Re: [PATCH v2] Cygwin: signal: Avoid frequent TLS lock/unlock for SIGCONT processing

[Corinna Vinschen]

On Jan 22 07:58, Takashi Yano wrote:
> On Tue, 21 Jan 2025 13:20:46 -0800 (PST)
> Jeremy Drake wrote:
> > dscho hooked me up with a workflow to run Git for Windows' test suite on
> > an msys2-runtime pull request's CI artifacts. With this patch and the
> > three you pushed (reverting the v2 patch we had applied already, and
> > applying the two from the cygwin-3_5-branch), the test suite no longer
> > hangs.
> 
> Thanks for testing! I'm happy to hear that.

Me too.  Please push.

Thanks,
Corinna

Re: [PATCH v3] Cygwin: Minor updates to load average calculations

[Corinna Vinschen]

Jon?  Are yuo going to review this one?

Thanks,
Corinna


On Jan 20 00:18, Mark Geisert wrote:
> Commentary wording now refers to tasks (i.e., threads) rather than
> processes.  This makes it somewhat easier to justify adding two kinds of
> counters together.  After researching what "load average" has meant over
> time, we have what seems like a reasonable implementation, modulo
> Windows differences to Linux.  The best resource I found is:
> https://www.brendangregg.com/blog/2017-08-08/linux-load-averages.html
> 
> At end of load_init(), obtain and discard the first measure of the
> counters to deal with the first call always returning error, no data.
> Follow this with a specific short delay so the next measure actually has
> data to report.
> 
> At least one older version of Windows, i.e. Win10 Pro 21H1, has a different
> name/location for the '% Processor Time' counter and is missing the
> 'Processor Queue Length' counter entirely.  Code is changed to support
> both possible locations of the former and treat the missing latter as always
> reporting 0.0.
> 
> A release note is added for 3.5.6.
> 
> Reported-by: Mark Liam Brown 
> Addresses: https://cygwin.com/pipermail/cygwin/2024-August/256361.html
> Signed-off-by: Mark Geisert 
> Fixes: 4dc982ddf60b (Cygwin: loadavg: improve debugging of load_init)
> 
> ---
>  winsup/cygwin/loadavg.cc| 50 ++---
>  winsup/cygwin/release/3.5.6 |  3 +++
>  2 files changed, 39 insertions(+), 14 deletions(-)
> 
> diff --git a/winsup/cygwin/loadavg.cc b/winsup/cygwin/loadavg.cc
> index 127591a2e..1c5e553a9 100644
> --- a/winsup/cygwin/loadavg.cc
> +++ b/winsup/cygwin/loadavg.cc
> @@ -15,17 +15,23 @@
>  
>A global load average estimate is maintained in shared memory.  Access to 
> that
>is guarded by a mutex.  This estimate is only updated at most every 5 
> seconds.
> +  The updates are done by any/all callers of the getloadavg() syscall.
>  
> -  We attempt to count running and runnable processes, but unlike linux we 
> don't
> -  count processes in uninterruptible sleep (blocked on I/O).
> +  We attempt to count running and runnable tasks (i.e., threads), but unlike
> +  Linux we don't count tasks in uninterruptible sleep (blocked on I/O).  
> There
> +  doesn't seem to be a kernel counter for the latter on Windows.
>  
> -  The number of running processes is estimated as (NumberOfProcessors) * (%
> -  Processor Time).  The number of runnable processes is estimated as
> -  ProcessorQueueLength.
> +  In the following text and code, "PDH" refers to Process Data Helper, a
> +  Windows component that arranges access to kernel counters.
> +
> +  The number of running tasks is estimated as
> +(the NumberOfProcessors counter) * (the % Processor Time counter).
> +  The number of runnable tasks is taken to be the ProcessorQueueLength 
> counter.
>  
>Note that PDH will only return data for '% Processor Time' afer the second
>call to PdhCollectQueryData(), as it's computed over an interval, so the 
> first
> -  attempt to estimate load will fail and 0.0 will be returned.
> +  attempt to estimate load will fail and 0.0 will be returned.  (This 
> nuisance
> +  is now worked-around near the end of load_init() below.)
>  
>We also assume that '% Processor Time' averaged over the interval since the
>last time getloadavg() was called is a good approximation of the 
> instantaneous
> @@ -68,8 +74,19 @@ static bool load_init (void)
>  if (status != STATUS_SUCCESS)
>{
>   debug_printf ("PdhAddEnglishCounterW(time), status %y", status);
> - return false;
> +
> + /* Windows 10 Pro 21H1, and maybe others, use an alternative name */
> +status = PdhAddEnglishCounterW (query,
> +L"\\Processor Information(_Total)\\% Processor Time",
> +0, &counter1);
> +if (status != STATUS_SUCCESS)
> +  {
> +debug_printf ("PdhAddEnglishCounterW(alt time), status %y", 
> status);
> +return false;
> +  }
>}
> +
> +/* Windows 10 Pro 21H1, and maybe others, are missing this counter */
>  status = PdhAddEnglishCounterW (query,
>   L"\\System\\Processor Queue Length",
>   0, &counter2);
> @@ -77,7 +94,7 @@ static bool load_init (void)
>  if (status != STATUS_SUCCESS)
>{
>   debug_printf ("PdhAddEnglishCounterW(queue length), status %y", status);
> - return false;
> + ; /* Ignore missing counter, just use zero in later calculations */
>}
>  
>  mutex = CreateMutex(&sec_all_nih, FALSE, "cyg.loadavg.mutex");
> @@ -87,6 +104,12 @@ static bool load_init (void)
>  }
>  
>  initialized = true;
> +
> +/* Do the first data collection (which always fails) here, rather than in
> +   get_load(). We wait at least one tick afterward so the collection done
> +   in get_load() is guaranteed to have data to work

