll CertGetCertificateChain function.

>From 276ca28ff98ed69920e3ad10c012a1bd3b022dd9 Mon Sep 17 00:00:00 2001
>From: Ben Deane <wine@elbeno.com>
>Date: Wed, 10 Jul 2013 16:47:27 -0700
>Subject: crypt32: Support HCCE_LOCAL_MACHINE.
>
>Add simple support for HCCE_LOCAL_MACHINE for
>CertGetCertificateChainEngine et al.
>---
> dlls/crypt32/chain.c           | 28 ++++++++++++++++------------
> dlls/crypt32/crypt32_private.h |  2 +-
> dlls/crypt32/main.c            |  5 +++--
> 3 files changed, 20 insertions(+), 15 deletions(-)
>
>diff --git a/dlls/crypt32/chain.c b/dlls/crypt32/chain.c
>index d112673..d142b9b 100644
>--- a/dlls/crypt32/chain.c
>+++ b/dlls/crypt32/chain.c
>@@ -33,7 +33,10 @@ WINE_DECLARE_DEBUG_CHANNEL(chain);
>
> #define DEFAULT_CYCLE_MODULUS 7
>
>-static HCERTCHAINENGINE CRYPT_defaultChainEngine;
>+/* There are two default chain engines which correspond to HCCE_CURRENT_USER and
>+ * HCCE_LOCAL_MACHINE.
>+*/
>+static HCERTCHAINENGINE CRYPT_defaultChainEngine[2] = { NULL, NULL };
>
> /* This represents a subset of a certificate chain engine:  it doesn't include
>  * the "hOther" store described by MSDN, because I'm not sure how that's used.
>@@ -212,7 +215,7 @@ VOID WINAPI CertFreeCertificateChainEngine(HCERTCHAINENGINE hChainEngine)
>
>     TRACE("(%p)\n", hChainEngine);
>
>-    if (engine && InterlockedDecrement(&engine->ref) == 0)
>+    if (engine > HCCE_LOCAL_MACHINE && InterlockedDecrement(&engine->ref) == 0)
>     {
>         CertCloseStore(engine->hWorld, 0);
>         CertCloseStore(engine->hRoot, 0);
>@@ -220,26 +223,28 @@ VOID WINAPI CertFreeCertificateChainEngine(HCERTCHAINENGINE hChainEngine)
>     }
> }
>
>-static HCERTCHAINENGINE CRYPT_GetDefaultChainEngine(void)
>+static HCERTCHAINENGINE CRYPT_GetDefaultChainEngine(HCERTCHAINENGINE h)
> {
>-    if (!CRYPT_defaultChainEngine)
>+    if (!CRYPT_defaultChainEngine[(int)h])
>     {
>         CERT_CHAIN_ENGINE_CONFIG config = { 0 };
>         HCERTCHAINENGINE engine;
>
>         config.cbSize = sizeof(config);
>+        if (h == HCCE_LOCAL_MACHINE)
>+            config.dwFlags = CERT_CHAIN_USE_LOCAL_MACHINE_STORE;
>         CertCreateCertificateChainEngine(&config, &engine);
>-        InterlockedCompareExchangePointer(&CRYPT_defaultChainEngine, engine,
>+        InterlockedCompareExchangePointer(&CRYPT_defaultChainEngine[(int)h], engine,
>          NULL);
>-        if (CRYPT_defaultChainEngine != engine)
>+        if (CRYPT_defaultChainEngine[(int)h] != engine)
>             CertFreeCertificateChainEngine(engine);
>     }
>-    return CRYPT_defaultChainEngine;
>+    return CRYPT_defaultChainEngine[(int)h];
> }
>
>-void default_chain_engine_free(void)
>+void default_chain_engine_free(HCERTCHAINENGINE h)
> {
>-    CertFreeCertificateChainEngine(CRYPT_defaultChainEngine);
>+    CertFreeCertificateChainEngine(CRYPT_defaultChainEngine[(int)h]);
> }
>
> typedef struct _CertificateChain
>@@ -2819,11 +2824,10 @@ BOOL WINAPI CertGetCertificateChain(HCERTCHAINENGINE hChainEngine,
>         return FALSE;
>     }
>
>-    if (!hChainEngine)
>-        hChainEngine = CRYPT_GetDefaultChainEngine();
>+    if (hChainEngine <= HCCE_LOCAL_MACHINE)
>+        hChainEngine = CRYPT_GetDefaultChainEngine(hChainEngine);
>     if (TRACE_ON(chain))
>         dump_chain_para(pChainPara);
>-    /* FIXME: what about HCCE_LOCAL_MACHINE? */
>     ret = CRYPT_BuildCandidateChainFromCert(hChainEngine, pCertContext, pTime,
>      hAdditionalStore, &chain);
>     if (ret)
>diff --git a/dlls/crypt32/crypt32_private.h b/dlls/crypt32/crypt32_private.h
>index ea85cbc..a07a830 100644
>--- a/dlls/crypt32/crypt32_private.h
>+++ b/dlls/crypt32/crypt32_private.h
>@@ -155,7 +155,7 @@ void crypt_oid_init(void) DECLSPEC_HIDDEN;
> void crypt_oid_free(void) DECLSPEC_HIDDEN;
> void crypt_sip_free(void) DECLSPEC_HIDDEN;
> void root_store_free(void) DECLSPEC_HIDDEN;
>-void default_chain_engine_free(void) DECLSPEC_HIDDEN;
>+void default_chain_engine_free(HCERTCHAINENGINE) DECLSPEC_HIDDEN;
>
> /* Some typedefs that make it easier to abstract which type of context we're
>  * working with.
>diff --git a/dlls/crypt32/main.c b/dlls/crypt32/main.c
>index 78f14f9..227e086 100644
>--- a/dlls/crypt32/main.c
>+++ b/dlls/crypt32/main.c
>@@ -49,8 +49,8 @@ BOOL WINAPI DllMain(HINSTANCE hInst, DWORD fdwReason, PVOID pvReserved)
>             crypt_oid_free();
>             crypt_sip_free();
>             root_store_free();
>-            default_chain_engine_free();
>+            default_chain_engine_free(HCCE_CURRENT_USER);
>+            default_chain_engine_free(HCCE_LOCAL_MACHINE);
>             if (hDefProv) CryptReleaseContext(hDefProv, 0);
>             break;
>     }
>1.8.1.2

