wine-devel October 2021

wine-devel@winehq.org

81 participants
772 discussions

[PATCH] kernelbase: Fix string size variable overflow in GetModuleFileNameW().
by Paul Gofman 04 Oct '21

04 Oct '21

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=51833 Signed-off-by: Paul Gofman <pgofman(a)codeweavers.com> --- dlls/kernel32/tests/module.c | 5 +++++ dlls/kernelbase/loader.c | 2 +- 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/dlls/kernel32/tests/module.c b/dlls/kernel32/tests/module.c index 2487b9d65ab..60654754302 100644 --- a/dlls/kernel32/tests/module.c +++ b/dlls/kernel32/tests/module.c @@ -190,6 +190,11 @@ static void testGetModuleFileName(const char* name) ok(len1A / 2 == len2A, "Correct length in GetModuleFilenameA with buffer too small (%d/%d)\n", len1A / 2, len2A); + + len1A = GetModuleFileNameA(hMod, bufA, 0x10000); + ok(len1A > 0, "Getting module filename for handle %p\n", hMod); + len1W = GetModuleFileNameW(hMod, bufW, 0x10000); + ok(len1W > 0, "Getting module filename for handle %p\n", hMod); } static void testGetModuleFileName_Wrong(void) diff --git a/dlls/kernelbase/loader.c b/dlls/kernelbase/loader.c index 145d721bc26..b2bbdc29234 100644 --- a/dlls/kernelbase/loader.c +++ b/dlls/kernelbase/loader.c @@ -311,7 +311,7 @@ DWORD WINAPI DECLSPEC_HOTPATCH GetModuleFileNameW( HMODULE module, LPWSTR filena } name.Buffer = filename; - name.MaximumLength = size * sizeof(WCHAR); + name.MaximumLength = min( size, (USHORT)~0 / sizeof(WCHAR) ) * sizeof(WCHAR); status = LdrGetDllFullName( module, &name ); if (!status || status == STATUS_BUFFER_TOO_SMALL) len = name.Length / sizeof(WCHAR); SetLastError( RtlNtStatusToDosError( status )); -- 2.31.1

1 0

[PATCH] winebus.sys: Use type and size of the current event, not the previous one.
by Rémi Bernon 04 Oct '21

04 Oct '21

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=51831 Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/winebus.sys/unixlib.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/dlls/winebus.sys/unixlib.c b/dlls/winebus.sys/unixlib.c index 49a42ea04ad..be030255705 100644 --- a/dlls/winebus.sys/unixlib.c +++ b/dlls/winebus.sys/unixlib.c @@ -350,8 +350,8 @@ BOOL bus_event_queue_pop(struct list *queue, struct bus_event *event) tmp = LIST_ENTRY(entry, struct bus_event, entry); list_remove(entry); - if (event->type != BUS_EVENT_TYPE_INPUT_REPORT) size = sizeof(*event); - else size = offsetof(struct bus_event, input_report.buffer[event->input_report.length]); + if (tmp->type != BUS_EVENT_TYPE_INPUT_REPORT) size = sizeof(*tmp); + else size = offsetof(struct bus_event, input_report.buffer[tmp->input_report.length]); memcpy(event, tmp, size); free(tmp); -- 2.33.0

1 0

[PATCH] hidclass.sys: Don't send WM_INPUT_DEVICE_CHANGE for WINEXINPUT devices.
by Rémi Bernon 04 Oct '21

04 Oct '21

From: Arkadiusz Hiler <ahiler(a)codeweavers.com> For the same reason we don't send WM_INPUT messages. Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/hidclass.sys/pnp.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/dlls/hidclass.sys/pnp.c b/dlls/hidclass.sys/pnp.c index a47af413dbd..dbf04c0a403 100644 --- a/dlls/hidclass.sys/pnp.c +++ b/dlls/hidclass.sys/pnp.c @@ -114,6 +114,8 @@ static void send_wm_input_device_change(BASE_DEVICE_EXTENSION *ext, LPARAM param RAWINPUT rawinput; INPUT input; + if (!IsEqualGUID( ext->class_guid, &GUID_DEVINTERFACE_HID )) return; + rawinput.header.dwType = RIM_TYPEHID; rawinput.header.dwSize = offsetof(RAWINPUT, data.hid.bRawData[2 * sizeof(USAGE)]); rawinput.header.hDevice = ULongToHandle(ext->u.pdo.rawinput_handle); -- 2.33.0

1 0

[PATCH] hidclass.sys: Overwrite queued reports as FIFO instead of LIFO.
by Rémi Bernon 04 Oct '21

