Signed-off-by: Mario Trangoni <mjtrangoni@gmail.com>

Attributes GH PR #818: https://github.com/ArcticaProject/nx-libs/pull/818

We can only free the xkbDevicePrivate because we do not know the
details of any other (possible) extension. So let's limit to that one
private for now and call the new xkbFreePrivates from dix (where such
a function is completely missing).

Attributes GH PR #815: https://github.com/ArcticaProject/nx-libs/pull/815

there's more work to do...

The whole nx-X11/extras dir is no longer checked since we only have
Mesa there and we do not want to check that mess...

All the exceptions in static-analysis-suppressions have been manually
checked.

[nx-X11/programs/Xserver/record/set.c:361]: (warning) Possible null pointer dereference: stackIntervals

stackIntervals is only NULL if nIntervals is 0, too. In that case
memcpy will do nothing and so it is ok to pass NULL as source. But it
is ugly nevertheless...

[Keyboard.c:559]: (error) Shifting signed 32-bit value by 31 bits is undefined behaviour

  commit 758393951233d1b2520cf4cefd33ec4288a3880a
  Author: Dave Airlie <airlied@redhat.com>
  Date:   Wed Sep 12 11:09:40 2018 +1000

    xkb: fix what looks to be a copy-paste error with first vs firstMM

    Pointed out by coverity.
Signed-off-by: Dave Airlie <airlied@redhat.com>

xkmread.c: In function ‘XkmReadFileSectionName’:
xkmread.c:1181:25: warning: ‘tmpTOC.type’ may be used uninitialized in this function [-Wmaybe-uninitialized]
     XkbConfigText(tmpTOC.type,XkbMessage),0);

  commit 83913de25d35709b3ab7b0ab124b73924145d2dd
  Author: Adam Jackson <ajax@redhat.com>
  Date:   Thu Apr 5 12:59:11 2018 -0400

    xkb: Silence some compiler warnings

    Of the form:

    ../xkb/XKBGAlloc.c: In function ‘SrvXkbAddGeomKeyAlias’:
    ../xkb/XKBGAlloc.c:591:13: warning: ‘strncpy’ specified bound 4 equals destination size [-Wstringop-truncation]
                 strncpy(alias->real, realStr, XkbKeyNameLength);
                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

    This is intentional; the code that reads from these fields never reads
    more than 4 bytes anyway. Rephrase things in terms of memcpy so that's
    clear. Obviously this is awful but in XKB awful is par.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Acked-by: Keith Packard <keithp@keithp.com>

  commit a4a2e814d5d0e6152307a301eda1d6fc1c555aaa
  Author: Alan Coopersmith <alan.coopersmith@oracle.com>
  Date:   Sun Feb 13 21:36:02 2011 -0800

    xkb: Use snprintf to measure string lengths instead of manual strlen math
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>

  commit 2391c409a2840d61fed93832650c0d6c82ebebdf
  Author: Eamon Walsh <ewalsh@tycho.nsa.gov>
  Date:   Fri Jun 13 22:48:17 2008 -0400

    Fix "warning: unused variable `s'".

  commit 534fc5140b039a8c98ab715d0a6740d513b41209
  Author: Daniel Stone <daniel@fooishbar.org>
  Date:   Sun Feb 3 23:30:22 2008 +1100

    XKB: Remove a bunch of mad ifdefs

    We have SEEK_SET and size_t, seriously.  Also use DebugF instead of
    ifdef DEBUG, and ditch a couple of random bits that were never used.

Fixes this cppcheck finding:
[nx-X11/programs/Xserver/xkb/xkbActions.c:1306]: (error) Uninitialized variable: oldState

  commit 35a4b8e7f4526a92d44cb16a783f21030cd1f6df
  Author: Peter Hutterer <peter.hutterer@who-t.net>
  Date:   Tue May 5 16:51:59 2009 +1000

    xkb: remove oldState from XkbHandleActions.

    I really don't know what the purpose of this variable is or was, aside from
    potentially clobbering up our key state since there's a path where it may be
    used uninitialised.

    Also, this means that xkbi->prev_state is now accessible from the DIX with
    meaningful data.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>

Move allocation of image_index close before first_use. This way we do
not need to free it if previous step fail. And we cannot forget that
free() call.

While at it replace malloc+memset by calloc.

Otherwise it will (falsely) report "Memory pointed to by 'sessionpath' is freed twice."

While at it create and use the freeDepths() helper function.

  commit 211d4c2d353b5e379716484055a3f58235ea65f4
  Author: Chris Wilson <chris@chris-wilson.co.uk>
  Date:   Wed Dec 14 15:55:22 2011 +0000

    render: Propagate allocation failure from createSourcePicture()

    All the callers were already checking for failure, except that
    createSourcePicture() itself was failing to check whether it
    successfully allocated the Picture.

