- Wine-Devel - list.winehq.org

[PATCH] ntdll: Implement idle time query for Mac OS.
by Chip Davis April 20, 2022

April 20, 2022

From: Chip Davis <cdavis(a)codeweavers.com> Signed-off-by: Chip Davis <cdavis5x(a)gmail.com> --- dlls/ntdll/unix/system.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/dlls/ntdll/unix/system.c b/dlls/ntdll/unix/system.c index 56a31f5e4757..1b92649ba284 100644 --- a/dlls/ntdll/unix/system.c +++ b/dlls/ntdll/unix/system.c @@ -1911,6 +1911,22 @@ static void get_performance_info( SYSTEM_PERFORMANCE_INFORMATION *info ) info->IdleTime.QuadPart = (ULONGLONG)ptimes[CP_IDLE] * 10000000 / clockrate.stathz; } } +#elif defined(__APPLE__) + { + host_name_port_t host = mach_host_self(); + struct host_cpu_load_info load_info; + mach_msg_type_number_t count; + + count = HOST_CPU_LOAD_INFO_COUNT; + if (host_statistics(host, HOST_CPU_LOAD_INFO, (host_info_t)&load_info, &count) == KERN_SUCCESS) + { + /* Believe it or not, based on my reading of XNU source, this is + * already in the units we want (100 ns). + */ + info->IdleTime.QuadPart = load_info.cpu_ticks[CPU_STATE_IDLE]; + } + mach_port_deallocate(mach_task_self(), host); + } #else { static ULONGLONG idle; -- 2.34.1

2 5

[PATCH v2] ntdll: Use the true maximum open FDs to set the limit on Mac.
by Chip Davis April 20, 2022

April 20, 2022

The actual upper limit is stored in the sysctl(3) "kern.maxfilesperproc". This value may be higher than the value of the OPEN_MAX constant. Using sysconf(3) with _SC_OPEN_MAX just returns the current value of RLIMIT_NOFILE, so that won't work here. Signed-off-by: Chip Davis <cdavis5x(a)gmail.com> --- v2: Remove unnecessary #ifdefs. --- dlls/ntdll/unix/loader.c | 22 ++++++++++++++++------ 1 file changed, 16 insertions(+), 6 deletions(-) diff --git a/dlls/ntdll/unix/loader.c b/dlls/ntdll/unix/loader.c index d5867e4aa754..0617fca631c1 100644 --- a/dlls/ntdll/unix/loader.c +++ b/dlls/ntdll/unix/loader.c @@ -452,13 +452,23 @@ static void set_max_limit( int limit ) rlimit.rlim_cur = rlimit.rlim_max; if (setrlimit( limit, &rlimit ) != 0) { -#if defined(__APPLE__) && defined(RLIMIT_NOFILE) && defined(OPEN_MAX) - /* On Leopard, setrlimit(RLIMIT_NOFILE, ...) fails on attempts to set - * rlim_cur above OPEN_MAX (even if rlim_max > OPEN_MAX). */ - if (limit == RLIMIT_NOFILE && rlimit.rlim_cur > OPEN_MAX) +#if defined(__APPLE__) && defined(RLIMIT_NOFILE) + if (limit == RLIMIT_NOFILE) { - rlimit.rlim_cur = OPEN_MAX; - setrlimit( limit, &rlimit ); + /* On Leopard, setrlimit(RLIMIT_NOFILE, ...) fails on attempts to set + * rlim_cur above OPEN_MAX (even if rlim_max > OPEN_MAX). */ + int open_max; + /* This is the true OPEN_MAX; it may be greater than the constant. + * sysconf(_SC_OPEN_MAX) won't work here: it just returns getrlimit(RLIMIT_NOFILE). */ + static int maxfilesperproc_oid[] = { CTL_KERN, KERN_MAXFILESPERPROC }; + size_t len = sizeof(open_max); + if (sysctl( maxfilesperproc_oid, ARRAY_SIZE(maxfilesperproc_oid), &open_max, &len, NULL, 0 ) != 0) + open_max = OPEN_MAX; + if (rlimit.rlim_cur > open_max) + { + rlimit.rlim_cur = open_max; + setrlimit( limit, &rlimit ); + } } #endif } -- 2.34.1