Created attachment 45189 [details]
Support HCCE_LOCAL_MACHINE in CertGetCertificateChain

Sorry, the first attachment had garbage characters which prevented the diff from applying. The patch expanded in the comment is correct.

Thanks so much. Confirming patch (attachment 45189 [details]) applies against git and fixes the instant crash.

Component changed to crypt32, patch keyword added.

I'm now getting BLZBNTBNA00000640 (SSL failed to load), but that'll be a separate bug.

Thank you for the patch, I can confirm that it works on debian sid wine-unstable,
and also on the git version and solve the nearly "instant crash" from the launcher desktop app.

But the launcher, give me a SSL error : BLZBNTBNA00000640 (1202),
I am not it is related so maybe we need a new BR.

I tested with with the 1.7 git version and got the same bug.

Some more test :

D3 can be launched, there is a strange *new* bug on my mouse witch seem's in a box, but the game is working.

WOW can not be launched, and stay freeze on "Connecting", maybe it is related to the SSL bug.

(In reply to comment #6)
Ben, have you sent the patch to wine-patches@winehq.org? Patches are not picked up from bugzilla.

http://wiki.winehq.org/SubmittingPatches

(Reply to comment 7, 8, 9)
The SSL bug is http://bugs.winehq.org/show_bug.cgi?id=27168

(In reply to comment #11)
> (Reply to comment 7, 8, 9)
> The SSL bug is http://bugs.winehq.org/show_bug.cgi?id=27168

Yes and there a patch provided there that fix the https/ssl bug. (http://sprunge.us/LQWS)

Anyone can update this patch for the latest source?

