- Wine-Devel - list.winehq.org

[PATCH v2 06/19] ntdll: Store stack pointer in x86_64 syscall frame.
by Jacek Caban Feb. 15, 2021

Feb. 15, 2021

Signed-off-by: Jacek Caban <jacek(a)codeweavers.com> --- dlls/ntdll/unix/signal_x86_64.c | 13 +++++-------- tools/winebuild/import.c | 2 ++ 2 files changed, 7 insertions(+), 8 deletions(-)

1 0

[PATCH v2 05/19] ntdll: Store instruction pointer register in x86_64 syscall frame.
by Jacek Caban Feb. 15, 2021

Feb. 15, 2021

Signed-off-by: Jacek Caban <jacek(a)codeweavers.com> --- dlls/ntdll/unix/signal_x86_64.c | 4 ++-- tools/winebuild/import.c | 4 +++- 2 files changed, 5 insertions(+), 3 deletions(-)

1 0

[PATCH v2 04/19] ntdll: Store eflags in x86_64 syscall frame.
by Jacek Caban Feb. 15, 2021

Feb. 15, 2021

Signed-off-by: Jacek Caban <jacek(a)codeweavers.com> --- dlls/ntdll/unix/signal_x86_64.c | 2 +- tools/winebuild/import.c | 4 +++- 2 files changed, 4 insertions(+), 2 deletions(-)

1 0

[PATCH v2 03/19] ntdll: Reserve space for all registers in x86_64 syscall frame.
by Jacek Caban Feb. 15, 2021

Feb. 15, 2021

Signed-off-by: Jacek Caban <jacek(a)codeweavers.com> --- dlls/ntdll/unix/signal_x86_64.c | 47 ++++++++++++++++++++++----------- tools/winebuild/import.c | 32 +++++++++++----------- 2 files changed, 48 insertions(+), 31 deletions(-)

1 0

[PATCH v2 02/19] winebuild: Remove no longer needed stack alignment in x86_64 syscall dispatcher.
by Jacek Caban Feb. 15, 2021

Feb. 15, 2021

Signed-off-by: Jacek Caban <jacek(a)codeweavers.com> --- tools/winebuild/import.c | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-)

1 0

[PATCH v2 01/19] ntdll: Store entire XMM context in x86_64 syscall frame.
by Jacek Caban Feb. 15, 2021

Feb. 15, 2021

Signed-off-by: Jacek Caban <jacek(a)codeweavers.com> --- v2 handles NtContinue better and picks fxsave/xsave/xsavec variant of syscall dispatcher in runtime, depending on host capabilities. dlls/ntdll/unix/signal_x86_64.c | 58 ++++++++++++++------------------- tools/winebuild/import.c | 18 +++------- 2 files changed, 30 insertions(+), 46 deletions(-)

1 0

[PATCH] plugplay: Broadcast WM_DEVICECHANGE message asynchronously.
by Rémi Bernon Feb. 15, 2021

Feb. 15, 2021

It may otherwise trigger a nasty race condition, where: 1) For explorer.exe to register the CLSID_ShellWindows classes, it needs RpcSS service to be started, 2) services.exe may start WinePlugPlay service group first, waiting for its startup to complete, 3) during startup and early device enumeration, hidclass.sys may call IoSetDeviceInterfaceState, which calls plugplay_send_event [1], 4) plugplay_send_event tries to broadcast a WM_DEVICECHANGE message with BSF_QUERY, waiting for the individual threads to reply, 5) which times-out because window threads are waiting on explorer.exe to create its desktop window and reply to the WM_NULL SendMessage. This happens more likely as there is threads with message queues being started, each waiting on the desktop window to reply. Usually explorer.exe is able to reply to some messages with the implicit message processing while creating its window, but not all of them. [1] Not completely sure how, it looks like some RPC too, but before RpcSs is started? Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- Not sure if this is right, but sending it at least for the description of the race condition. programs/plugplay/main.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/programs/plugplay/main.c b/programs/plugplay/main.c index 19e02286312..1156cb9f036 100644 --- a/programs/plugplay/main.c +++ b/programs/plugplay/main.c @@ -142,8 +142,8 @@ void __cdecl plugplay_send_event( DWORD code, const BYTE *data, unsigned int siz struct listener *listener; struct event *event; - BroadcastSystemMessageW( BSF_FORCEIFHUNG | BSF_QUERY, NULL, WM_DEVICECHANGE, code, (LPARAM)data ); - BroadcastSystemMessageW( BSF_FORCEIFHUNG | BSF_QUERY, NULL, WM_DEVICECHANGE, DBT_DEVNODES_CHANGED, 0 ); + BroadcastSystemMessageW( 0, NULL, WM_DEVICECHANGE, code, (LPARAM)data ); + BroadcastSystemMessageW( 0, NULL, WM_DEVICECHANGE, DBT_DEVNODES_CHANGED, 0 ); EnterCriticalSection( &plugplay_cs ); -- 2.30.0