Re: [PATCH v7 3/5] Cygwin: winsup/doc/posix.xml: SUS V5 POSIX 2024 not implemented new additions

[Corinna Vinschen]

On Jan 17 10:01, Brian Inglis wrote:
> Add unavailable POSIX additions to Not Implemented section.
> 
> Signed-off-by: Brian Inglis 
> ---
>  winsup/doc/posix.xml | 26 --
>  1 file changed, 24 insertions(+), 2 deletions(-)
> 
> diff --git a/winsup/doc/posix.xml b/winsup/doc/posix.xml
> index ac05657d2246..7e66cb8fc1c0 100644
> --- a/winsup/doc/posix.xml
> +++ b/winsup/doc/posix.xml
> @@ -16,6 +16,9 @@ ISO/IEC DIS 9945 Information technology
>  - Issue 8.
>  
>  
> +CMPLX
> +CMPLXF   
> +CMPLXL   
>  FD_CLR
>  FD_ISSET
>  FD_SET
> @@ -554,6 +557,7 @@ ISO/IEC DIS 9945 Information technology
>  jn
>  jrand48
>  kill
> +kill_dependency  
>  killpg
>  l64a
>  labs
> @@ -1466,6 +1470,8 @@ ISO/IEC DIS 9945 Information technology
>  mempcpy
>  memrchr
>  mkostemps
> +posix_spawn_file_actions_addchdir_np
> +posix_spawn_file_actions_addfchdir_np
>  pow10
>  pow10f
>  pow10l

These first three hunks belong into patch #1 of the set, don't they?

> @@ -1677,9 +1683,14 @@ ISO/IEC DIS 9945 Information technology
>  
>  
>  
> -NOT implemented system interfaces from the 
> Single Unix Specification, Volume 7:
> +NOT implemented system interfaces from the 
> Single UNIX® Specification Version 5:
>  
>  
> +_Fork
> +dcgettext_l  
> +dcngettext_l 
> +dgettext_l   
> +dngettext_l  
>  endnetent
>  fattach
>  fmtmsg
> @@ -1691,17 +1702,28 @@ ISO/IEC DIS 9945 Information technology
>  getnetbyname
>  getnetent
>  getpmsg
> +getresgid
> +getresuid
> +gettext_l
>  isastream
>  mlockall
>  munlockall
> +ngettext_l   
> +posix_close  
> +posix_devctl (prototyped in cygwin-devel "devctl.h" header)

Hmm, I guess we should really keep this hint.

Note to self: However, given this is POSIX-1.2024 now, I think the
header in newlib is now incorrect.  We should rectify this.