04 Oct '21

Based on a patch from Ivo Ivanov <logos128(a)gmail.com>. The issue causes severe skipping and non smooth movement tracking in apps/games, when the HidP/HidD APIs are used to control the device (joysticks, controllers, steering wheels, etc.). Usually such devices use constant stream of INPUT reports to report their coords, so any report skipping or change of the sequence, when the interested apps are reading, would lead to such issues. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=51824 Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/hidclass.sys/device.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/dlls/hidclass.sys/device.c b/dlls/hidclass.sys/device.c index ae8d80c2e95..39ae678d111 100644 --- a/dlls/hidclass.sys/device.c +++ b/dlls/hidclass.sys/device.c @@ -190,10 +190,12 @@ static void hid_queue_push_report( struct hid_queue *queue, struct hid_report *r KeAcquireSpinLock( &queue->lock, &irql ); prev = queue->reports[i]; queue->reports[i] = report; - if (next != queue->read_idx) queue->write_idx = next; + if (next == queue->read_idx) queue->read_idx = next + 1; + if (queue->read_idx >= queue->length) queue->read_idx = 0; KeReleaseSpinLock( &queue->lock, irql ); hid_report_decref( prev ); + queue->write_idx = next; } static struct hid_report *hid_queue_pop_report( struct hid_queue *queue ) -- 2.33.0

1 0

[PATCH 1/3] hidclass.sys: Use the report length when copying report data.
by Rémi Bernon 04 Oct '21

04 Oct '21

From: Ivo Ivanov <logos128(a)gmail.com> Instead of the HID descriptor input report length. Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=51828 Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/hidclass.sys/device.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/dlls/hidclass.sys/device.c b/dlls/hidclass.sys/device.c index 026ec1efeb1..632e98ab740 100644 --- a/dlls/hidclass.sys/device.c +++ b/dlls/hidclass.sys/device.c @@ -216,7 +216,6 @@ static struct hid_report *hid_queue_pop_report( struct hid_queue *queue ) static void hid_device_queue_input( DEVICE_OBJECT *device, HID_XFER_PACKET *packet ) { BASE_DEVICE_EXTENSION *ext = device->DeviceExtension; - HIDP_COLLECTION_DESC *desc = ext->u.pdo.device_desc.CollectionDesc; const BOOL polled = ext->u.pdo.information.Polled; struct hid_report *last_report, *report; struct hid_queue *queue; @@ -270,7 +269,7 @@ static void hid_device_queue_input( DEVICE_OBJECT *device, HID_XFER_PACKET *pack if (!(irp = hid_queue_pop_irp( queue ))) break; if (!(report = hid_queue_pop_report( queue ))) hid_report_incref( (report = last_report) ); - memcpy( irp->AssociatedIrp.SystemBuffer, report->buffer, desc->InputLength ); + memcpy( irp->AssociatedIrp.SystemBuffer, report->buffer, report->length ); irp->IoStatus.Information = report->length; irp->IoStatus.Status = STATUS_SUCCESS; hid_report_decref( report ); @@ -652,7 +651,7 @@ NTSTATUS WINAPI pdo_read(DEVICE_OBJECT *device, IRP *irp) irp->IoStatus.Information = 0; if ((report = hid_queue_pop_report( queue ))) { - memcpy( irp->AssociatedIrp.SystemBuffer, report->buffer, desc->InputLength ); + memcpy( irp->AssociatedIrp.SystemBuffer, report->buffer, report->length ); irp->IoStatus.Information = report->length; irp->IoStatus.Status = STATUS_SUCCESS; hid_report_decref( report ); -- 2.33.0

1 2

[PATCH 1/7] dinput8/tests: Add some HID joystick IDirectInputDevice8_EnumCreatedEffectObjects tests.
by Rémi Bernon 04 Oct '21