It seems the SSL patch is already integrated so that should not be an issue but this immediate crash and the attached patch do not work.

Thank you

Created attachment 46553 [details]
Updated HCCE patch against current git.

OK, this one... It applies against current version fetched from git. Compiles, builds, WoW runs from launcher as before.

"Battle.Net Launcher.exe" waits forever at 'looking for updates'
BUT!
"wine Battle.Net.exe" runs correctly and updates your games as expected. So... it's progress. Someone with more coding chops than I have should check this out.

Created attachment 46554 [details]
Actually get the HCCE patch right this time

OK this time I actully have the right patch. First time using the git-format-patch thing. This one is right.

Thank you. This was really helpful. 

This patch worked perfectly. I know it might not be at the level of submission to the main wine source branch but it would be nice if someone follows that step and validates this for inclusion.

Thank you

This patch resolved the issue for me.

Still an issue in wine 1.7.8

FYI:
The battle.net desktop client can also be tested in the Diablo 3 Public Test Realm (PTR) now. This is no closed beta, but you need a Diablo 3 license.
http://eu.battle.net/d3/en/blog/11954349/patch-201-ptr-now-available-12-12-2013

I couldn't find the patch submitted to to the wine-patches ML, but I found a review of the patch here:
http://eu.battle.net/d3/en/blog/11954349/patch-201-ptr-now-available-12-12-2013

It would be nice if somebody could check this review,
write an updated patch and submit it to the mailing list.
There is a working patch for half a year now, but still nothing submitted.

I pasted the same link twice. Sorry for that.

The link with the review is:
http://www.winehq.org/pipermail/wine-devel/2013-September/101473.html

And now I also understood that patches and review are on different MLs,
so I also found the patch that was reviewed:
http://www.winehq.org/pipermail/wine-patches/2013-September/127042.html

So @Ben Deane, please have a look at that review if you missed it.

So I read the review and as far as I can tell, it's sound. The changes it suggests are reasonable. I just am not a good enough programmer nor familiar enough with the project code to implement it myself. All I did was hack the hack to compensate for some renamed variables that were making one of the files fail to patch correctly.

Someone with some better chops than I have will have to do this if we want it actually included.

On wine-1.7.8-184-g965d0ca in 64 bit environment (untested in 32) the patch seems to prevent the bnet app from launching at all  (at least the client; the agent seems to start). 

err:ntdll:RtlpWaitForCriticalSection section 0x7e8182e0 "?" wait timed out in thread 0064, blocked by 0000, retrying (60 sec)
err:ntdll:RtlpWaitForCriticalSection section 0x7e8182e0 "?" wait timed out in thread 0064, blocked by 0000, retrying (60 sec)

wine-1.7.8-88-gfb75292 did work for me with https://bugs.winehq.org/attachment.cgi?id=46554&action=diff as only patch applied. (64 bit Arch Linux)
I tested this build again to make sure nothing changed in the Battle.net Beta and it still works.

Current git (wine-1.7.8-142-gb87b9f0 plus patch) seems to have the same issues for me as for you (no windows opens, these RtlWaitForCriticalSection lines at the end).
Battle.net.exe keeps running until killed (though no Window opens and the Agent.exe is closed after inactivity automatically).
This is *not* the bug that the patch was fixing, but a different bug.

I will try current git without patch, too.
It does look a lot like you found a new bug introduced in current git that doesn't have to do anything with this patch. (-> new bug report, regression)

I tested newest git again without the patch from this issue and get the same problems you are stating (comment 22).

So please open a new bug and link it here.
You can start regression with fb75292 (good), b87b9f0 (bad) (56 commits).

