Fix build against xorg-server >= 1.20.

Patch taken from upstream:
https://cgit.freedesktop.org/xorg/driver/xf86-video-savage/commit/?id=0ece556daa8a88771b669d8104396abd9166d2d0

diff --git a/src/savage_driver.c b/src/savage_driver.c
index 58a294d..3cda923 100644
--- a/src/savage_driver.c
+++ b/src/savage_driver.c
@@ -2034,8 +2034,6 @@ static Bool SavagePreInit(ScrnInfoPtr pScrn, int flags)
     xf86DrvMsg(pScrn->scrnIndex, X_PROBED, "Detected current MCLK value of %1.3f MHz\n",
 	       mclk / 1000.0);
 
-    pScrn->maxHValue = 2048 << 3;	/* 11 bits of h_total 8-pixel units */
-    pScrn->maxVValue = 2048;		/* 11 bits of v_total */
     pScrn->virtualX = pScrn->display->virtualX;
     pScrn->virtualY = pScrn->display->virtualY;
 
@@ -3637,6 +3635,14 @@ static ModeStatus SavageValidMode(SCRN_ARG_TYPE arg, DisplayModePtr pMode,
        (pMode->VDisplay > psav->PanelY)))
 	    return MODE_PANEL;
 
+    /* 11 bits of h_total 8-pixel units */
+    if (pMode->HTotal > (2048 << 3))
+	return MODE_BAD_HVALUE;
+
+    /* 11 bits of v_total */
+    if (pMode->VTotal > 2048)
+	return MODE_BAD_VVALUE;
+
     if (psav->UseBIOS) {
 	refresh = SavageGetRefresh(pMode);
         return (SavageMatchBiosMode(pScrn,pMode->HDisplay,
s.scm?id=6839095af7f6c8a8d5ad481db22cd4ece0a1b5e3'>tree</a><a href='/guix/commit/tests/packages.scm?id=6839095af7f6c8a8d5ad481db22cd4ece0a1b5e3'>commit</a><a href='/guix/diff/tests/packages.scm?id=6839095af7f6c8a8d5ad481db22cd4ece0a1b5e3'>diff</a></td><td class='form'><form class='right' method='get' action='/guix/log/tests/packages.scm'>
<input type='hidden' name='id' value='6839095af7f6c8a8d5ad481db22cd4ece0a1b5e3'/><select name='qt'>
<option value='grep'>log msg</option>
<option value='author'>author</option>
<option value='committer'>committer</option>
<option value='range'>range</option>
</select>
<input class='txt' type='search' size='10' name='q' value=''/>
<input type='submit' value='search'/>
</form>
</td></tr></table>
<div class='path'>path: <a href='/guix/log/?id=6839095af7f6c8a8d5ad481db22cd4ece0a1b5e3'>root</a>/<a href='/guix/log/tests?id=6839095af7f6c8a8d5ad481db22cd4ece0a1b5e3'>tests</a>/<a href='/guix/log/tests/packages.scm?id=6839095af7f6c8a8d5ad481db22cd4ece0a1b5e3'>packages.scm</a></div><div class='content'><table class='list nowrap'><tr class='nohover'><th class='left'>Age</th><th class='left'>Commit message (<a href='/guix/log/tests/packages.scm?id=6839095af7f6c8a8d5ad481db22cd4ece0a1b5e3&amp;showmsg=1'>Expand</a>)</th><th class='left'>Author</th></tr>
<tr><td><span title='2019-09-06 14:41:58 +0200'>2019-09-06</span></td><td><a href='/guix/commit/tests/packages.scm?id=bc60349b5bc58a0b803df5adce1de6db82453744'>packages: 'supported-package?' binds '%current-system' for graph traversal.</a><span class='msg-avail'>...<span class='msg-tooltip'>Previously, (supported-package? coreutils "armhf-linux")
with (%current-system) = "x86_64-linux" would return false.  That's
because 'supported-package?' would traverse the x86_64 dependency graph,
which contains 'tcc-boot0', which supports x86 only.

Consequently, 'supported-package?' would match only 53 packages for
"armhf-linux" when running on x86, as is the case during continuous
integration.

* guix/packages.scm (package-transitive-supported-systems): Add an
optional 'system' parameter.  Use 'mlambda' instead of 'mlambdaq' for
memoization.
(supported-package?): Pass 'system' to 'package-transitive-supported-systems'.
* tests/packages.scm ("package-transitive-supported-systems, implicit inputs")
("package-transitive-supported-systems: reduced binary seed, implicit inputs"):
Remove calls to 'invalidate-memoization!', which no longer work and were
presumably introduced to work around the bug we're fixing (see commit
0db65c168fd6dec57a357735fe130c80feba5460).
* tests/packages.scm ("supported-package?"): Rewrite test to use only
existing system name since otherwise 'bootstrap-executable' raises an
exception.
("supported-package? vs. system-dependent graph"): New test.
</span></span></td><td>Ludovic Courtès</td></tr>
<tr><td><span title='2019-06-27 23:33:48 +0200'>2019-06-27</span></td><td><a href='/guix/commit/tests/packages.scm?id=5cc1075a76392666d3d733837f5c6252b1e48002'>Merge branch 'master' into core-updates</a></td><td>Ludovic Courtès</td></tr>
<tr><td><span title='2019-06-27 11:14:41 +0200'>2019-06-27</span></td><td><a href='/guix/commit/tests/packages.scm?id=066eeae1a197b1926c7ffc42919d684686f56fdb'>packages: 'specification-&gt;package+output' distinguishes "no output specified".</a><span class='msg-avail'>...<span class='msg-tooltip'>Until now the caller couldn't tell the different between a spec like
"foo:out" and one like "foo".  This change allows users to distinguish
between these two cases.

* gnu/packages.scm (specification-&gt;package+output): Disable output
membership test when OUTPUT = #f and SUB-DRV = #f.
* tests/packages.scm ("specification-&gt;package+output")
("specification-&gt;package+output invalid output")
("specification-&gt;package+output no default output")
("specification-&gt;package+output invalid output, no default"): New tests.
</span></span></td><td>Ludovic Courtès</td></tr>
<tr><td><span title='2019-06-14 21:57:39 +0200'>2019-06-14</span></td><td><a href='/guix/commit/tests/packages.scm?id=03d76577b96ba81c9921eb3a297d42db8644280b'>tests: Make builds less expensive.</a><span class='msg-avail'>...<span class='msg-tooltip'>The switch to the reduced bootstrap broke build time assumptions made by
tests, notably the assumption that GNU-MAKE-BOOT0 was cheap to build.
This commit adjusts this to make these tests cheaper.

* gnu/packages/bootstrap.scm (%bootstrap-inputs-for-tests): New variable.
* guix/tests.scm (gnu-make-for-tests): New variable.
* tests/guix-environment.sh: Use GNU-MAKE-FOR-TESTS instead of
GNU-MAKE-BOOT0.  Remove test with FINDUTILS-BOOT0.
* tests/guix-package-net.sh (boot_make): Use GNU-MAKE-FOR-TESTS.
* tests/packages.scm ("GNU Make, bootstrap"): Likewise.
* tests/profiles.scm ("profile-derivation relative symlinks, two entries"):
Likewise.
* tests/union.scm (%bootstrap-inputs): Remove.
("union-build"): Use %BOOTSTRAP-INPUTS-FOR-TESTS instead of
%BOOTSTRAP-INPUTS.
</span></span></td><td>Ludovic Courtès</td></tr>