04 Oct '21

Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/dinput8/tests/hid.c | 46 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 46 insertions(+) diff --git a/dlls/dinput8/tests/hid.c b/dlls/dinput8/tests/hid.c index 6bf887e4c7a..cc2f3268e6f 100644 --- a/dlls/dinput8/tests/hid.c +++ b/dlls/dinput8/tests/hid.c @@ -3422,6 +3422,32 @@ static BOOL CALLBACK check_effect_count( const DIEFFECTINFOW *effect, void *args return DIENUM_CONTINUE; } +static BOOL CALLBACK check_no_created_effect_objects( IDirectInputEffect *effect, void *context ) +{ + ok( 0, "unexpected effect %p\n", effect ); + return DIENUM_CONTINUE; +} + +struct check_created_effect_params +{ + IDirectInputEffect *expect_effect; + DWORD count; +}; + +static BOOL CALLBACK check_created_effect_objects( IDirectInputEffect *effect, void *context ) +{ + struct check_created_effect_params *params = context; + ULONG ref; + + ok( effect == params->expect_effect, "got effect %p, expected %p\n", effect, params->expect_effect ); + params->count++; + + IDirectInputEffect_AddRef( effect ); + ref = IDirectInputEffect_Release( effect ); + ok( ref == 1, "got ref %u, expected 1\n", ref ); + return DIENUM_CONTINUE; +} + static void test_simple_joystick(void) { #include "psh_hid_macros.h" @@ -4721,6 +4747,15 @@ static void test_simple_joystick(void) hr = IDirectInputDevice8_CreateEffect( device, &GUID_Sine, NULL, &effect, NULL ); ok( hr == DIERR_UNSUPPORTED, "IDirectInputDevice8_CreateEffect returned %#x\n", hr ); + hr = IDirectInputDevice8_EnumCreatedEffectObjects( device, NULL, effect, 0 ); + ok( hr == DIERR_INVALIDPARAM, "IDirectInputDevice8_EnumCreatedEffectObjects returned %#x\n", hr ); + hr = IDirectInputDevice8_EnumCreatedEffectObjects( device, check_no_created_effect_objects, effect, 0xdeadbeef ); + todo_wine + ok( hr == DIERR_INVALIDPARAM, "IDirectInputDevice8_EnumCreatedEffectObjects returned %#x\n", hr ); + hr = IDirectInputDevice8_EnumCreatedEffectObjects( device, check_no_created_effect_objects, (void *)0xdeadbeef, 0 ); + todo_wine + ok( hr == DI_OK, "IDirectInputDevice8_EnumCreatedEffectObjects returned %#x\n", hr ); + hr = IDirectInputDevice8_Escape( device, NULL ); todo_wine ok( hr == E_POINTER, "IDirectInputDevice8_Escape returned: %#x\n", hr ); @@ -5098,6 +5133,7 @@ static BOOL test_device_types(void) static void test_periodic_effect( IDirectInputDevice8W *device, HANDLE file ) { + struct check_created_effect_params check_params = {0}; IDirectInputEffect *effect; HRESULT hr; ULONG ref; @@ -5117,6 +5153,16 @@ static void test_periodic_effect( IDirectInputDevice8W *device, HANDLE file ) ok( hr == DI_OK, "IDirectInputDevice8_CreateEffect returned %#x\n", hr ); if (hr != DI_OK) return; + hr = IDirectInputDevice8_EnumCreatedEffectObjects( device, check_no_created_effect_objects, effect, 0xdeadbeef ); + todo_wine + ok( hr == DIERR_INVALIDPARAM, "IDirectInputDevice8_EnumCreatedEffectObjects returned %#x\n", hr ); + check_params.expect_effect = effect; + hr = IDirectInputDevice8_EnumCreatedEffectObjects( device, check_created_effect_objects, &check_params, 0 ); + todo_wine + ok( hr == DI_OK, "IDirectInputDevice8_EnumCreatedEffectObjects returned %#x\n", hr ); + todo_wine + ok( check_params.count == 1, "got count %u, expected 1\n", check_params.count ); + ref = IDirectInputEffect_Release( effect ); ok( ref == 0, "IDirectInputDeviceW_Release returned %d\n", ref ); } -- 2.33.0

1 6

[PATCH] wininet: Clear cached credentials after the first failed attempt.
by Alistair Leslie-Hughes 04 Oct '21

04 Oct '21

Windows when using cached credentials will use them on the first challenge, if then a second 403 (ACCESS_DENIED) is received, the user is prompted again but this time with the password and save checkbox cleared. Signed-off-by: Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> --- dlls/wininet/dialogs.c | 21 ++++++++++++++++++++- dlls/wininet/internet.h | 1 + 2 files changed, 21 insertions(+), 1 deletion(-) diff --git a/dlls/wininet/dialogs.c b/dlls/wininet/dialogs.c index a48c75865ac..94def22350e 100644 --- a/dlls/wininet/dialogs.c +++ b/dlls/wininet/dialogs.c @@ -152,9 +152,25 @@ static BOOL WININET_GetSetPassword( HWND hdlg, LPCWSTR szServer, p = wcschr( szUserPass, ':' ); if( p ) { + struct WININET_ErrorDlgParams *params; + HWND hwnd; + + params = (struct WININET_ErrorDlgParams*) + GetWindowLongPtrW( hdlg, GWLP_USERDATA ); + *p = 0; SetWindowTextW( hUserItem, szUserPass ); - SetWindowTextW( hPassItem, p+1 ); + + if (params->req->retrycnt == 0) + { + SetWindowTextW( hPassItem, p+1 ); + + hwnd = GetDlgItem( hdlg, IDC_SAVEPASSWORD ); + if (hwnd) + SendMessageW(hwnd, BM_SETCHECK, BST_CHECKED, 0); + } + else + WININET_GetSetPassword( hdlg, szServer, szRealm, TRUE ); } return TRUE; @@ -264,6 +280,8 @@ static INT_PTR WINAPI WININET_ProxyPasswordDialog( WININET_GetSetPassword( hdlg, params->req->session->appInfo->proxy, szRealm, TRUE ); WININET_SetAuthorization( params->req, username, password, TRUE ); + params->req->retrycnt++; + EndDialog( hdlg, ERROR_INTERNET_FORCE_RETRY ); return TRUE; } @@ -340,6 +358,7 @@ static INT_PTR WINAPI WININET_PasswordDialog( WININET_GetSetPassword( hdlg, params->req->session->hostName, szRealm, TRUE ); } WININET_SetAuthorization( params->req, username, password, FALSE ); + params->req->retrycnt++; EndDialog( hdlg, ERROR_INTERNET_FORCE_RETRY ); return TRUE; diff --git a/dlls/wininet/internet.h b/dlls/wininet/internet.h index 8538699811c..cccfc24898d 100644 --- a/dlls/wininet/internet.h +++ b/dlls/wininet/internet.h @@ -361,6 +361,7 @@ typedef struct LPWSTR statusText; DWORD bytesToWrite; DWORD bytesWritten; + INT retrycnt; /* Count of number of Auth retries */ CRITICAL_SECTION headers_section; /* section to protect the headers array */ HTTPHEADERW *custHeaders; -- 2.33.0