I have no clue where you take wine-1.7.8-184-g965d0ca from, by the way. I couldn't find this commit in master and it has 42 commits more on top of 1.7.8 as the current master.

I opened a report for the regression:
https://bugs.winehq.org/show_bug.cgi?id=35189

I suspect it's their internal commit for the patches.

* Plain 1.7.8+hcce patch runs fine for those folks just wanting something that works.

Here is a hotfix for ppl who can't compile wine.
http://appdb.winehq.org/objectManager.php?sClass=version&iId=28875
(See Installation Fix)

I can confirm that adding "dbghelp (disable)" to the dll overrides works as workaround (without patching wine).
It still doesn't show the news (which the patch does), but it doesn't crash.
The disabled dbghelp workaround shows this where the news should be:

"Whoops! We're having problems loading some content.
Please try reloading this page using the button below.
Thanks!
(Reload) More help: BLZBNTBNA000003E8 (1007)"
(the last thing is probably a link to
https://nydus.battle.net/App/enGB/client/error/BLZBNTBNA000003E8?targetRegion=EU&code=1007, (I wasn't prepared to open another FF window there))

This shows up in the terminal (with the dbghelp disabled):
[Browser] {} https://nydus.battle.net/App/enGB/client/play?targetRegion=EU (68):The page at https://nydus.battle.net/App/enGB/client/play?targetRegion=EU displayed insecure content from resources://client/errors/crabby/hats.png.

Confirming the bug with 1.7.10. And I can report that the patches from attachment 46554 [details] fix this bug (no crash and news load). (Debian users can just test this with my packages <http://dev.carbon-project.org/debian/wine-unstable/>.)

@Ryan Davis (comment 15): have you submitted the patch series to wine-patches? If not, please do so (see <http://wiki.winehq.org/SubmittingPatches> for details on how to submit patches). If your patches were derived from Ben's you should probably give some credit in your commit messages before sending them to the mailing list.

Thanks for your work!

Well... the commit message was my own internal one, so perhaps wasn't all clear. All credit goes to Ben. All I did was massage it a little so it applied after some variable names got changed. I was actually surprised that it compiled afterward... so it's Ben's patch and all credit should go where due.

If you look back in the issue log here, you'll see that the patch was submitted earlier and found unsuitable. There was a bunch of good criticism, and I changed nothing to make it any different. I'm not a good enough coder to bring the patch up to spec, so re-submitting it to be incorporated upstream seemed useless. Again, someone with better chops than I have should look over the criticism and incorporate it into something that a) meets project standards & best practices b) doesn't throw warnings during compilation.

With this patch http://bugs.winehq.org/attachment.cgi?id=46554&action=diff Battle net app work's perfectly, but StarCraft II closed after login.

(In reply to comment #31)
> With this patch http://bugs.winehq.org/attachment.cgi?id=46554&action=diff
> Battle net app work's perfectly, but StarCraft II closed after login.

Why do you think the StarCraft II problem is related?

Did you try the dbhelp (disabled) workaround? Do you still have the StartCraft II problem? In that case the problem probably isn't related to the patch.


What does "Battle net app work's perfectly" mean?
Does this also fix bug 33943 for you?

(In reply to comment #32)
> (In reply to comment #31)
> > With this patch http://bugs.winehq.org/attachment.cgi?id=46554&action=diff
> > Battle net app work's perfectly, but StarCraft II closed after login.
> 
> Why do you think the StarCraft II problem is related?
> 
> Did you try the dbhelp (disabled) workaround? Do you still have the
> StartCraft II problem? In that case the problem probably isn't related to
> the patch.
> 
> 
> What does "Battle net app work's perfectly" mean?
> Does this also fix bug 33943 for you?

Battle.net application, Diablo III, Hearthstone works, when i 
launch StarCraft II, press login button, StarCraft II closing. Without patch(i'm testing in PlayOnLinux) StarCraft II logining and not closing.

