[PATCH v3 0/3] use wincap in format_proc_cpuinfo for user_shstk

[PATCH v3 0/3] use wincap in format_proc_cpuinfo for user_shstk

[Brian Inglis]

Fixes: 41fdb869f998 "fhandler/proc.cc(format_proc_cpuinfo): Add Linux 6.3 cpuinfo"

In test for for AMD/Intel Control flow Enforcement Technology user mode
shadow stack support replace Windows version tests with test of wincap
member addition has_user_shstk with Windows version dependent value

Signed-off-by: Brian Inglis <Brian.Inglis@Shaw.ca>

Brian Inglis (3):
  wincap.h: add wincap member has_user_shstk
  wincap.cc: set wincap member has_user_shstk true for 2004+
  fhandler/proc.cc: use wincap.has_user_shstk

 winsup/cygwin/fhandler/proc.cc        |  8 ++++----
 winsup/cygwin/local_includes/wincap.h |  2 ++
 winsup/cygwin/wincap.cc               | 10 ++++++++++
 3 files changed, 16 insertions(+), 4 deletions(-)

-- 
2.39.0

[PATCH v3 1/3] wincap.h: add wincap member has_user_shstk

[Brian Inglis]

Signed-off-by: Brian Inglis <Brian.Inglis@Shaw.ca>
---
 winsup/cygwin/local_includes/wincap.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/winsup/cygwin/local_includes/wincap.h b/winsup/cygwin/local_includes/wincap.h
index 29a7a63de7f6..c14872787ca2 100644
--- a/winsup/cygwin/local_includes/wincap.h
+++ b/winsup/cygwin/local_includes/wincap.h
@@ -32,6 +32,7 @@ struct wincaps
     unsigned has_linux_tcp_keepalive_sockopts			: 1;
     unsigned has_tcp_maxrtms					: 1;
     unsigned has_con_broken_tabs				: 1;
+    unsigned has_user_shstk					: 1;
   };
 };
 
@@ -84,6 +85,7 @@ public:
   bool	IMPLEMENT (has_linux_tcp_keepalive_sockopts)
   bool	IMPLEMENT (has_tcp_maxrtms)
   bool	IMPLEMENT (has_con_broken_tabs)
+  bool	IMPLEMENT (has_user_shstk)
 
   void disable_case_sensitive_dirs ()
   {
-- 
2.39.0

[PATCH v3 2/3] wincap.cc: set wincap member has_user_shstk true for 2004+

[Brian Inglis]

Signed-off-by: Brian Inglis <Brian.Inglis@Shaw.ca>
---
 winsup/cygwin/wincap.cc | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/winsup/cygwin/wincap.cc b/winsup/cygwin/wincap.cc
index 91d5d9df8889..30d9c14e8d3b 100644
--- a/winsup/cygwin/wincap.cc
+++ b/winsup/cygwin/wincap.cc
@@ -31,6 +31,7 @@ static const wincaps wincap_8_1 = {
     has_linux_tcp_keepalive_sockopts:false,
     has_tcp_maxrtms:false,
     has_con_broken_tabs:false,
+    has_user_shstk:false,
   },
 };
 
@@ -52,6 +53,7 @@ static const wincaps  wincap_10_1507 = {
     has_linux_tcp_keepalive_sockopts:false,
     has_tcp_maxrtms:false,
     has_con_broken_tabs:false,
+    has_user_shstk:false,
   },
 };
 
@@ -73,6 +75,7 @@ static const wincaps  wincap_10_1607 = {
     has_linux_tcp_keepalive_sockopts:false,
     has_tcp_maxrtms:true,
     has_con_broken_tabs:false,
+    has_user_shstk:false,
   },
 };
 
@@ -94,6 +97,7 @@ static const wincaps wincap_10_1703 = {
     has_linux_tcp_keepalive_sockopts:false,
     has_tcp_maxrtms:true,
     has_con_broken_tabs:true,
+    has_user_shstk:false,
   },
 };
 
@@ -115,6 +119,7 @@ static const wincaps wincap_10_1709 = {
     has_linux_tcp_keepalive_sockopts:true,
     has_tcp_maxrtms:true,
     has_con_broken_tabs:true,
+    has_user_shstk:false,
   },
 };
 