1 0

[PATCH 1/3] winebuild: Move the CALL32_CBClient[Ex] implementation to krnl386.
by Zebediah Figura 04 Oct '21

04 Oct '21

Except for the return thunk. Signed-off-by: Zebediah Figura <zfigura(a)codeweavers.com> --- dlls/krnl386.exe16/thunk.c | 164 ++++++++++++++++++++++++++++++++++++- tools/winebuild/relay.c | 155 ----------------------------------- 2 files changed, 162 insertions(+), 157 deletions(-) diff --git a/dlls/krnl386.exe16/thunk.c b/dlls/krnl386.exe16/thunk.c index 98dbf31a770..4a373b6c7cb 100644 --- a/dlls/krnl386.exe16/thunk.c +++ b/dlls/krnl386.exe16/thunk.c @@ -1951,10 +1951,134 @@ void WINAPI CBClientGlueSL( CONTEXT *context ) context->Eip = OFFSETOF ( glue ); } +/******************************************************************* + * CALL32_CBClient + * + * Call a CBClient relay stub from 32-bit code (KERNEL.620). + * + * Since the relay stub is itself 32-bit, this should not be a problem; + * unfortunately, the relay stubs are expected to switch back to a + * 16-bit stack (and 16-bit code) after completion :-( + * + * This would conflict with our 16- vs. 32-bit stack handling, so + * we simply switch *back* to our 32-bit stack before returning to + * the caller ... + * + * The CBClient relay stub expects to be called with the following + * 16-bit stack layout, and with ebp and ebx pointing into the 16-bit + * stack at the designated places: + * + * ... + * (ebp+14) original arguments to the callback routine + * (ebp+10) far return address to original caller + * (ebp+6) Thunklet target address + * (ebp+2) Thunklet relay ID code + * (ebp) BP (saved by CBClientGlueSL) + * (ebp-2) SI (saved by CBClientGlueSL) + * (ebp-4) DI (saved by CBClientGlueSL) + * (ebp-6) DS (saved by CBClientGlueSL) + * + * ... buffer space used by the 16-bit side glue for temp copies + * + * (ebx+4) far return address to 16-bit side glue code + * (ebx) saved 16-bit ss:sp (pointing to ebx+4) + * + * The 32-bit side glue code accesses both the original arguments (via ebp) + * and the temporary copies prepared by the 16-bit side glue (via ebx). + * After completion, the stub will load ss:sp from the buffer at ebx + * and perform a far return to 16-bit code. + * + * To trick the relay stub into returning to us, we replace the 16-bit + * return address to the glue code by a cs:ip pair pointing to our + * return entry point (the original return address is saved first). + * Our return stub thus called will then reload the 32-bit ss:esp and + * return to 32-bit code (by using and ss:esp value that we have also + * pushed onto the 16-bit stack before and a cs:eip values found at + * that position on the 32-bit stack). The ss:esp to be restored is + * found relative to the 16-bit stack pointer at: + * + * (ebx-4) ss (flat) + * (ebx-8) sp (32-bit stack pointer) + * + * The second variant of this routine, CALL32_CBClientEx, which is used + * to implement KERNEL.621, has to cope with yet another problem: Here, + * the 32-bit side directly returns to the caller of the CBClient thunklet, + * restoring registers saved by CBClientGlueSL and cleaning up the stack. + * As we have to return to our 32-bit code first, we have to adapt the + * layout of our temporary area so as to include values for the registers + * that are to be restored, and later (in the implementation of KERNEL.621) + * we *really* restore them. The return stub restores DS, DI, SI, and BP + * from the stack, skips the next 8 bytes (CBClient relay code / target), + * and then performs a lret NN, where NN is the number of arguments to be + * removed. Thus, we prepare our temporary area as follows: + * + * (ebx+22) 16-bit cs (this segment) + * (ebx+20) 16-bit ip ('16-bit' return entry point) + * (ebx+16) 32-bit ss (flat) + * (ebx+12) 32-bit sp (32-bit stack pointer) + * (ebx+10) 16-bit bp (points to ebx+24) + * (ebx+8) 16-bit si (ignored) + * (ebx+6) 16-bit di (ignored) + * (ebx+4) 16-bit ds (we actually use the flat DS here) + * (ebx+2) 16-bit ss (16-bit stack segment) + * (ebx+0) 16-bit sp (points to ebx+4) + * + * Note that we ensure that DS is not changed and remains the flat segment, + * and the 32-bit stack pointer our own return stub needs fits just + * perfectly into the 8 bytes that are skipped by the Windows stub. + * One problem is that we have to determine the number of removed arguments, + * as these have to be really removed in KERNEL.621. Thus, the BP value + * that we place in the temporary area to be restored, contains the value + * that SP would have if no arguments were removed. By comparing the actual + * value of SP with this value in our return stub we can compute the number + * of removed arguments. This is then returned to KERNEL.621. + * + * The stack layout of this function: + * (ebp+20) nArgs pointer to variable receiving nr. of args (Ex only) + * (ebp+16) esi pointer to caller's esi value + * (ebp+12) arg ebp value to be set for relay stub + * (ebp+8) func CBClient relay stub address + * (ebp+4) ret addr + * (ebp) ebp + */ +extern DWORD CALL32_CBClient( FARPROC proc, LPWORD args, WORD *stackLin, DWORD *esi ); +__ASM_GLOBAL_FUNC( CALL32_CBClient, + "pushl %ebp\n\t" + __ASM_CFI(".cfi_adjust_cfa_offset 4\n\t") + __ASM_CFI(".cfi_rel_offset %ebp,0\n\t") + "movl %esp,%ebp\n\t" + __ASM_CFI(".cfi_def_cfa_register %ebp\n\t") + "pushl %edi\n\t" + __ASM_CFI(".cfi_rel_offset %edi,-4\n\t") + "pushl %esi\n\t" + __ASM_CFI(".cfi_rel_offset %esi,-8\n\t") + "pushl %ebx\n\t" + __ASM_CFI(".cfi_rel_offset %ebx,-12\n\t") + "movl 16(%ebp),%ebx\n\t" + "leal -8(%esp),%eax\n\t" + "movl %eax,-8(%ebx)\n\t" + "movl 20(%ebp),%esi\n\t" + "movl (%esi),%esi\n\t" + "movl 8(%ebp),%eax\n\t" + "movl 12(%ebp),%ebp\n\t" + "pushl %cs\n\t" + "call *%eax\n\t" + "movl 32(%esp),%edi\n\t" + "movl %esi,(%edi)\n\t" + "popl %ebx\n\t" + __ASM_CFI(".cfi_same_value %ebx\n\t") + "popl %esi\n\t" + __ASM_CFI(".cfi_same_value %esi\n\t") + "popl %edi\n\t" + __ASM_CFI(".cfi_same_value %edi\n\t") + "popl %ebp\n\t" + __ASM_CFI(".cfi_def_cfa %esp,4\n\t") + __ASM_CFI(".cfi_same_value %ebp\n\t") + "ret\n\t" ) + /*********************************************************************** * CBClientThunkSL (KERNEL.620) */ -extern DWORD CALL32_CBClient( FARPROC proc, LPWORD args, WORD *stackLin, DWORD *esi ); void WINAPI CBClientThunkSL( CONTEXT *context ) { /* Call 32-bit relay code */ @@ -1976,10 +2100,46 @@ void WINAPI CBClientThunkSL( CONTEXT *context ) stack16_pop( 12 ); } +extern DWORD CALL32_CBClientEx( FARPROC proc, LPWORD args, WORD *stackLin, DWORD *esi, INT *nArgs ); +__ASM_GLOBAL_FUNC( CALL32_CBClientEx, + "pushl %ebp\n\t" + __ASM_CFI(".cfi_adjust_cfa_offset 4\n\t") + __ASM_CFI(".cfi_rel_offset %ebp,0\n\t") + "movl %esp,%ebp\n\t" + __ASM_CFI(".cfi_def_cfa_register %ebp\n\t") + "pushl %edi\n\t" + __ASM_CFI(".cfi_rel_offset %edi,-4\n\t") + "pushl %esi\n\t" + __ASM_CFI(".cfi_rel_offset %esi,-8\n\t") + "pushl %ebx\n\t" + __ASM_CFI(".cfi_rel_offset %ebx,-12\n\t") + "movl 16(%ebp),%ebx\n\t" + "leal -8(%esp),%eax\n\t" + "movl %eax,12(%ebx)\n\t" + "movl 20(%ebp),%esi\n\t" + "movl (%esi),%esi\n\t" + "movl 8(%ebp),%eax\n\t" + "movl 12(%ebp),%ebp\n\t" + "pushl %cs\n\t" + "call *%eax\n\t" + "movl 32(%esp),%edi\n\t" + "movl %esi,(%edi)\n\t" + "movl 36(%esp),%ebx\n\t" + "movl %ebp,(%ebx)\n\t" + "popl %ebx\n\t" + __ASM_CFI(".cfi_same_value %ebx\n\t") + "popl %esi\n\t" + __ASM_CFI(".cfi_same_value %esi\n\t") + "popl %edi\n\t" + __ASM_CFI(".cfi_same_value %edi\n\t") + "popl %ebp\n\t" + __ASM_CFI(".cfi_def_cfa %esp,4\n\t") + __ASM_CFI(".cfi_same_value %ebp\n\t") + "ret\n\t" ) + /*********************************************************************** * CBClientThunkSLEx (KERNEL.621) */ -extern DWORD CALL32_CBClientEx( FARPROC proc, LPWORD args, WORD *stackLin, DWORD *esi, INT *nArgs ); void WINAPI CBClientThunkSLEx( CONTEXT *context ) { /* Call 32-bit relay code */ diff --git a/tools/winebuild/relay.c b/tools/winebuild/relay.c index abe2bdae6ed..eaa70888986 100644 --- a/tools/winebuild/relay.c +++ b/tools/winebuild/relay.c @@ -564,163 +564,8 @@ static void BuildRet16Func(void) } -/******************************************************************* - * BuildCallTo32CBClient - * - * Call a CBClient relay stub from 32-bit code (KERNEL.620). - * - * Since the relay stub is itself 32-bit, this should not be a problem; - * unfortunately, the relay stubs are expected to switch back to a - * 16-bit stack (and 16-bit code) after completion :-( - * - * This would conflict with our 16- vs. 32-bit stack handling, so - * we simply switch *back* to our 32-bit stack before returning to - * the caller ... - * - * The CBClient relay stub expects to be called with the following - * 16-bit stack layout, and with ebp and ebx pointing into the 16-bit - * stack at the designated places: - * - * ... - * (ebp+14) original arguments to the callback routine - * (ebp+10) far return address to original caller - * (ebp+6) Thunklet target address - * (ebp+2) Thunklet relay ID code - * (ebp) BP (saved by CBClientGlueSL) - * (ebp-2) SI (saved by CBClientGlueSL) - * (ebp-4) DI (saved by CBClientGlueSL) - * (ebp-6) DS (saved by CBClientGlueSL) - * - * ... buffer space used by the 16-bit side glue for temp copies - * - * (ebx+4) far return address to 16-bit side glue code - * (ebx) saved 16-bit ss:sp (pointing to ebx+4) - * - * The 32-bit side glue code accesses both the original arguments (via ebp) - * and the temporary copies prepared by the 16-bit side glue (via ebx). - * After completion, the stub will load ss:sp from the buffer at ebx - * and perform a far return to 16-bit code. - * - * To trick the relay stub into returning to us, we