2 1

[PATCH] ntdll: Use the true maximum open FDs to set the limit on Mac.
by Chip Davis April 19, 2022

April 19, 2022

The actual upper limit is stored in the sysctl(3) "kern.maxfilesperproc". This value may be higher than the value of the OPEN_MAX constant. Using sysconf(3) with _SC_OPEN_MAX just returns the current value of RLIMIT_NOFILE, so that won't work here. Signed-off-by: Chip Davis <cdavis5x(a)gmail.com> --- Fixing this in the place it *actually* needs to be fixed this time... --- dlls/ntdll/unix/loader.c | 28 ++++++++++++++++++++++------ 1 file changed, 22 insertions(+), 6 deletions(-) diff --git a/dlls/ntdll/unix/loader.c b/dlls/ntdll/unix/loader.c index d5867e4aa754..dc54a54a225f 100644 --- a/dlls/ntdll/unix/loader.c +++ b/dlls/ntdll/unix/loader.c @@ -452,13 +452,29 @@ static void set_max_limit( int limit ) rlimit.rlim_cur = rlimit.rlim_max; if (setrlimit( limit, &rlimit ) != 0) { -#if defined(__APPLE__) && defined(RLIMIT_NOFILE) && defined(OPEN_MAX) - /* On Leopard, setrlimit(RLIMIT_NOFILE, ...) fails on attempts to set - * rlim_cur above OPEN_MAX (even if rlim_max > OPEN_MAX). */ - if (limit == RLIMIT_NOFILE && rlimit.rlim_cur > OPEN_MAX) +#if defined(__APPLE__) && defined(RLIMIT_NOFILE) && (defined(OPEN_MAX) || defined(KERN_MAXFILESPERPROC)) + if (limit == RLIMIT_NOFILE) { - rlimit.rlim_cur = OPEN_MAX; - setrlimit( limit, &rlimit ); + /* On Leopard, setrlimit(RLIMIT_NOFILE, ...) fails on attempts to set + * rlim_cur above OPEN_MAX (even if rlim_max > OPEN_MAX). */ + int open_max = INT_MAX; +#ifdef KERN_MAXFILESPERPROC + /* This is the true OPEN_MAX; it may be greater than the constant. + * sysconf(_SC_OPEN_MAX) won't work here: it just returns getrlimit(RLIMIT_NOFILE). */ + static int maxfilesperproc_oid[] = { CTL_KERN, KERN_MAXFILESPERPROC }; + size_t len = sizeof(open_max); + if (sysctl(maxfilesperproc_oid, ARRAY_SIZE(maxfilesperproc_oid), &open_max, &len, NULL, 0) != 0) + open_max = INT_MAX; +#endif +#ifdef OPEN_MAX + if (open_max == INT_MAX) + open_max = OPEN_MAX; +#endif + if (rlimit.rlim_cur > open_max) + { + rlimit.rlim_cur = open_max; + setrlimit( limit, &rlimit ); + } } #endif } -- 2.34.1

3 2

[PATCH resend] winex11: Call XFlush() after XRenderComposite().
by Connor McAdams April 19, 2022