    [ajax: Rebase, fix line wrap of preceding line]
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Jeremy Huddleston <jeremyhu@apple.com>

  commit 2aa935bc5cc1e2d5365a97b8c5bb3d33eb5fc758
  Author: Tiago Vignatti <tiago.vignatti@nokia.com>
  Date:   Fri Mar 25 22:10:55 2011 +0200

    fb: fix memory leak in fbOverlayFinishScreenInit
Signed-off-by: Tiago Vignatti <tiago.vignatti@nokia.com>
Reviewed-by: Nicolas Peninguy <nico@lostgeeks.org>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>

==12976==ERROR: LeakSanitizer: detected memory leaks

Direct leak of 6 byte(s) in 1 object(s) allocated from:
    #0 0x7f510b3ac810 in strdup (/usr/lib/x86_64-linux-gnu/libasan.so.5+0x3a810)
    #1 0x559ca29c5035 in nxagentKeyboardProc /home/uli/work/nx/ArcticaProject/nx-libs/nx-X11/programs/Xserver/hw/nxagent/Keyboard.c:866
    #2 0x7a29bff07  (<unknown module>)

Direct leak of 1 byte(s) in 1 object(s) allocated from:
    #0 0x7f510b3ac810 in strdup (/usr/lib/x86_64-linux-gnu/libasan.so.5+0x3a810)
    #1 0x559ca29c509a in nxagentKeyboardProc /home/uli/work/nx/ArcticaProject/nx-libs/nx-X11/programs/Xserver/hw/nxagent/Keyboard.c:870
    #2 0x7a29bff07  (<unknown module>)

Direct leak of 1 byte(s) in 1 object(s) allocated from:
    #0 0x7f510b3ac810 in strdup (/usr/lib/x86_64-linux-gnu/libasan.so.5+0x3a810)
    #1 0x559ca29c507f in nxagentKeyboardProc /home/uli/work/nx/ArcticaProject/nx-libs/nx-X11/programs/Xserver/hw/nxagent/Keyboard.c:869
    #2 0x7a29bff07  (<unknown module>)

SUMMARY: AddressSanitizer: 8 byte(s) leaked in 3 allocation(s).

If compiled with -fsanitize=address this showed up when running startlxde:

==11551==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x60d000018fbc at pc 0x7f270a9ed57b bp 0x7fff30ef3050 sp 0x7fff30ef2800
READ of size 204 at 0x60d000018fbc thread T0
    #0 0x7f270a9ed57a  (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xb857a)
    #1 0x559dafcd5c93 in FindGlyphRef ../../render/glyph.c:179
    #2 0x559dafcd705d in AddGlyph /work/nx-libs/nx-X11/programs/Xserver/hw/nxagent/NXglyph.c:71
    #3 0x559dafccc0ff in ProcRenderAddGlyphs ../../mi/../render/render.c:1186
    #4 0x559dafcbd5a5 in ProcRenderDispatch /work/nx-libs/nx-X11/programs/Xserver/hw/nxagent/NXrender.c:1689
    #5 0x559dafcbc4ea in Dispatch /work/nx-libs/nx-X11/programs/Xserver/hw/nxagent/NXdispatch.c:476
    #6 0x559dafc4e9b0 in main /work/nx-libs/nx-X11/programs/Xserver/dix/main.c:353
    #7 0x7f2708e1d09a in __libc_start_main ../csu/libc-start.c:308
    #8 0x559dafc4f5d9 in _start (/work/nx-libs/nx-X11/programs/Xserver/nxagent+0x6e5d9)

0x60d000018fbc is located 0 bytes to the right of 140-byte region [0x60d000018f30,0x60d000018fbc)
allocated by thread T0 here:
    #0 0x7f270aa1e330 in __interceptor_malloc (/usr/lib/x86_64-linux-gnu/libasan.so.5+0xe9330)
    #1 0x559dafcd646c in AllocateGlyph ../../render/glyph.c:348

This happens when two glyphs are compared via memcmp and the smaller
one happens to be identical to the beginning of the bigger one.

Newer render implementations use a sha1 hash instead of memcmp so this
patch will (hopefully) be obsolete once render gets updated.

==10226== 3,337 bytes in 1 blocks are definitely lost in loss record 295 of 307
==10226==    at 0x483577F: malloc (vg_replace_malloc.c:299)
==10226==    by 0x6281DB9: strdup (strdup.c:42)
==10226==    by 0x2ABA9E: __glXClientInfo (glxcmds.c:2170)
==10226==    by 0x17CA3E: __glXDispatch (NXglxext.c:128)
==10226==    by 0x16EE77: Dispatch (NXdispatch.c:476)
==10226==    by 0x14DCE0: main (main.c:353)

There's no point in trying to free cl->* after memset(0).