replace the 16-bit - * return address to the glue code by a cs:ip pair pointing to our - * return entry point (the original return address is saved first). - * Our return stub thus called will then reload the 32-bit ss:esp and - * return to 32-bit code (by using and ss:esp value that we have also - * pushed onto the 16-bit stack before and a cs:eip values found at - * that position on the 32-bit stack). The ss:esp to be restored is - * found relative to the 16-bit stack pointer at: - * - * (ebx-4) ss (flat) - * (ebx-8) sp (32-bit stack pointer) - * - * The second variant of this routine, CALL32_CBClientEx, which is used - * to implement KERNEL.621, has to cope with yet another problem: Here, - * the 32-bit side directly returns to the caller of the CBClient thunklet, - * restoring registers saved by CBClientGlueSL and cleaning up the stack. - * As we have to return to our 32-bit code first, we have to adapt the - * layout of our temporary area so as to include values for the registers - * that are to be restored, and later (in the implementation of KERNEL.621) - * we *really* restore them. The return stub restores DS, DI, SI, and BP - * from the stack, skips the next 8 bytes (CBClient relay code / target), - * and then performs a lret NN, where NN is the number of arguments to be - * removed. Thus, we prepare our temporary area as follows: - * - * (ebx+22) 16-bit cs (this segment) - * (ebx+20) 16-bit ip ('16-bit' return entry point) - * (ebx+16) 32-bit ss (flat) - * (ebx+12) 32-bit sp (32-bit stack pointer) - * (ebx+10) 16-bit bp (points to ebx+24) - * (ebx+8) 16-bit si (ignored) - * (ebx+6) 16-bit di (ignored) - * (ebx+4) 16-bit ds (we actually use the flat DS here) - * (ebx+2) 16-bit ss (16-bit stack segment) - * (ebx+0) 16-bit sp (points to ebx+4) - * - * Note that we ensure that DS is not changed and remains the flat segment, - * and the 32-bit stack pointer our own return stub needs fits just - * perfectly into the 8 bytes that are skipped by the Windows stub. - * One problem is that we have to determine the number of removed arguments, - * as these have to be really removed in KERNEL.621. Thus, the BP value - * that we place in the temporary area to be restored, contains the value - * that SP would have if no arguments were removed. By comparing the actual - * value of SP with this value in our return stub we can compute the number - * of removed arguments. This is then returned to KERNEL.621. - * - * The stack layout of this function: - * (ebp+20) nArgs pointer to variable receiving nr. of args (Ex only) - * (ebp+16) esi pointer to caller's esi value - * (ebp+12) arg ebp value to be set for relay stub - * (ebp+8) func CBClient relay stub address - * (ebp+4) ret addr - * (ebp) ebp - */ static void BuildCallTo32CBClient( int isEx ) { - function_header( isEx ? "CALL32_CBClientEx" : "CALL32_CBClient" ); - - /* Entry code */ - - output_cfi( ".cfi_startproc" ); - output( "\tpushl %%ebp\n" ); - output_cfi( ".cfi_adjust_cfa_offset 4" ); - output_cfi( ".cfi_rel_offset %%ebp,0" ); - output( "\tmovl %%esp,%%ebp\n" ); - output_cfi( ".cfi_def_cfa_register %%ebp" ); - output( "\tpushl %%edi\n" ); - output_cfi( ".cfi_rel_offset %%edi,-4" ); - output( "\tpushl %%esi\n" ); - output_cfi( ".cfi_rel_offset %%esi,-8" ); - output( "\tpushl %%ebx\n" ); - output_cfi( ".cfi_rel_offset %%ebx,-12" ); - - /* Get pointer to temporary area and save the 32-bit stack pointer */ - - output( "\tmovl 16(%%ebp), %%ebx\n" ); - output( "\tleal -8(%%esp), %%eax\n" ); - - if ( !isEx ) - output( "\tmovl %%eax, -8(%%ebx)\n" ); - else - output( "\tmovl %%eax, 12(%%ebx)\n" ); - - /* Set up registers and call CBClient relay stub (simulating a far call) */ - - output( "\tmovl 20(%%ebp), %%esi\n" ); - output( "\tmovl (%%esi), %%esi\n" ); - - output( "\tmovl 8(%%ebp), %%eax\n" ); - output( "\tmovl 12(%%ebp), %%ebp\n" ); - - output( "\tpushl %%cs\n" ); - output( "\tcall *%%eax\n" ); - - /* Return new esi value to caller */ - - output( "\tmovl 32(%%esp), %%edi\n" ); - output( "\tmovl %%esi, (%%edi)\n" ); - - /* Return argument size to caller */ - if ( isEx ) - { - output( "\tmovl 36(%%esp), %%ebx\n" ); - output( "\tmovl %%ebp, (%%ebx)\n" ); - } - - /* Restore registers and return */ - - output( "\tpopl %%ebx\n" ); - output_cfi( ".cfi_same_value %%ebx" ); - output( "\tpopl %%esi\n" ); - output_cfi( ".cfi_same_value %%esi" ); - output( "\tpopl %%edi\n" ); - output_cfi( ".cfi_same_value %%edi" ); - output( "\tpopl %%ebp\n" ); - output_cfi( ".cfi_def_cfa %%esp,4" ); - output_cfi( ".cfi_same_value %%ebp" ); - output( "\tret\n" ); - output_cfi( ".cfi_endproc" ); - output_function_size( isEx ? "CALL32_CBClientEx" : "CALL32_CBClient" ); - /* '16-bit' return stub */ function_header( isEx ? "CALL32_CBClientEx_Ret" : "CALL32_CBClient_Ret" ); -- 2.33.0