April 19, 2022

Make sure to flush the output buffer after drawing. Helps applications that draw to HWNDs belonging to another process. Signed-off-by: Connor McAdams <cmcadams(a)codeweavers.com> --- This fixes an issue with Chromium based applications that draw on HWNDs that belong to other processes. Without calling XFlush, we often end up one frame behind, with the XFlush call in free_xrender_picture causing the update. This may not be the correct way to solve the issue I'm dealing with, so if not, I'd also appreciate feedback/ideas on a proper way to solve the problem. --- dlls/winex11.drv/xrender.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/dlls/winex11.drv/xrender.c b/dlls/winex11.drv/xrender.c index be64da90290..10a446b0f6b 100644 --- a/dlls/winex11.drv/xrender.c +++ b/dlls/winex11.drv/xrender.c @@ -1512,6 +1512,8 @@ static void xrender_blit( int op, Picture src_pict, Picture mask_pict, Picture d } pXRenderComposite( gdi_display, op, src_pict, mask_pict, dst_pict, x_offset, y_offset, 0, 0, x_dst, y_dst, width_dst, height_dst ); + + XFlush( gdi_display ); } /* Helper function for (stretched) mono->color blitting using xrender */ -- 2.25.1

1 0

[PATCH 1/7] kernel32/tests: Tweak tests results for w7pro64.
by Rémi Bernon April 19, 2022

April 19, 2022

Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/kernel32/tests/heap.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/dlls/kernel32/tests/heap.c b/dlls/kernel32/tests/heap.c index 53cff284c6f..a131a7f56e0 100644 --- a/dlls/kernel32/tests/heap.c +++ b/dlls/kernel32/tests/heap.c @@ -328,10 +328,12 @@ static void test_HeapCreate(void) ok( !!heap, "HeapCreate failed, error %lu\n", GetLastError() ); ok( !((ULONG_PTR)heap & 0xffff), "wrong heap alignment\n" ); - ptr = HeapAlloc( heap, 0, alloc_size - (0x400 + 0x80 * sizeof(void *)) ); + /* theshold between failure and success varies, and w7pro64 has a much larger overhead. */ + + ptr = HeapAlloc( heap, 0, alloc_size - (0x400 + 0x100 * sizeof(void *)) ); ok( !!ptr, "HeapAlloc failed, error %lu\n", GetLastError() ); size = HeapSize( heap, 0, ptr ); - ok( size == alloc_size - (0x400 + 0x80 * sizeof(void *)), + ok( size == alloc_size - (0x400 + 0x100 * sizeof(void *)), "HeapSize returned %#Ix, error %lu\n", size, GetLastError() ); ret = HeapFree( heap, 0, ptr ); ok( ret, "HeapFree failed, error %lu\n", GetLastError() ); @@ -438,7 +440,7 @@ static void test_HeapCreate(void) "got Region.dwCommittedSize %#lx\n", entries[0].Region.dwCommittedSize ); todo_wine ok( entries[0].Region.dwUnCommittedSize == 0x10000 - entries[0].Region.dwCommittedSize || - entries[0].Region.dwUnCommittedSize == 0x40000 - entries[0].Region.dwCommittedSize /* win7 */, + entries[0].Region.dwUnCommittedSize == 0x10000 * sizeof(void *) - entries[0].Region.dwCommittedSize /* win7 */, "got Region.dwUnCommittedSize %#lx\n", entries[0].Region.dwUnCommittedSize ); todo_wine ok( (BYTE *)entries[0].Region.lpFirstBlock == (BYTE *)entries[0].lpData + entries[0].cbData + 2 * sizeof(void *) || @@ -493,7 +495,7 @@ static void test_HeapCreate(void) ok( entries[3].lpData == ptr, "got lpData %p\n", entries[3].lpData ); todo_wine ok( entries[3].cbData == 5 * alloc_size, "got cbData %#lx\n", entries[3].cbData ); - ok( entries[3].cbOverhead == 0 || entries[3].cbOverhead == 0x20 /* win7 */, + ok( entries[3].cbOverhead == 0 || entries[3].cbOverhead == 8 * sizeof(void *) /* win7 */, "got cbOverhead %#x\n", entries[3].cbOverhead ); todo_wine ok( entries[3].iRegionIndex == 64, "got iRegionIndex %d\n", entries[3].iRegionIndex ); @@ -520,7 +522,7 @@ static void test_HeapCreate(void) ok( entries[4].lpData == ptr1, "got lpData %p\n", entries[4].lpData ); todo_wine ok( entries[4].cbData == 5 * alloc_size, "got cbData %#lx\n", entries[4].cbData ); - ok( entries[4].cbOverhead == 0 || entries[4].cbOverhead == 0x20 /* win7 */, + ok( entries[4].cbOverhead == 0 || entries[4].cbOverhead == 8 * sizeof(void *) /* win7 */, "got cbOverhead %#x\n", entries[4].cbOverhead ); todo_wine ok( entries[4].iRegionIndex == 64, "got iRegionIndex %d\n", entries[4].iRegionIndex ); -- 2.35.1