(In reply to comment #33)
> (In reply to comment #32)
> > (In reply to comment #31)
> > > With this patch http://bugs.winehq.org/attachment.cgi?id=46554&action=diff
> > > Battle net app work's perfectly, but StarCraft II closed after login.
> > 
> > Why do you think the StarCraft II problem is related?
> > 
> > Did you try the dbhelp (disabled) workaround? Do you still have the
> > StartCraft II problem? In that case the problem probably isn't related to
> > the patch.
> > 
> > 
> > What does "Battle net app work's perfectly" mean?
> > Does this also fix bug 33943 for you?
> 
> Battle.net application, Diablo III, Hearthstone works, when i 
> launch StarCraft II, press login button, StarCraft II closing. Without
> patch(i'm testing in PlayOnLinux) StarCraft II logining and not closing.

Well, in that case I just can't confirm this. I am playing (and can login) fine with the Patch and wine 1.7.11+59, starcraft II starter edition. Started using the old launcher (I don't use the BN-App yet to start Starcraft.

Did you start Starcraft the same way, when trying with and without patch? (both without BN-App, or both with (meaning you'd have to use dbghelp (disable) when not using the patch)).

Additionally:
There have been several login problems these days, at Blizzard end.

(In reply to comment #32)
> (In reply to comment #31)
> > With this patch http://bugs.winehq.org/attachment.cgi?id=46554&action=diff
> > Battle net app work's perfectly, but StarCraft II closed after login.
> 
> Why do you think the StarCraft II problem is related?
> 
> Did you try the dbhelp (disabled) workaround? Do you still have the
> StartCraft II problem? In that case the problem probably isn't related to
> the patch.
> 
> 
> What does "Battle net app work's perfectly" mean?
> Does this also fix bug 33943 for you?

Battle.net application, Diablo III, Hearthstone works, when i 
launch StarCraft II, press login button, StarCraft II closing. Without patch(i'm testing in PlayOnLinux) StarCraft II logining and not closing.(In reply to comment #34)
> (In reply to comment #33)
> > (In reply to comment #32)
> > > (In reply to comment #31)
> > > > With this patch http://bugs.winehq.org/attachment.cgi?id=46554&action=diff
> > > > Battle net app work's perfectly, but StarCraft II closed after login.
> > > 
> > > Why do you think the StarCraft II problem is related?
> > > 
> > > Did you try the dbhelp (disabled) workaround? Do you still have the
> > > StartCraft II problem? In that case the problem probably isn't related to
> > > the patch.
> > > 
> > > 
> > > What does "Battle net app work's perfectly" mean?
> > > Does this also fix bug 33943 for you?
> > 
> > Battle.net application, Diablo III, Hearthstone works, when i 
> > launch StarCraft II, press login button, StarCraft II closing. Without
> > patch(i'm testing in PlayOnLinux) StarCraft II logining and not closing.
> 
> Well, in that case I just can't confirm this. I am playing (and can login)
> fine with the Patch and wine 1.7.11+59, starcraft II starter edition.
> Started using the old launcher (I don't use the BN-App yet to start
> Starcraft.
> 
> Did you start Starcraft the same way, when trying with and without patch?
> (both without BN-App, or both with (meaning you'd have to use dbghelp
> (disable) when not using the patch)).
> 
> Additionally:
> There have been several login problems these days, at Blizzard end.

I'm record video.
http://rghost.ru/private/52099871/0cc56110959eddd75bc9ec9d01ef9c00
First i'm launch starcraft with patch, after i'm launch with playonlinux.