1 2

[PATCH] user32: Implement WindowFromPhysicalPoint
by Jinoh Kang 03 Oct '21

03 Oct '21

physical<->logical is an identity mapping in other functions and also >win8, so make it consistent with those. Signed-off-by: Jinoh Kang <jinoh.kang.kr(a)gmail.com> --- .../ext-ms-win-ntuser-uicontext-ext-l1-1-0.spec | 2 +- .../ext-ms-win-ntuser-window-l1-1-1.spec | 2 +- dlls/user32/sysparams.c | 8 ++++++++ dlls/user32/user32.spec | 1 + 4 files changed, 11 insertions(+), 2 deletions(-) diff --git a/dlls/ext-ms-win-ntuser-uicontext-ext-l1-1-0/ext-ms-win-ntuser-uicontext-ext-l1-1-0.spec b/dlls/ext-ms-win-ntuser-uicontext-ext-l1-1-0/ext-ms-win-ntuser-uicontext-ext-l1-1-0.spec index ea155c098d9..acc03f665de 100644 --- a/dlls/ext-ms-win-ntuser-uicontext-ext-l1-1-0/ext-ms-win-ntuser-uicontext-ext-l1-1-0.spec +++ b/dlls/ext-ms-win-ntuser-uicontext-ext-l1-1-0/ext-ms-win-ntuser-uicontext-ext-l1-1-0.spec @@ -51,5 +51,5 @@ @ stub SoundSentry @ stdcall UnhookWindowsHookEx(long) user32.UnhookWindowsHookEx @ stdcall UpdateLayeredWindow(long long ptr ptr long ptr long ptr long) user32.UpdateLayeredWindow -@ stub WindowFromPhysicalPoint +@ stdcall WindowFromPhysicalPoint(int64) user32.WindowFromPhysicalPoint @ stdcall WindowFromPoint(int64) user32.WindowFromPoint diff --git a/dlls/ext-ms-win-ntuser-window-l1-1-1/ext-ms-win-ntuser-window-l1-1-1.spec b/dlls/ext-ms-win-ntuser-window-l1-1-1/ext-ms-win-ntuser-window-l1-1-1.spec index ea155c098d9..acc03f665de 100644 --- a/dlls/ext-ms-win-ntuser-window-l1-1-1/ext-ms-win-ntuser-window-l1-1-1.spec +++ b/dlls/ext-ms-win-ntuser-window-l1-1-1/ext-ms-win-ntuser-window-l1-1-1.spec @@ -51,5 +51,5 @@ @ stub SoundSentry @ stdcall UnhookWindowsHookEx(long) user32.UnhookWindowsHookEx @ stdcall UpdateLayeredWindow(long long ptr ptr long ptr long ptr long) user32.UpdateLayeredWindow -@ stub WindowFromPhysicalPoint +@ stdcall WindowFromPhysicalPoint(int64) user32.WindowFromPhysicalPoint @ stdcall WindowFromPoint(int64) user32.WindowFromPoint diff --git a/dlls/user32/sysparams.c b/dlls/user32/sysparams.c index d6f08654ab0..733acd3d3c6 100644 --- a/dlls/user32/sysparams.c +++ b/dlls/user32/sysparams.c @@ -4506,6 +4506,14 @@ BOOL WINAPI SetPhysicalCursorPos( INT x, INT y ) return SetCursorPos( x, y ); } +/*********************************************************************** + * WindowFromPhysicalPoint (USER32.@) + */ +HWND WINAPI WindowFromPhysicalPoint( POINT pt ) +{ + return WindowFromPoint( pt ); +} + /*********************************************************************** * LogicalToPhysicalPoint (USER32.@) */ diff --git a/dlls/user32/user32.spec b/dlls/user32/user32.spec index 8d3ac70210c..8d5f1bc4145 100644 --- a/dlls/user32/user32.spec +++ b/dlls/user32/user32.spec @@ -820,6 +820,7 @@ # @ stub WinOldAppHackoMatic @ stdcall WindowFromDC(long) @ stdcall WindowFromPoint(int64) +@ stdcall WindowFromPhysicalPoint(int64) # @ stub YieldTask # @ stub _SetProcessDefaultLayout @ stdcall keybd_event(long long long long) -- 2.31.1