Thanks,
Corinna

Re: [PATCH v7 4/5] Cygwin: winsup/doc/posix.xml: SUS V5 POSIX 2024 move or remove dropped entries

[Corinna Vinschen]

On Jan 17 10:01, Brian Inglis wrote:
> Move entries no longer in POSIX from the SUS/POSIX section to
> Deprecated Interfaces section and mark with (SUSv4).
> Remove entries no longer in POSIX from the NOT Implemented section.

This looks good, but I just realized that a bunch of functions are
missing.

> -Other UNIX system interfaces, not in 
> POSIX.1-2008 or deprecated:
> +Other UNIX® system interfaces, not in 
> POSIX.1-2024, or deprecated:

When I introduced the ACL functions from the abandoned POSIX.1e draft,
I missed to add them to the docs.

Well, fortunately I'm now noticing this a mere 8 years later... *facepalm*

Sigh.  I'll create a patch to add them on top of your patches later on.


Thanks,
Corinna

Re: [PATCH v7 1/5] Cygwin: winsup/doc/posix.xml: SUS V5 POSIX 2024 TOG Issue 8 ISO 9945 move new

[Corinna Vinschen]

Hi Brian,

your patch doesn't apply.  Apparently you didn't rebase the patch to the
latest state in the main branch, because your patch contains the changes
from commit 0813644661e3, which I wrote I'll fix separately in
https://sourceware.org/pipermail/cygwin-patches/2025q1/013265.html


Thanks,
Corinna


On Jan 17 10:01, Brian Inglis wrote:
> Update anchor id and description to current version, year, issue, etc.
> Move new POSIX entries in other sections to the SUS/POSIX section.
> 
> Signed-off-by: Brian Inglis 
> ---
>  winsup/doc/posix.xml | 145 ++-
>  1 file changed, 75 insertions(+), 70 deletions(-)
> 
> diff --git a/winsup/doc/posix.xml b/winsup/doc/posix.xml
> index 2782beb86547..9a8890936875 100644
> --- a/winsup/doc/posix.xml
> +++ b/winsup/doc/posix.xml
> @@ -5,10 +5,15 @@
>  http://www.w3.org/2001/XInclude";>
>  Compatibility
>  
> -System interfaces compatible with the Single 
> Unix Specification, Version 7:
> +System interfaces compatible with the Single 
> UNIX® Specification Version 5:
>  
> -Note that the core of the Single Unix Specification, Version 7 is
> -also IEEE Std 1003.1-2017 (POSIX.1-2017).
> +Note that the core of the Single UNIX® Specification Version 5 is
> +POSIX®.1-2024 also simultaneously IEEE Std 1003.1™-2024 Edition,
> +The Open Group Base Specifications Issue 8
> +(see https://pubs.opengroup.org/onlinepubs/9799919799/), and
> +ISO/IEC DIS 9945 Information technology
> +- Portable Operating System Interface (POSIX®) base specifications
> +- Issue 8.
>  
>  
>  FD_CLR
> @@ -25,6 +30,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  abort
>  abs
>  accept
> +accept4
>  access
>  acos
>  acosf
> @@ -40,6 +46,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  aio_suspend
>  aio_write
>  alarm
> +aligned_alloc
>  alphasort
>  asctime
>  asctime_r
> @@ -49,6 +56,9 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  asinhf
>  asinhl
>  asinl
> +asprintf
> +assert
> +at_quick_exit
>  atan
>  atan2
>  atan2f
> @@ -68,6 +78,8 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  bind
>  bsearch
>  btowc
> +c16rtomb
> +c32rtomb
>  cabs
>  cabsf
>  cabsl
> @@ -77,6 +89,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  cacoshf
>  cacoshl
>  cacosl
> +call_once
>  calloc
>  carg
>  cargf
> @@ -134,6 +147,12 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  close
>  closedir
>  closelog
> +cnd_broadcast
> +cnd_destroy
> +cnd_init
> +cnd_signal
> +cnd_timedwait
> +cnd_wait
>  confstr
>  conj
>  conjf
> @@ -158,7 +177,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  crealf
>  creall
>  creat
> -crypt(available in external "crypt" library)
> +crypt(available in external "libcrypt" library)
>  csin
>  csinf
>  csinh
> @@ -191,6 +210,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  dirfd
>  dirname
>  div
> +dladdr   (see chapter 
> "Implementation Notes")
>  dlclose
>  dlerror
>  dlopen
> @@ -199,8 +219,9 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  drand48
>  dup
>  dup2
> +dup3
>  duplocale
> -encrypt  (available in external "crypt" library)
> +encrypt  (available in external "libcrypt" library)
>  endgrent
>  endhostent
>  endprotoent
> @@ -265,6 +286,8 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  fexecve
>  fflush
>  ffs
> +ffsl
> +ffsll
>  fgetc
>  fgetpos
>  fgets
> @@ -541,6 +564,8 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  malloc
>  mblen
>  mbrlen
> +mbrtoc16
> +mbrtoc32
>  mbrtowc
>  mbsinit
>  mbsnrtowcs
> @@ -551,6 +576,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  memchr
>  memcmp
>  memcpy
> +memmem
>  memmove
>  memset
>  mkdir
> @@ -560,6 +586,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  mkfifoat
>  mknod
>  mknodat
> +mkostemp
>  mkstemp
>  mktime
>  mlock
> @@ -584,6 +611,12 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  msgrcv   (see chapter 
> "Implementation Notes")
>  msgsnd   (see chapter 
> "Implementation Notes")
>  msync
> +mtx_destroy
> +mtx_init
> +mtx_lock
> +mtx_timedlock
> +mtx_trylock
> +mtx_unlock
>  munlock
>  munmap
>  nan
> @@ -622,6 +655,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  pclose
>  perror
>  pipe
> +pipe2
>  poll
>  popen
>  posix_fadvise
> @@ -630,8 +664,10 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
>  posix_memalign
>  posix_openpt
>  posix_spawn
> +posix_spawn_file_actions_addchdir
>  posix_spawn_file_actions_addclose
>  posix_spawn_file_actions_adddup2