This one is a bug that is found identically in xorg upstream and has
only been fixed during rework of the whole client resource freeing
stuff. So we fix it in glxext.c.

Fix these memory leaks:

==30021== 128 bytes in 1 blocks are definitely lost in loss record 230 of 302
==30021==    at 0x483577F: malloc (vg_replace_malloc.c:299)
==30021==    by 0x2EF89C: init_visuals (xf86glx.c:390)
==30021==    by 0x2EF89C: __MESA_initVisuals (xf86glx.c:541)
==30021==    by 0x17C922: GlxInitVisuals (glxext.c:317)
==30021==    by 0x218E73: fbInitVisuals (fbcmap.c:668)
==30021==    by 0x20BEB1: fbFinishScreenInit (fbscreen.c:229)
==30021==    by 0x20C275: fbScreenInit (fbscreen.c:273)
==30021==    by 0x1E0317: nxagentOpenScreen (Screen.c:1357)
==30021==    by 0x16D848: AddScreen (dispatch.c:4171)
==30021==    by 0x1DB7FF: InitOutput (Init.c:396)
==30021==    by 0x14DB12: main (main.c:279)
==30021==
==30021== 3,072 (192 direct, 2,880 indirect) bytes in 1 blocks are definitely lost in loss record 290 of 302
==30021==    at 0x483577F: malloc (vg_replace_malloc.c:299)
==30021==    by 0x2CCCC7: _gl_context_modes_create (glcontextmodes.c:364)
==30021==    by 0x2EF87C: init_visuals (xf86glx.c:381)
==30021==    by 0x2EF87C: __MESA_initVisuals (xf86glx.c:541)
==30021==    by 0x17C922: GlxInitVisuals (glxext.c:317)
==30021==    by 0x218E73: fbInitVisuals (fbcmap.c:668)
==30021==    by 0x20BEB1: fbFinishScreenInit (fbscreen.c:229)
==30021==    by 0x20C275: fbScreenInit (fbscreen.c:273)
==30021==    by 0x1E0317: nxagentOpenScreen (Screen.c:1357)
==30021==    by 0x16D848: AddScreen (dispatch.c:4171)
==30021==    by 0x1DB7FF: InitOutput (Init.c:396)
==30021==    by 0x14DB12: main (main.c:279)

The problem here is that GlxInitVisuals is called twice. First via
fbScreenInit and then again via nxagentInitGlxExtension. We remove the
first one to ensure the code in nxagenOpenScreen works as initially
intended.

There's an xorg upstream patch that does the same
(7d74690536b64f7b8e8036507ab7790807349c50), but it also cleans up
other stuff we do not even have in out source (yet?).

==12280== 0 bytes in 5 blocks are definitely lost in loss record 1 of 304
==12280==    at 0x483577F: malloc (vg_replace_malloc.c:299)
==12280==    by 0x2EFC29: init_visuals (xf86glx.c:489)
==12280==    by 0x2EFC29: __MESA_initVisuals (xf86glx.c:540)
==12280==    by 0x17C902: GlxInitVisuals (glxext.c:317)
==12280==    by 0x218C03: fbInitVisuals (fbcmap.c:668)
==12280==    by 0x20BC41: fbFinishScreenInit (fbscreen.c:229)
==12280==    by 0x20C005: fbScreenInit (fbscreen.c:273)
==12280==    by 0x1E024C: nxagentOpenScreen (Screen.c:1356)
==12280==    by 0x16D828: AddScreen (dispatch.c:4171)
==12280==    by 0x1DB7DF: InitOutput (Init.c:396)
==12280==    by 0x14DB12: main (main.c:279)
==12280==
==12280== 64 bytes in 2 blocks are definitely lost in loss record 223 of 304
==12280==    at 0x483577F: malloc (vg_replace_malloc.c:299)
==12280==    by 0x2EFA05: init_visuals (xf86glx.c:489)
==12280==    by 0x2EFA05: __MESA_initVisuals (xf86glx.c:540)
==12280==    by 0x17C902: GlxInitVisuals (glxext.c:317)
==12280==    by 0x218C03: fbInitVisuals (fbcmap.c:668)
==12280==    by 0x20BC41: fbFinishScreenInit (fbscreen.c:229)
==12280==    by 0x20C005: fbScreenInit (fbscreen.c:273)
==12280==    by 0x1E024C: nxagentOpenScreen (Screen.c:1356)
==12280==    by 0x16D828: AddScreen (dispatch.c:4171)
==12280==    by 0x1DB7DF: InitOutput (Init.c:396)
==12280==    by 0x14DB12: main (main.c:279)