2 12

[PATCH resend 1/5] mf/tests: Simplify handling of broken Win7 results.
by Rémi Bernon April 19, 2022

April 19, 2022

Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- Supersedes: 231858, 231871-231874 FWIW I'm also happy to move the tests elsewhere if it's a burden. I now think that, though there's no guarantee the modules won't change, the transform / dmo classes and their tests should perhaps better live in the same modules where native has them. It is in theory possible that some applications load them and instantiate the classes directly, and I don't think it is likely that windows will remove these modules. The main blocker for that is winegstreamer being a dll and the classes being registered in it. Perhaps it should be a static lib instead, or expose the unixlib interface as its PE entry points. dlls/mf/tests/mf.c | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-) diff --git a/dlls/mf/tests/mf.c b/dlls/mf/tests/mf.c index 31a84e47bc1..b2dd807dd33 100644 --- a/dlls/mf/tests/mf.c +++ b/dlls/mf/tests/mf.c @@ -6631,6 +6631,7 @@ static void test_h264_decoder(void) IMFMediaType *media_type; IMFTransform *transform; DWORD status, length; + BOOL is_win7 = FALSE; ULONG i, ret, flags; HANDLE output_file; IMFSample *sample; @@ -6741,10 +6742,9 @@ static void test_h264_decoder(void) winetest_push_context("out %lu", i); ok(hr == S_OK, "GetOutputAvailableType returned %#lx\n", hr); check_media_type(media_type, default_outputs[i], -1); - if (FAILED(hr = IMFMediaType_GetItem(media_type, &MF_MT_VIDEO_ROTATION, NULL))) - check_media_type(media_type, default_outputs_win7[i], -1); - else - check_media_type(media_type, default_outputs_extra[i], -1); + hr = IMFMediaType_GetItem(media_type, &MF_MT_VIDEO_ROTATION, NULL); + is_win7 = broken(FAILED(hr)); + check_media_type(media_type, is_win7 ? default_outputs_win7[i] : default_outputs_extra[i], -1); ret = IMFMediaType_Release(media_type); ok(ret == 0, "Release returned %lu\n", ret); winetest_pop_context(); @@ -6772,11 +6772,9 @@ static void test_h264_decoder(void) init_media_type(media_type, output_type_desc, i + 1); } hr = IMFTransform_SetOutputType(transform, 0, media_type, 0); - if (broken(hr == MF_E_INVALIDMEDIATYPE)) - { - init_media_type(media_type, output_type_desc_win7, ARRAY_SIZE(output_type_desc_win7) - 1); - hr = IMFTransform_SetOutputType(transform, 0, media_type, 0); - } + ok(hr == (is_win7 ? MF_E_INVALIDMEDIATYPE : S_OK), "SetOutputType returned %#lx.\n", hr); + init_media_type(media_type, is_win7 ? output_type_desc_win7 : output_type_desc, -1); + hr = IMFTransform_SetOutputType(transform, 0, media_type, 0); ok(hr == S_OK, "SetOutputType returned %#lx.\n", hr); ret = IMFMediaType_Release(media_type); ok(ret == 1, "Release returned %lu\n", ret); -- 2.35.1