Re: [PATCH v3] Cygwin: Minor updates to load average calculations

[Jon Turney]

On 22/01/2025 10:13, Corinna Vinschen wrote:

Jon?  Are yuo going to review this one?


This looks fine to me. No notes. Please apply.

Thanks Mark!

(and a particular thank you for fixing the process vs. thread vagueness 
I'd made in the comments)

Re: [PATCH v7 3/5] Cygwin: winsup/doc/posix.xml: SUS V5 POSIX 2024 not implemented new additions

[Corinna Vinschen]

On Jan 22 12:04, Corinna Vinschen wrote:
> On Jan 17 10:01, Brian Inglis wrote:
> > Add unavailable POSIX additions to Not Implemented section.
> > 
> > Signed-off-by: Brian Inglis 
> > ---
> >  winsup/doc/posix.xml | 26 --
> >  1 file changed, 24 insertions(+), 2 deletions(-)
> > 
> > diff --git a/winsup/doc/posix.xml b/winsup/doc/posix.xml
> > index ac05657d2246..7e66cb8fc1c0 100644
> > --- a/winsup/doc/posix.xml
> > +++ b/winsup/doc/posix.xml
> > @@ -16,6 +16,9 @@ ISO/IEC DIS 9945 Information technology
> >  - Issue 8.
> >  
> >  
> > +CMPLX  
> > +CMPLXF 
> > +CMPLXL 
> >  FD_CLR
> >  FD_ISSET
> >  FD_SET
> > @@ -554,6 +557,7 @@ ISO/IEC DIS 9945 Information technology
> >  jn
> >  jrand48
> >  kill
> > +kill_dependency
> >  killpg
> >  l64a
> >  labs
> > @@ -1466,6 +1470,8 @@ ISO/IEC DIS 9945 Information technology
> >  mempcpy
> >  memrchr
> >  mkostemps
> > +posix_spawn_file_actions_addchdir_np
> > +posix_spawn_file_actions_addfchdir_np
> >  pow10
> >  pow10f
> >  pow10l
> 
> These first three hunks belong into patch #1 of the set, don't they?

No, sorry, into patch #2, of course.