2 1

RFC: Split the test.w.org Windows 10 page
by Francois Gouget Feb. 15, 2021

Feb. 15, 2021

So the Windows 10 page is getting a ton of results, in part because now there are 11 different Windows 10 versions (of which we test 7). So the idea would be to split it into multiple pages. For reference here are the number of test results we typically get per Windows 10 version: 1507 7 1511 1607 1 1703 1709 4 1803 1809 6 1903 1909 5 2004 11 (w10pro64 locales, always current spring) 2009 9 (mostly w1064, always current fall) Note: my goal going forward is to always keep the Windows 10 VM running the locale tests (w10pro64) on the latest spring release; while the VM (w1064) testing the different Windows 10 versions and other 'esoteric' configurations (dual screen, test signing) would only get the fall releases. The patch below creates four-release groups, so one new group every other year, except for the latest releases. So Win10.15, Win10.17, Win10.19 and Win10.20. It does mean that Win10.20 gets 20+ results. (maybe better names would be Win1507+, Win1709+, Win1909+ and Win2004+) Another approach would be to have one group per fall release and put the next fall release with it. Though we would not have much in the Win1607 group. Whatever groups we decide on, one question is whether to keep the results from the latest spring and fall releases together as that means 20+ results. To me it makes sense to have the all the results for the 'latest' version together. But it may also be good to have the locale results on one page and all the 'normal' ones on another. Then we'd presumably just create separate groups for the latest spring and fall releases; which would be simpler than having dissect check the locale. I have also not tested what happens when a test result gets migrated from one group to another (the patch does not account for the existing results in the 'win10' group, though that should be easy). At least dissect prefixes the results with the Windows version so we can then decide to group the results any way we want without having to re-run dissect. In the worst case all will be good once the old results expire. But we may be able to move existing results from one page to another by touching or deleting some summary files. Also let me know if you want more spring releases. --- winetest/build-index | 12 ++++++--- winetest/dissect | 61 +++++++++++++++++++++++++++----------------- winetest/gather | 22 +++++++++++++--- 3 files changed, 64 insertions(+), 31 deletions(-) diff --git a/winetest/build-index b/winetest/build-index index 58f28c0c1..203871c71 100755 --- a/winetest/build-index +++ b/winetest/build-index @@ -96,7 +96,10 @@ my %vista = (name => "Vista"); my %w2k8 = (name => "2008"); my %win7 = (name => "Win7"); my %win8 = (name => "Win8"); -my %win10 = (name => "Win10"); +my %win10_15= (name => "Win10.15"); +my %win10_17= (name => "Win10.17"); +my %win10_19= (name => "Win10.19"); +my %win10_20= (name => "Win10.20"); my %unknown = (name => "Other"); my %linux = (name => "Linux"); my %mac = (name => "Mac"); @@ -105,7 +108,10 @@ my %solaris = (name => "Solaris"); my %wine = (name => "Wine"); # Define the order of version groups in the summary -my @groups = (\%w95, \%w98, \%me, \%nt3, \%nt4, \%w2k, \%xp, \%w2k3, \%vista, \%w2k8, \%win7, \%win8, \%win10, +my @groups = (\%w95, \%w98, \%me, + \%nt3, \%nt4, \%w2k, \%xp, \%w2k3, + \%vista, \%w2k8, \%win7, \%win8, + \%win10_15, \%win10_17, \%win10_19, \%win10_20, \%unknown, \%linux, \%mac, \%bsd, \%solaris, \%wine); @@ -356,7 +362,7 @@ foreach my $build (@builds) { while (<TOTAL>) { - if (/^([A-Za-z0-9]+)\s+(\d+)\s+(\d+)\s+(\d+)\s+(\d+)(?:\s+(\d+))?