2 7

[tools 1/4] testbot/cgi: Add support for importing JavaScript files.
by Francois Gouget April 19, 2022

April 19, 2022

Signed-off-by: Francois Gouget <fgouget(a)codeweavers.com> --- testbot/lib/ObjectModel/CGI/Page.pm | 7 +++++++ testbot/lib/WineTestBot/CGI/PageBase.pm | 11 +++++++++++ 2 files changed, 18 insertions(+) diff --git a/testbot/lib/ObjectModel/CGI/Page.pm b/testbot/lib/ObjectModel/CGI/Page.pm index d7463b5a8..406ca4ff8 100644 --- a/testbot/lib/ObjectModel/CGI/Page.pm +++ b/testbot/lib/ObjectModel/CGI/Page.pm @@ -369,6 +369,13 @@ sub GetTitle($) return undef; } +sub GenerateImportJS($$) +{ + my ($self, $Filename) = @_; + + $self->{PageBase}->GenerateImportJS($Filename); +} + sub GenerateHttpHeaders($) { my ($self) = @_; diff --git a/testbot/lib/WineTestBot/CGI/PageBase.pm b/testbot/lib/WineTestBot/CGI/PageBase.pm index 477a2ff15..0dacbf6dd 100644 --- a/testbot/lib/WineTestBot/CGI/PageBase.pm +++ b/testbot/lib/WineTestBot/CGI/PageBase.pm @@ -397,6 +397,17 @@ sub GetPageTitle($$) return $Title; } +sub GenerateImportJS($$) +{ + my ($self, $Filename) = @_; + + if (!$self->{js}->{$Filename}) + { + print "<script type='text/javascript' src='$Filename'></script>\n"; + $self->{js}->{$Filename} = 1; + } +} + sub GenerateHttpHeaders($) { my ($self) = @_; -- 2.30.2

1 3

[PATCH 1/8] jscript: Initialize hres in String constructor.
by Gabriel Ivăncescu April 19, 2022

April 19, 2022

Signed-off-by: Gabriel Ivăncescu <gabrielopcode(a)gmail.com> --- Sorry about this. c464c1bd4781ca650a5397ff6a2cfd574c1877f0 made it possibly used uninitialized when argc was 0 and return value was not used. dlls/jscript/string.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/dlls/jscript/string.c b/dlls/jscript/string.c index e73b9e8..dac89c9 100644 --- a/dlls/jscript/string.c +++ b/dlls/jscript/string.c @@ -1628,7 +1628,7 @@ static HRESULT StringConstr_fromCharCode(script_ctx_t *ctx, jsval_t vthis, WORD static HRESULT StringConstr_value(script_ctx_t *ctx, jsval_t vthis, WORD flags, unsigned argc, jsval_t *argv, jsval_t *r) { - HRESULT hres; + HRESULT hres = S_OK; TRACE("\n"); -- 2.34.1

3 20

[PATCH 9/9] win32u: Move NtUserFindWindowEx implementation from user32.
by Jacek Caban April 19, 2022

April 19, 2022

Signed-off-by: Jacek Caban <jacek(a)codeweavers.com> --- dlls/user32/win.c | 61 +++++++---------------------------------- dlls/win32u/syscall.c | 1 + dlls/win32u/win32u.spec | 2 +- dlls/win32u/window.c | 49 +++++++++++++++++++++++++++++++++ dlls/wow64win/syscall.h | 1 + dlls/wow64win/user.c | 16 +++++++++++ include/ntuser.h | 2 ++ 7 files changed, 80 insertions(+), 52 deletions(-)

2 1

[PATCH 8/9] winex11: Directly use ntdll in display_device_init mutex.
by Jacek Caban April 19, 2022

April 19, 2022

Signed-off-by: Jacek Caban <jacek(a)codeweavers.com> --- dlls/winex11.drv/display.c | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-)

2 1