Corinna

Re: [PATCH] Cygwin: mmap: use 64K pages for bookkeeping

[Corinna Vinschen]

On Jan 20 10:36, Ken Brown wrote:
> On 1/20/2025 6:49 AM, Corinna Vinschen wrote:
> > Nice idea, but this may not do what is expected if the mapping is an
> > anonymous mapping, leaving the protection or mapping of trailing pages
> > in a wrong state, isn't it?
> > 
> > Can we easily make sure the type of mapping (file vs anon) is known
> > at the time of rounding, so the rounding is performed differently?
> I hadn't thought of that.  Actually, I *think* the record length is already
> a multiple of 64K for an anonymous mapping.

You're right, of course.  For an anonymous mapping, mmap() forces len to
be a multiple of wincap.allocation_granularity() anyway.

So yeah, looks like your previous patch does the trick.  Feel free to
push it, but I would love to see a comment in mmap_record::match() to
explain why we do it this way.


Thanks,
Corinna

Re: [PATCH v7 1/5] Cygwin: winsup/doc/posix.xml: SUS V5 POSIX 2024 TOG Issue 8 ISO 9945 move new

[Brian Inglis]

Hi Corinna,

No - never occurred to me other changes would be made here - I have enough 
problems tracking rebasing from *MY OWN* upstream!


On 2025-01-22 03:51, Corinna Vinschen wrote:

Hi Brian,

your patch doesn't apply.  Apparently you didn't rebase the patch to the
latest state in the main branch, because your patch contains the changes
from commit 0813644661e3, which I wrote I'll fix separately in
https://sourceware.org/pipermail/cygwin-patches/2025q1/013265.html


Thanks,
Corinna


On Jan 17 10:01, Brian Inglis wrote:

Update anchor id and description to current version, year, issue, etc.
Move new POSIX entries in other sections to the SUS/POSIX section.

Signed-off-by: Brian Inglis 
---
  winsup/doc/posix.xml | 145 ++-
  1 file changed, 75 insertions(+), 70 deletions(-)

diff --git a/winsup/doc/posix.xml b/winsup/doc/posix.xml
index 2782beb86547..9a8890936875 100644
--- a/winsup/doc/posix.xml
+++ b/winsup/doc/posix.xml
@@ -5,10 +5,15 @@
  http://www.w3.org/2001/XInclude";>
  Compatibility
  
-System interfaces compatible with the Single Unix Specification, Version 7:

+System interfaces compatible with the Single UNIX® 
Specification Version 5:
  
-Note that the core of the Single Unix Specification, Version 7 is

-also IEEE Std 1003.1-2017 (POSIX.1-2017).
+Note that the core of the Single UNIX® Specification Version 5 is
+POSIX®.1-2024 also simultaneously IEEE Std 1003.1™-2024 Edition,
+The Open Group Base Specifications Issue 8
+(see https://pubs.opengroup.org/onlinepubs/9799919799/), and
+ISO/IEC DIS 9945 Information technology
+- Portable Operating System Interface (POSIX®) base specifications
+- Issue 8.
  
  