/) + if (/^([A-Za-z0-9.]+)\s+(\d+)\s+(\d+)\s+(\d+)\s+(\d+)(?:\s+(\d+))?/) { my ($name, $runs, $tests, $errors, $todos, $successes) = ($1, $2, $3, $4, $5, $6); $versions{$name}++; diff --git a/winetest/dissect b/winetest/dissect index dd7241631..9a2f24e6e 100755 --- a/winetest/dissect +++ b/winetest/dissect @@ -305,7 +305,7 @@ $line =~ /^Operating system version:\r?$/ or mydie "no OS header: $line"; $box->{data} .= "<h2>Operating system version</h2>\n"; $box->{data} .= "<table class=\"output\">\n"; -my ($wine, $wine_build, $major, $minor, $plid, $product, $host); +my ($wine, $wine_build, $major, $minor, $winbuild, $plid, $product, $host); while ($line = <IN> || "") { last if ($line !~ /^\s*([0-9a-zA-Z ]+)=(.*?)\r?$/); @@ -320,6 +320,8 @@ while ($line = <IN> || "") $major = $2; } elsif ($1 eq "dwMinorVersion") { $minor = $2; + } elsif ($1 eq "dwBuildNumber") { + $winbuild = $2; } elsif ($1 eq "PlatformId") { $plid = $2; } elsif ($1 eq "wProductType") { @@ -341,25 +343,35 @@ if (!defined $plid or !defined $major or !defined $minor or !defined $product) { my @idmatch = ( # Describes how to match a platform's version information # with a dissect platform id: - # dissect id plid major minor product prediluvian - [ "95", 1, 4, 0, undef, 1 ], - [ "98", 1, 4, 10, undef, 1 ], - [ "me", 1, 4, 90, undef, 1 ], - [ "nt3", 2, 3, 51, undef, 1 ], - [ "2000", 2, 5, 0, undef, 1 ], - [ "xp", 2, 5, 1, 1, 0 ], - [ "xp", 2, 5, 2, 1, 0 ], - [ "2003", 2, 5, 2, undef, 0 ], - [ "vista", 2, 6, 0, 1, 0 ], - [ "2008", 2, 6, 0, 3, 0 ], - [ "win7", 2, 6, 1, 1, 0 ], - [ "2008", 2, 6, 1, 3, 0 ], - [ "win8", 2, 6, 2, undef, 0 ], - [ "win8", 2, 6, 3, undef, 0 ], - [ "win10", 2, 10, 0, 1, 0 ], -# [ "ce", 3, undef, undef, undef, 0 ], - [ "unknown", 2, undef, undef, undef, 0 ], - [ "unknown", undef, undef, undef, undef, 1 ], + # dissect id plid major minor build product prediluvian + [ "95", 1, 4, 0, undef, undef, 1 ], + [ "98", 1, 4, 10, undef, undef, 1 ], + [ "me", 1, 4, 90, undef, undef, 1 ], + [ "nt3", 2, 3, 51, undef, undef, 1 ], + [ "2000", 2, 5, 0, undef, undef, 1 ], + [ "xp", 2, 5, 1, undef, 1, 0 ], + [ "xp", 2, 5, 2, undef, 1, 0 ], + [ "2003", 2, 5, 2, undef, undef, 0 ], + [ "vista", 2, 6, 0, undef, 1, 0 ], + [ "2008", 2, 6, 0, undef, 3, 0 ], + [ "win7", 2, 6, 1, undef, 1, 0 ], + [ "2008", 2, 6, 1, undef, 3, 0 ], + [ "win8", 2, 6, 2, undef, undef, 0 ], + [ "win8", 2, 6, 3, undef, undef, 0 ], + [ "win1507", 2, 10, 0, 10240, 1, 0 ], + [ "win1511", 2, 10, 0, 10586, 1, 0 ], + [ "win1607", 2, 10, 0, 14393, 1, 0 ], + [ "win1703", 2, 10, 0, 15063, 1, 0 ], + [ "win1709", 2, 10, 0, 16299, 1, 0 ], + [ "win1803", 2, 10, 0, 17134, 1, 0 ], + [ "win1809", 2, 10, 0, 17763, 1, 0 ], + [ "win1903", 2, 10, 0, 18362, 1, 0 ], + [ "win1909", 2, 10, 0, 18363, 1, 0 ], + [ "win2004", 