Re: [xserver,1/4] os: move xf86PrivsElevated here

[Antoine Martin]

On 09/03/18 05:11, Ben Crocker wrote:
> Sent that a little too soon; please consider "Reviewed-by: Ben Crocker
> >" to be
> added after each of the "Signed-off-by: Nicolai Haehnle
> >" lines.
You can add:
Reviewed-by: Antoine Martin 

You should probably re-send the patch series as individual emails.

Cheers
Antoine


> 
> Also please note that I rebased Nicolai's original patch,
> https://patchwork.freedesktop.org/series/18684/,
> against a very recent copy of master.
> 
>   Thanks,
>   Ben
> 
> 
> On Thu, Mar 8, 2018 at 4:53 PM, Ben Crocker  > wrote:
> 
> From: Nicolai Hähnle >
> 
> From: Nicolai Hähnle  >
> 
> Having different types of code all trying to check for elevated
> privileges
> is a bad idea. This implementation is the most thorough one.
> 
> Signed-off-by: Nicolai Hähnle  >
> ---
>  hw/xfree86/common/xf86Init.c | 59
> +
>  include/os.h                 |  3 +++
>  os/utils.c                   | 63
> 
>  3 files changed, 67 insertions(+), 58 deletions(-)
> 
> diff --git a/hw/xfree86/common/xf86Init.c b/hw/xfree86/common/xf86Init.c
> index e61fe66..758b926 100644
> --- a/hw/xfree86/common/xf86Init.c
> +++ b/hw/xfree86/common/xf86Init.c
> @@ -230,78 +230,21 @@ xf86PrintBanner(void)
>      xf86ErrorFVerb(0, "Current version of pixman: %s\n",
>                     pixman_version_string());
>      xf86ErrorFVerb(0, "\tBefore reporting problems, check "
>                     "" __VENDORDWEBSUPPORT__ "\n"
>                     "\tto make sure that you have the latest
> version.\n");
>  }
> 
>  Bool
>  xf86PrivsElevated(void)
>  {
> -    static Bool privsTested = FALSE;
> -    static Bool privsElevated = TRUE;
> -
> -    if (!privsTested) {
> -#if defined(WIN32)
> -        privsElevated = FALSE;
> -#else
> -        if ((getuid() != geteuid()) || (getgid() != getegid())) {
> -            privsElevated = TRUE;
> -        }
> -        else {
> -#if defined(HAVE_ISSETUGID)
> -            privsElevated = issetugid();
> -#elif defined(HAVE_GETRESUID)
> -            uid_t ruid, euid, suid;
> -            gid_t rgid, egid, sgid;
> -
> -            if ((getresuid(, , ) == 0) &&
> -                (getresgid(, , ) == 0)) {
> -                privsElevated = (euid != suid) || (egid != sgid);
> -            }
> -            else {
> -                printf("Failed getresuid or getresgid");
> -                /* Something went wrong, make defensive assumption */
> -                privsElevated = TRUE;
> -            }
> -#else
> -            if (getuid() == 0) {
> -                /* running as root: uid==euid==0 */
> -                privsElevated = FALSE;
> -            }
> -            else {
> -                /*
> -                 * If there are saved ID's the process might still
> be privileged
> -                 * even though the above test succeeded. If
> issetugid() and
> -                 * getresgid() aren't available, test this by
> trying to set
> -                 * euid to 0.
> -                 */
> -                unsigned int oldeuid;
> -
> -                oldeuid = geteuid();
> -
> -                if (seteuid(0) != 0) {
> -                    privsElevated = FALSE;
> -                }
> -                else {
> -                    if (seteuid(oldeuid) != 0) {
> -                        FatalError("Failed to drop privileges. 
> Exiting\n");
> -                    }
> -                    privsElevated = TRUE;
> -                }
> -            }
> -#endif
> -        }
> -#endif
> -        privsTested = TRUE;
> -    }
> -    return privsElevated;
> +    return PrivsElevated();
>  }
> 
>  static void
>  TrapSignals(void)
>  {
>      if (xf86Info.notrapSignals) {
>          OsSignal(SIGSEGV, SIG_DFL);
>          OsSignal(SIGABRT, SIG_DFL);
>          OsSignal(SIGILL, SIG_DFL);
>  #ifdef SIGEMT
> diff --git a/include/os.h b/include/os.h
> index d2c41b4..686f6d6 100644
> --- a/include/os.h
> +++ b/include/os.h
> @@ -355,20 +355,23 @@ Fclose(void *);
>  extern const char *
>  Win32TempDir(void);
> 
>  extern int
>  System(const char *cmdline);
> 
>  #define Fopen(a,b) fopen(a,b)
>  #define Fclose(a) fclose(a)
>  #endif
> 
> 