  FD_CLR
@@ -25,6 +30,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  abort
  abs
  accept
+accept4
  access
  acos
  acosf
@@ -40,6 +46,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  aio_suspend
  aio_write
  alarm
+aligned_alloc
  alphasort
  asctime
  asctime_r
@@ -49,6 +56,9 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  asinhf
  asinhl
  asinl
+asprintf
+assert
+at_quick_exit
  atan
  atan2
  atan2f
@@ -68,6 +78,8 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  bind
  bsearch
  btowc
+c16rtomb
+c32rtomb
  cabs
  cabsf
  cabsl
@@ -77,6 +89,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  cacoshf
  cacoshl
  cacosl
+call_once
  calloc
  carg
  cargf
@@ -134,6 +147,12 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  close
  closedir
  closelog
+cnd_broadcast
+cnd_destroy
+cnd_init
+cnd_signal
+cnd_timedwait
+cnd_wait
  confstr
  conj
  conjf
@@ -158,7 +177,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  crealf
  creall
  creat
-crypt  (available in external "crypt" library)
+crypt  (available in external "libcrypt" library)
  csin
  csinf
  csinh
@@ -191,6 +210,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  dirfd
  dirname
  div
+dladdr (see chapter "Implementation 
Notes")
  dlclose
  dlerror
  dlopen
@@ -199,8 +219,9 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  drand48
  dup
  dup2
+dup3
  duplocale
-encrypt(available in external "crypt" library)
+encrypt(available in external "libcrypt" library)
  endgrent
  endhostent
  endprotoent
@@ -265,6 +286,8 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  fexecve
  fflush
  ffs
+ffsl
+ffsll
  fgetc
  fgetpos
  fgets
@@ -541,6 +564,8 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  malloc
  mblen
  mbrlen
+mbrtoc16
+mbrtoc32
  mbrtowc
  mbsinit
  mbsnrtowcs
@@ -551,6 +576,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  memchr
  memcmp
  memcpy
+memmem
  memmove
  memset
  mkdir
@@ -560,6 +586,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  mkfifoat
  mknod
  mknodat
+mkostemp
  mkstemp
  mktime
  mlock
@@ -584,6 +611,12 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  msgrcv(see chapter "Implementation 
Notes")
  msgsnd(see chapter "Implementation 
Notes")
  msync
+mtx_destroy
+mtx_init
+mtx_lock
+mtx_timedlock
+mtx_trylock
+mtx_unlock
  munlock
  munmap
  nan
@@ -622,6 +655,7 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  pclose
  perror
  pipe
+pipe2
  poll
  popen
  posix_fadvise
@@ -630,8 +664,10 @@ also IEEE Std 1003.1-2017 (POSIX.1-2017).
  posix_memalign
  posix_openpt
  posix_spawn
+posix_spawn_file_actions_addchdir
  posix_spawn_file_actions_addclose
  posix_spawn_file_actions_adddup2
+posix_spawn_file_actions_addfchdir
  posix_sp

Re: [PATCH v7 3/5] Cygwin: winsup/doc/posix.xml: SUS V5 POSIX 2024 not implemented new additions

[Jon Turney]

On 22/01/2025 17:36, Brian Inglis wrote:
[...]

"Please note some changes are displaced due to rebase conflicts."

which only appear after commit --amend/rebase --continue!

As I said before, I am now noticing the limitations of interactive 
rebasing and cherry-picking, sometimes skipping patches or hunks, 
especially after conflicts!


If you have any advice for avoiding those conflicts, please hit me with 
your clue stick! ;^>
 If a 'git rebase' reports merge conflicts, I would suggest using 'git 
mergetool' to invoke your favorite 3-way graphical merge tool (I like 
kdiff3, but you might have a different preference).

Re: [PATCH v7 3/5] Cygwin: winsup/doc/posix.xml: SUS V5 POSIX 2024 not implemented new additions

[Brian Inglis]

On 2025-01-22 04:08, Corinna Vinschen wrote:

On Jan 22 12:04, Corinna Vinschen wrote:

On Jan 17 10:01, Brian Inglis wrote:

Add unavailable POSIX additions to Not Implemented section.

Signed-off-by: Brian Inglis 
---
  winsup/doc/posix.xml | 26 --
  1 file changed, 24 insertions(+), 2 deletions(-)

diff --git a/winsup/doc/posix.xml b/winsup/doc/posix.xml
index ac05657d2246..7e66cb8fc1c0 100644
--- a/winsup/doc/posix.xml
+++ b/winsup/doc/posix.xml
@@ -16,6 +16,9 @@ ISO/IEC DIS 9945 Information technology
  - Issue 8.
  
  

+CMPLX  
+CMPLXF 
+CMPLXL 
  FD_CLR
  FD_ISSET
  FD_SET
@@ -554,6 +557,7 @@ ISO/IEC DIS 9945 Information technology
  jn
  jrand48
  kill
+kill_dependency
  killpg
  l64a
  labs
@@ -1466,6 +1470,8 @@ ISO/IEC DIS 9945 Information technology
  mempcpy
  memrchr
  mkostemps
+posix_spawn_file_actions_addchdir_np
+posix_spawn_file_actions_addfchdir_np
  pow10
  pow10f
  pow10l


These first three hunks belong into patch #1 of the set, don't they?


No, sorry, into patch #2, of course.


Hi Corinna,

Re previous comment: I thought I was to apply those lib+crypt changes!

As I said in 0/5:

"Please note some changes are displaced due to rebase conflicts."

which only appear after commit --amend/rebase --continue!