1 0

GCC 11
by Eric Pouech 03 Oct '21

03 Oct '21

For sharing : - newest gcc version (11) sets by default of bunch of new tests for code correctness - it spits out around ten thousand lines of warnings (*), making it quite impossible to look into compiler output there are a couple of genuine issues reported by gcc11 - Rémi and I already sent a serie of patches for addressing those - a strange indentation waiting for a volunteer (in dlls/wininet/internet.c:2878, the line lpwhh->ErrorMask = *(ULONG*)lpBuffer; at the end of the if/else chain is misindented) unfortunately, most of the generated warnings are false positives, that need to be fixed (or removerd) before GCC11 becomes mainstream - Fedora 35 with GCC11 is in beta right now, and GA is mid of October - I haven't checked other distros though, but I expect similar evolution in coming weeks/months - in Fedora both gcc and mingw-gcc have been upgraded to version 11 The false positives we need to tackle: A) misleading indentation =================== GCC11 complains about code like: foo; todo_wine ok(tst, ...); bar; (gcc11 warns about 'bar ' not being properly indented wrt todo_wine) there are (as of today) 817 occurrences of those in wine tree (spread across 127 files) and counts for ~90% (*) of the line of warnings generated by GCC only 2 true positives are generated by this GCC warning Remarks: - constructs on a single line don't generate the warning todo_wine ok(tst, ...); // ok no warning, whatever the indentation of the line wrt the rest of the code - having the ok(tst, ...) inside a block after the todo_wine doesn't generate the warning possible solutions: - disable the warning in configure cons: will not trigger on true positive pros: minor impact on code base (either committed and under progress in dev:s trees) - merge the todo_wine and ok() line into a single one pros: keep current indentation, coherent with todo_wine + block cons: harder to read, esp. for todo_wine_if cons: for todo_wine_if, could generate too long lines cons: large impact on code base pros/cons: git blame on ok() line will tell who removed the todo_wine, not who wrote the test - reindent only problematic todo_wine foo; todo_wine ok(tst, ...); bar; ^ without indenting the line ok(tst, ...); ^ without indenting the "todo_wine {\n<...>\n}" nor the "todo_wine ok()" on a single line constructs as they don't generate warnings pros/cons: indentation isn't used for adding comments to the code (that's what comments are for) cons: large impact on code base - reindent all todo_wine same as above, but also reindenting the "todo_wine {\n<...>\n}" and the "todo_wine ok()" on a single line constructs pros/cons: indentation isn't used for adding comments to the code (that's what comments are for) cons: large impact on code base B) uninitialized objects =============== GCC11 warns on code like: void foo(const TYPE* s); ..... TYPE s; foo(&s); because, s isn't initialized, and is passed as const ptr, hence could be read by function foo without being initialized proposal: - depending on context, different fixes are at hand: either initialize s, or when applicable, play with aliasing C) partial structures ============ In a bunch of places, we use a pointer to a structure, where underlying allocated object is smaller than the actual size of the structure. (roughly 400 lines of warnings (*)) It's used: - when dealing with two structures sharing their first fields (like in BITMAP for example) - in tests, to test behavior will various sizes of objects as input... proposal: disable this warning in configure D) sizeof/sizeof ========= GCC11 warns when sizeof(foo)/sizeof(bar) is used and cannot be reduced to computing the number of elements of an array proposal: rewrite as sizeof(foo)/(sizeof(bar)) (note the parenthesis around sizeof(bar), as suggested by GCC) A+ (*) those values are rough estimates, only consider the order of magnitude

3 8

← Newer
1
...
70
71
72
73
74
75
76
77
78
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

wine-devel October 2021