2, 10, 0, 19041, 1, 0 ], + [ "win2009", 2, 10, 0, 19042, 1, 0 ], +# [ "ce", 3, undef, undef, undef, undef, 0 ], + [ "unknown", 2, undef, undef, undef, undef, 0 ], + [ "unknown", undef, undef, undef, undef, undef, 1 ], ); my ($version, $prediluvian); @@ -368,15 +380,16 @@ foreach my $entry (@idmatch) if ((!defined $entry->[1] or $entry->[1] eq $plid) and (!defined $entry->[2] or $entry->[2] eq $major) and (!defined $entry->[3] or $entry->[3] eq $minor) and - (!defined $entry->[4] or $entry->[4] eq $product)) + (!defined $entry->[4] or $entry->[4] eq $winbuild) and + (!defined $entry->[5] or $entry->[5] eq $product)) { $version = $entry->[0]; - $prediluvian = $entry->[5]; + $prediluvian = $entry->[6]; last; } } -# Give a little slack to the Windows 10 1709. -$maxfailedtests += 20 if ($version eq "win10"); +# Give a little slack to the latest Windows 10. +$maxfailedtests += 20 if ($version eq "win10_20"); if ($prediluvian and not $acceptprediluvianwin) { diff --git a/winetest/gather b/winetest/gather index 31ba8fa03..2c8900122 100755 --- a/winetest/gather +++ b/winetest/gather @@ -157,7 +157,10 @@ my %vista = (name => "Vista"); my %w2k8 = (name => "2008"); my %win7 = (name => "Win7"); my %win8 = (name => "Win8"); -my %win10 = (name => "Win10"); +my %win10_15= (name => "Win10.15"); +my %win10_17= (name => "Win10.17"); +my %win10_19= (name => "Win10.19"); +my %win10_20= (name => "Win10.20"); my %unknown = (name => "Other"); my %linux = (name => "Linux"); my %mac = (name => "Mac"); @@ -166,12 +169,23 @@ my %solaris = (name => "Solaris"); my %wine = (name => "Wine"); # Define the order of version groups in the summary -my @groups = (\%w95, \%w98, \%me, \%nt3, \%nt4, \%w2k, \%xp, \%w2k3, \%vista, \%w2k8, \%win7, \%win8, \%win10, +my @groups = (\%w95, \%w98, \%me, + \%nt3, \%nt4, \%w2k, \%xp, \%w2k3, + \%vista, \%w2k8, \%win7, \%win8, + \%win10_15, \%win10_17, \%win10_19, \%win10_20, \%unknown, \%linux, \%mac, \%bsd, \%solaris, \%wine); # Map dissect's IDs to the above hashes -my %idmap = (95=>\%w95, 98=>\%w98, me=>\%me, nt3=>\%nt3, nt4=>\%nt4, 2000=>\%w2k, - xp=>\%xp, 2003=>\%w2k3, vista=>\%vista, 2008=>\%w2k8, win7=>\%win7, win8=>\%win8, win10=>\%win10, +my %idmap = (95=>\%w95, 98=>\%w98, me=>\%me, + nt3=>\%nt3, nt4=>\%nt4, 2000=>\%w2k, + xp=>\%xp, 2003=>\%w2k3, vista=>\%vista, 2008=>\%w2k8, + win7=>\%win7, win8=>\%win8, + win1507=>\%win10_15, win1511=>\%win10_15, + win1607=>\%win10_15, + win1703=>\%win10_17, win1709=>\%win10_17, + win1803=>\%win10_17, win1809=>\%win10_17, + win1903=>\%win10_19, win1909=>\%win10_19, + win2004=>\%win10_20, win2009=>\%win10_20, unknown=>\%unknown, wine=>\%wine, linux=>\%linux, mac=>\%mac, bsd=>\%bsd, solaris=>\%solaris); -- 2.20.1

1 0

[PATCH v5] msxml3/tests: Test mxwriter domdoc output.
by Jefferson Carpenter Feb. 15, 2021

Feb. 15, 2021

Removed todo_wine from passing tests as per Francois's comment.

1 0

[PATCH v4] msxml3/tests: Test mxwriter domdoc output.
by Jefferson Carpenter Feb. 15, 2021

Feb. 15, 2021

Now tests that the destination document is cleared when startDocument is called, that createElement no longer succeeds, and that changes are visible more or less immediately after the SAX events that create them.

3 4