As I said before, I am now noticing the limitations of interactive rebasing and 
cherry-picking, sometimes skipping patches or hunks, especially after conflicts!


If you have any advice for avoiding those conflicts, please hit me with your 
clue stick! ;^>


I will return with v8 - caffeine+sugar-powered rather than veggie-powered!

--
Take care. Thanks, Brian Inglis  Calgary, Alberta, Canada

La perfection est atteinte   Perfection is achieved
non pas lorsqu'il n'y a plus rien à ajouter  not when there is no more to add
mais lorsqu'il n'y a plus rien à retrancher  but when there is no more to cut
-- Antoine de Saint-Exupéry

[PATCH 1/1] replace undocumented Nt* calls with documented win32 apis

[Sebastian Hernandez via Cygwin-patches]

Hi all,

Attaching the patch from the separate discussion in cygwin-developers - Re: 
[Proposed changes/Incoming patch] Undocumented call to 
ntdll.dll!NtAssignProcessToJobObject in msys2.0.dll / cygwin - included in 
mingit/Git for Windows.

The following patch includes the replacement of the following undocumented APIs 
and their documented counterparts.

  *   NtOpenJobObject - 
OpenJobObjectA
  *   NtCreateJobObject - 
CreateJobObjectA
  *   NtQueryInformationJobObject - 
QueryInformationJobObject
  *   NtSetInformationJobObject - 
SetInformationJobObject
  *
NTAssignProcessToJobObject - 
AssignProcessToJobObject

Please let us know if you have any further questions or feedback regarding the 
changes.

Whitney & Sebastian

P.S.
Unfortunately the use of git send-email was not possible due to firewall issues.



>From 2cf2176ee8dd35bc4fca8d420533dfb385241d24 Mon Sep 17 00:00:00 2001
From: "Sebastian Hernandez (from Dev Box)" 
Date: Wed, 22 Jan 2025 10:20:59 -0800
Subject: [PATCH 1/1] replace undocumented Nt* calls with documented win32 apis

---
 winsup/cygwin/local_includes/ntdll.h |  6 --
 winsup/cygwin/resource.cc| 93 ++--
 2 files changed, 61 insertions(+), 38 deletions(-)