[PATCH v3 app/xrandr] Document that --dpi and --fbmm options set DPI of whole X screen

[Pali Rohár]

Explicitly document and make it clear that those options do not change
DPI of some monitor output. Also state that these options have no useful
meaning for multi-monitor configuration.

Signed-off-by: Pali Rohár 
---
 man/xrandr.man | 19 +++
 1 file changed, 15 insertions(+), 4 deletions(-)

diff --git a/man/xrandr.man b/man/xrandr.man
index aa82724..1291bad 100644
--- a/man/xrandr.man
+++ b/man/xrandr.man
@@ -257,14 +257,25 @@ fit within this size. When this option is not provided, 
xrandr computes the
 smallest screen size that will hold the set of configured outputs; this
 option provides a way to override that behaviour.
 .IP "\-\-fbmm \fIwidth\fPx\fIheight\fP"
-Sets the reported values for the physical size of the screen. Normally,
+Sets the value reported as physical size of the X screen as a whole
+(union of all configured monitors). In configurations with multiple
+monitors with different DPIs, the value has no physical meaning, but
+it may be used by some legacy clients which do not support RandR
+version 1.2 to compute a reference font scaling. Normally,
 xrandr resets the reported physical size values to keep the DPI constant.
-This overrides that computation.
+This overrides that computation. Default DPI value is 96.
 .IP "\-\-dpi \fIdpi\fP"
 .IP "\-\-dpi \fIfrom-output\fP"
-This also sets the reported physical size values of the screen, it uses either
+This also sets the value reported as physical size of the X screen as a whole
+(union of all configured monitors). In configurations with multiple
+monitors with different DPIs, the value has no physical meaning, but
+it may be used by some legacy clients which do not support RandR
+version 1.2 to compute a reference font scaling. This option uses either
 the specified DPI value, or the DPI of the given output, to compute an 
appropriate
-physical size using whatever pixel size will be set.
+physical size using whatever pixel size will be set. Typical values are
+the default (96 DPI), the DPI of the only monitor in single-monitor
+configurations, or the DPI of the primary monitor in multi-monitor
+configurations.
 .IP "\-\-newmode \fIname\fP \fImode\fP"
 New modelines can be added to the server and then associated with outputs.
 This option does the former. The \fImode\fP is specified using the ModeLine
-- 
2.11.0

___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel

Missing swapl() in Xext/saver.c?

[Mihai Moldovan]

Hi


In
https://cgit.freedesktop.org/xorg/xserver/commit/?id=9edcae78c46286baff42e74bfe26f6ae4d00fe01
the swapl() call for stuff->suspend was removed. That sounds like a bad idea to 
me?

xScreenSaverSuspendReq to this day defines it as a Bool, which is typedef'd to
int everywhere and includes the (ignored) B32 width specifier, so using swapl()
on the member sounds like a good idea. Definitely isn't a CARD8 object, despite
its name.



Mihai



signature.asc
Description: OpenPGP digital signature
___
xorg-devel@lists.x.org: X.Org development
Archives: http://lists.x.org/archives/xorg-devel
Info: https://lists.x.org/mailman/listinfo/xorg-devel