@@ -136,6 +141,7 @@ static const wincaps wincap_10_1803 = {
     has_linux_tcp_keepalive_sockopts:true,
     has_tcp_maxrtms:true,
     has_con_broken_tabs:true,
+    has_user_shstk:false,
   },
 };
 
@@ -157,6 +163,7 @@ static const wincaps wincap_10_1809 = {
     has_linux_tcp_keepalive_sockopts:true,
     has_tcp_maxrtms:true,
     has_con_broken_tabs:true,
+    has_user_shstk:false,
   },
 };
 
@@ -178,6 +185,7 @@ static const wincaps wincap_10_1903 = {
     has_linux_tcp_keepalive_sockopts:true,
     has_tcp_maxrtms:true,
     has_con_broken_tabs:true,
+    has_user_shstk:false,
   },
 };
 
@@ -199,6 +207,7 @@ static const wincaps wincap_10_2004 = {
     has_linux_tcp_keepalive_sockopts:true,
     has_tcp_maxrtms:true,
     has_con_broken_tabs:true,
+    has_user_shstk:true,
   },
 };
 
@@ -220,6 +229,7 @@ static const wincaps wincap_11 = {
     has_linux_tcp_keepalive_sockopts:true,
     has_tcp_maxrtms:true,
     has_con_broken_tabs:false,
+    has_user_shstk:true,
   },
 };
 
-- 
2.39.0

[PATCH v3 3/3] fhandler/proc.cc: use wincap.has_user_shstk

[Brian Inglis]

Signed-off-by: Brian Inglis <Brian.Inglis@Shaw.ca>
---
 winsup/cygwin/fhandler/proc.cc | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/winsup/cygwin/fhandler/proc.cc b/winsup/cygwin/fhandler/proc.cc
index 3c79762e0fbd..cbc49a12a417 100644
--- a/winsup/cygwin/fhandler/proc.cc
+++ b/winsup/cygwin/fhandler/proc.cc
@@ -1486,12 +1486,12 @@ format_proc_cpuinfo (void *, char *&destbuf)
 
 /*	  ftcprint (features1,  6, "split_lock_detect");*//* MSR_TEST_CTRL split lock */
 
-      /* cpuid 0x00000007 ecx & Windows [20]20H1/[20]2004+ */
-      if (maxf >= 0x00000007 && wincap.osname () >= "10.0"
-					 && wincap.build_number () >= 19041)
+      /* Windows [20]20H1/[20]2004/19041 user shadow stack */
+      if (maxf >= 0x00000007 && wincap.has_user_shstk ())
         {
+	  /* cpuid 0x00000007 ecx CET shadow stack */
 	  cpuid (&unused, &unused, &features1, &unused, 0x00000007, 0);
-	  ftcprint (features1,  7, "user_shstk");	/* "user shadow stack" */
+	  ftcprint (features1,  7, "user_shstk");	/* user shadow stack */
 	}
 
       /* cpuid 0x00000007:1 eax */
-- 
2.39.0

Re: [PATCH v3 0/3] use wincap in format_proc_cpuinfo for user_shstk

[Corinna Vinschen]

Hi Brian,

On Jun 16 11:17, Brian Inglis wrote:
> Fixes: 41fdb869f998 "fhandler/proc.cc(format_proc_cpuinfo): Add Linux 6.3 cpuinfo"
> 
> In test for for AMD/Intel Control flow Enforcement Technology user mode
> shadow stack support replace Windows version tests with test of wincap
> member addition has_user_shstk with Windows version dependent value
> 
> Signed-off-by: Brian Inglis <Brian.Inglis@Shaw.ca>
> 
> Brian Inglis (3):
>   wincap.h: add wincap member has_user_shstk
>   wincap.cc: set wincap member has_user_shstk true for 2004+
>   fhandler/proc.cc: use wincap.has_user_shstk
> 
>  winsup/cygwin/fhandler/proc.cc        |  8 ++++----
>  winsup/cygwin/local_includes/wincap.h |  2 ++
>  winsup/cygwin/wincap.cc               | 10 ++++++++++
>  3 files changed, 16 insertions(+), 4 deletions(-)
> 
> -- 
> 2.39.0

Is that actually the final version?  It's still missing the commit
message text explaining things and the "Fixes" line...