diff --git a/winsup/cygwin/local_includes/ntdll.h 
b/winsup/cygwin/local_includes/ntdll.h
index 4497fe53f..9f07b37e4 100644
--- a/winsup/cygwin/local_includes/ntdll.h
+++ b/winsup/cygwin/local_includes/ntdll.h
@@ -1451,7 +1451,6 @@ extern "C"
   NTSTATUS NtAdjustPrivilegesToken (HANDLE, BOOLEAN, PTOKEN_PRIVILEGES, ULONG,
PTOKEN_PRIVILEGES, PULONG);
   NTSTATUS NtAllocateLocallyUniqueId (PLUID);
-  NTSTATUS NtAssignProcessToJobObject (HANDLE, HANDLE);
   NTSTATUS NtCancelTimer (HANDLE, PBOOLEAN);
   NTSTATUS NtClose (HANDLE);
   NTSTATUS NtCommitTransaction (HANDLE, BOOLEAN);
@@ -1461,7 +1460,6 @@ extern "C"
   NTSTATUS NtCreateFile (PHANDLE, ACCESS_MASK, POBJECT_ATTRIBUTES,
  PIO_STATUS_BLOCK, PLARGE_INTEGER, ULONG, ULONG, ULONG,
  ULONG, PVOID, ULONG);
-  NTSTATUS NtCreateJobObject (PHANDLE, ACCESS_MASK, POBJECT_ATTRIBUTES);
   NTSTATUS NtCreateKey (PHANDLE, ACCESS_MASK, POBJECT_ATTRIBUTES, ULONG,
 PUNICODE_STRING, ULONG, PULONG);
   NTSTATUS NtCreateMutant (PHANDLE, ACCESS_MASK, POBJECT_ATTRIBUTES, BOOLEAN);
@@ -1498,7 +1496,6 @@ extern "C"
   NTSTATUS NtOpenEvent (PHANDLE, ACCESS_MASK, POBJECT_ATTRIBUTES);
   NTSTATUS NtOpenFile (PHANDLE, ACCESS_MASK, POBJECT_ATTRIBUTES,
 PIO_STATUS_BLOCK, ULONG, ULONG);
-  NTSTATUS NtOpenJobObject (PHANDLE, ACCESS_MASK, POBJECT_ATTRIBUTES);
   NTSTATUS NtOpenKey (PHANDLE, ACCESS_MASK, POBJECT_ATTRIBUTES);
   NTSTATUS NtOpenMutant (PHANDLE, ACCESS_MASK, POBJECT_ATTRIBUTES);
   NTSTATUS NtOpenProcessToken (HANDLE, ACCESS_MASK, PHANDLE);
@@ -1520,8 +1517,6 @@ extern "C"
   NTSTATUS NtQueryEvent (HANDLE, EVENT_INFORMATION_CLASS, PVOID, ULONG, 
PULONG);
   NTSTATUS NtQueryInformationFile (HANDLE, PIO_STATUS_BLOCK, PVOID, ULONG,
   FILE_INFORMATION_CLASS);
-  NTSTATUS NtQueryInformationJobObject (HANDLE, JOBOBJECTINFOCLASS, PVOID,
-  ULONG, PULONG);
   NTSTATUS NtQueryInformationProcess (HANDLE, PROCESSINFOCLASS, PVOID, ULONG,
  PULONG);
   NTSTATUS NtQueryInformationThread (HANDLE, THREADINFOCLASS, PVOID, ULONG,
@@ -1555,7 +1550,6 @@ extern "C"
   NTSTATUS NtSetEvent (HANDLE, PULONG);
   NTSTATUS NtSetInformationFile (HANDLE, PIO_STATUS_BLOCK, PVOID, ULONG,
 FILE_INFORMATION_CLASS);
-  NTSTATUS NtSetInformationJobObject (HANDLE, JOBOBJECTINFOCLASS, PVOID, 
ULONG);
   NTSTATUS NtSetInformationThread (HANDLE, THREADINFOCLASS, PVOID, ULONG);
   NTSTATUS NtSetInformationToken (HANDLE, TOKEN_INFORMATION_CLASS, PVOID,
  ULONG);
diff --git a/winsup/cygwin/resource.cc b/winsup/cygwin/resource.cc
index 5ec436c2c..64c26c14d 100644
--- a/winsup/cygwin/resource.cc
+++ b/winsup/cygwin/resource.cc
@@ -177,27 +177,43 @@ job_shared_name (PWCHAR buf, LONG num)
 static void
 __get_rlimit_as (struct rlimit *rlp)
 {
-  UNICODE_STRING uname;
   WCHAR jobname[32];
-  OBJECT_ATTRIBUTES attr;
+  char jobnameA[32];
   HANDLE job = NULL;
-  NTSTATUS status;
+  BOOL result;
+  DWORD winError;
   JOBOBJECT_EXTENDED_LIMIT_INFORMATION jobinfo;

   if (cygheap->rlim_as_id)
 {
-  RtlInitUnicodeString (&uname,
-job_shared_name (jobname,
-   cygheap->rlim_as_id));
-  InitializeObjectAttributes (&attr, &uname, 0,
- get_session_parent_dir (), NULL);
+  /* Get the wide

Re: [PATCH v7 4/5] Cygwin: winsup/doc/posix.xml: SUS V5 POSIX 2024 move or remove dropped entries

[Jon Turney]

On 22/01/2025 11:23, Corinna Vinschen wrote:

On Jan 17 10:01, Brian Inglis wrote:

Move entries no longer in POSIX from the SUS/POSIX section to
Deprecated Interfaces section and mark with (SUSv4).
Remove entries no longer in POSIX from the NOT Implemented section.


This looks good, but I just realized that a bunch of functions are
missing.


-Other UNIX system interfaces, not in POSIX.1-2008 or 
deprecated:
+Other UNIX® system interfaces, not in POSIX.1-2024, or 
deprecated:


When I introduced the ACL functions from the abandoned POSIX.1e draft,
I missed to add them to the docs.

Well, fortunately I'm now noticing this a mere 8 years later... *facepalm*

Sigh.  I'll create a patch to add them on top of your patches later on.

I happen to notice the other day that we don't mention eaccess() (which 
the glibc extension euidaccess() is a synonym for) here, which I think 
belongs in this section.