(In reply to comment #33)
> (In reply to comment #32)
> > (In reply to comment #31)
> > > With this patch http://bugs.winehq.org/attachment.cgi?id=46554&action=diff
> > > Battle net app work's perfectly, but StarCraft II closed after login.
> > 
> > Why do you think the StarCraft II problem is related?
> > 
> > Did you try the dbhelp (disabled) workaround? Do you still have the
> > StartCraft II problem? In that case the problem probably isn't related to
> > the patch.
> > 
> > 
> > What does "Battle net app work's perfectly" mean?
> > Does this also fix bug 33943 for you?
> 
> Battle.net application, Diablo III, Hearthstone works, when i 
> launch StarCraft II, press login button, StarCraft II closing. Without
> patch(i'm testing in PlayOnLinux) StarCraft II logining and not closing.

PlayOnLinux (or any wine 3rd-party) is *NOT SUPPORTED* here. You must only try with plain Wine.

(In reply to comment #36)
> (In reply to comment #33)
> > (In reply to comment #32)
> > > (In reply to comment #31)
> > > > With this patch http://bugs.winehq.org/attachment.cgi?id=46554&action=diff
> > > > Battle net app work's perfectly, but StarCraft II closed after login.
> > > 
> > > Why do you think the StarCraft II problem is related?
> > > 
> > > Did you try the dbhelp (disabled) workaround? Do you still have the
> > > StartCraft II problem? In that case the problem probably isn't related to
> > > the patch.
> > > 
> > > 
> > > What does "Battle net app work's perfectly" mean?
> > > Does this also fix bug 33943 for you?
> > 
> > Battle.net application, Diablo III, Hearthstone works, when i 
> > launch StarCraft II, press login button, StarCraft II closing. Without
> > patch(i'm testing in PlayOnLinux) StarCraft II logining and not closing.
> 
> PlayOnLinux (or any wine 3rd-party) is *NOT SUPPORTED* here. You must only
> try with plain Wine.

Wine 1.7.11 without patch launch Starcraft II, and all ok. Wine 1.7.11 with this patch (http://bugs.winehq.org/attachment.cgi?id=46554&action=diff) can't login in Starcraft, he is closing...

(In reply to comment #37)
> (In reply to comment #36)
> > (In reply to comment #33)
> > > (In reply to comment #32)
> > > > (In reply to comment #31)
> > > > > With this patch http://bugs.winehq.org/attachment.cgi?id=46554&action=diff
> > > > > Battle net app work's perfectly, but StarCraft II closed after login.
> > > > 
> > > > Why do you think the StarCraft II problem is related?
> > > > 
> > > > Did you try the dbhelp (disabled) workaround? Do you still have the
> > > > StartCraft II problem? In that case the problem probably isn't related to
> > > > the patch.
> > > > 
> > > > 
> > > > What does "Battle net app work's perfectly" mean?
> > > > Does this also fix bug 33943 for you?
> > > 
> > > Battle.net application, Diablo III, Hearthstone works, when i 
> > > launch StarCraft II, press login button, StarCraft II closing. Without
> > > patch(i'm testing in PlayOnLinux) StarCraft II logining and not closing.
> > 
> > PlayOnLinux (or any wine 3rd-party) is *NOT SUPPORTED* here. You must only
> > try with plain Wine.
> 
> Wine 1.7.11 without patch launch Starcraft II, and all ok. Wine 1.7.11 with
> this patch (http://bugs.winehq.org/attachment.cgi?id=46554&action=diff)
> can't login in Starcraft, he is closing...

I'm try install wine from <http://dev.carbon-project.org/debian/wine-unstable/>, wineprefix cleared and i have same trouble.
sadly, i must rebuild wine without this patch.

> I'm try install wine from
> <http://dev.carbon-project.org/debian/wine-unstable/>, wineprefix cleared
> and i have same trouble.
> sadly, i must rebuild wine without this patch.
I don't know why, problem be in video driver....

I can confirm the patch fixed the crash for me as well.
I patched against git.

This seems to be fixed with today's git unpatched (probably ca2e1c164fcd892dfbeaa865914f97ce3ae3c1a5).

Confirming fixed. Thanks, Jacek.

*** Bug 35724 has been marked as a duplicate of this bug. ***

Closing bugs fixed in 1.7.14.