Thanks,
Corinna

Re: [PATCH v3 0/3] use wincap in format_proc_cpuinfo for user_shstk

[Brian Inglis]

On 2023-06-16 13:51, Corinna Vinschen wrote:
> Hi Brian,
> 
> On Jun 16 11:17, Brian Inglis wrote:

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
>> Fixes: 41fdb869f998 "fhandler/proc.cc(format_proc_cpuinfo): Add Linux 6.3 cpuinfo"
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
>> In test for for AMD/Intel Control flow Enforcement Technology user mode
>> shadow stack support replace Windows version tests with test of wincap
>> member addition has_user_shstk with Windows version dependent value
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

> Is that actually the final version?  It's still missing the commit
> message text explaining things and the "Fixes" line...
Hi Corinna,

Is more required above?

-- 
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 retirer     but when there is no more to cut
                                 -- Antoine de Saint-Exupéry

Re: [PATCH v3 0/3] use wincap in format_proc_cpuinfo for user_shstk

[Corinna Vinschen]

On Jun 16 15:26, Brian Inglis wrote:
> On 2023-06-16 13:51, Corinna Vinschen wrote:
> > Hi Brian,
> > 
> > On Jun 16 11:17, Brian Inglis wrote:
> 
> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
> > > Fixes: 41fdb869f998 "fhandler/proc.cc(format_proc_cpuinfo): Add Linux 6.3 cpuinfo"
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
> > > In test for for AMD/Intel Control flow Enforcement Technology user mode
> > > shadow stack support replace Windows version tests with test of wincap
> > > member addition has_user_shstk with Windows version dependent value
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> > Is that actually the final version?  It's still missing the commit
> > message text explaining things and the "Fixes" line...
> Hi Corinna,
> 
> Is more required above?

No, it's fine, albeit "Fixes:" is supposed to be kind of like a footer,
just where the "Signed-off-by:" is, too.

But it's still only in the cover letter.  As I wrote, it needs to go
into the actual patch, otherwise all the nice info doesn't make it into
the git repo.


Thanks,
Corinna

Re: [PATCH v3 0/3] use wincap in format_proc_cpuinfo for user_shstk

[Brian Inglis]

On 2023-06-19 02:55, Corinna Vinschen wrote:
> On Jun 16 15:26, Brian Inglis wrote:
>> On 2023-06-16 13:51, Corinna Vinschen wrote:
>>> Hi Brian,
>>>
>>> On Jun 16 11:17, Brian Inglis wrote:
>>
>> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
>>>> Fixes: 41fdb869f998 "fhandler/proc.cc(format_proc_cpuinfo): Add Linux 6.3 cpuinfo"
>> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>>
>> vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv
>>>> In test for for AMD/Intel Control flow Enforcement Technology user mode
>>>> shadow stack support replace Windows version tests with test of wincap
>>>> member addition has_user_shstk with Windows version dependent value
>> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>>
>>> Is that actually the final version?  It's still missing the commit
>>> message text explaining things and the "Fixes" line...
>> Hi Corinna,
>>
>> Is more required above?
> 
> No, it's fine, albeit "Fixes:" is supposed to be kind of like a footer,
> just where the "Signed-off-by:" is, too.
> 
> But it's still only in the cover letter.  As I wrote, it needs to go
> into the actual patch, otherwise all the nice info doesn't make it into
> the git repo.

Ahah - finally got the point, although I wondered about whether a cover could be 
added as a note to the patches, and found that since early 2022 git 2.35+ allows

	$ git ... am --empty=keep --allow-empty ...

to create an empty commit, with the contents of the e-mail message as its log.
As we are on 2.39 you may already know this and prefer to keep the log clear. YMMV
Emailed v4.

-- 
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 retirer     but when there is no more to cut
                                 -- Antoine de Saint-Exupéry