[Bug 45290] New: [bisected r200] fdo23670-drawpix_stencil fails and crashes

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=45290

             Bug #: 45290
           Summary: [bisected r200] fdo23670-drawpix_stencil fails and
                    crashes
    Classification: Unclassified
           Product: Mesa
           Version: git
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: medium
         Component: Drivers/DRI/r200
        AssignedTo: dri-devel at lists.freedesktop.org
        ReportedBy: idr at freedesktop.org


There seem to be at least two problems with fdo23670-drawpix_stencil (and other
stencil tests) on r200.  There is incorrect rendering, which bisects to the
first commit listed below, and a crash, which bisects to the second.  The
backtrace of the segfault is:

(gdb) bt
#0  0x00007ffff5a2fd5f in radeon_unmap_renderbuffer_s8z24 (ctx=0x6992d0, 
    rb=0x85d2d0) at radeon_fbo.c:270
#1  0x00007ffff5a2fe31 in radeon_unmap_renderbuffer (ctx=0x6992d0, rb=0x85d2d0)
    at radeon_fbo.c:290
#2  0x00007ffff5a40726 in radeon_renderbuffer_unmap (ctx=0x6992d0, rb=0x85d2d0)
    at radeon_span.c:510
#3  0x00007ffff5a408a8 in radeon_unmap_framebuffer (ctx=0x6992d0, fb=0x85cc50)
    at radeon_span.c:545
#4  0x00007ffff5a40a67 in radeonSpanRenderFinish (ctx=0x6992d0)
    at radeon_span.c:579
#5  0x00007ffff5d0d249 in swrast_render_finish (ctx=0x6992d0)
    at swrast/s_context.h:335
#6  0x00007ffff5d0f5ae in _swrast_DrawPixels (ctx=0x6992d0, x=50, y=50, 
    width=20, height=20, format=6401, type=5121, unpack=0x6a8bb0, 
    pixels=0x7fffffffdf30) at swrast/s_drawpix.c:750
#7  0x00007ffff5c0c9a6 in _mesa_meta_DrawPixels (ctx=0x6992d0, x=50, y=50, 
    width=20, height=20, format=6401, type=5121, unpack=0x6a8bb0, 
    pixels=0x7fffffffdf30) at drivers/common/meta.c:2217
#8  0x00007ffff5ca6311 in _mesa_DrawPixels (width=20, height=20, format=6401, 
    type=5121, pixels=0x7fffffffdf30) at main/drawpix.c:131
#9  0x0000000000429ba0 in piglit_display ()
    at /home/idr/devel/graphics/piglit/tests/bugs/fdo23670-drawpix_stencil.c:68
#10 0x0000000000429c51 in display ()
    at /home/idr/devel/graphics/piglit/tests/util/piglit-framework.c:56
#11 0x00007ffff7b0db07 in ?? () from /usr/lib64/libglut.so.3
#12 0x00007ffff7b11269 in fgEnumWindows () from /usr/lib64/libglut.so.3
#13 0x00007ffff7b0dfe2 in glutMainLoopEvent () from /usr/lib64/libglut.so.3
#14 0x00007ffff7b0e8d5 in glutMainLoop () from /usr/lib64/libglut.so.3
#15 0x000000000042a3c0 in main (argc=1, argv=0x7fffffffe4b8)
    at /home/idr/devel/graphics/piglit/tests/util/piglit-framework.c:304
(gdb) print tiled_s8z24_map
$1 = (uint32_t *) 0x7ffff4d09000
(gdb) print dst_offset
$2 = 53288
(gdb) print untiled_s8z24_map 
$3 = (uint32_t *) 0x868b10
(gdb) print src_offset
$4 = 256
(gdb) print tiled_s8z24_map[dst_offset/4]
Cannot access memory at address 0x7ffff4d16028
(gdb) 

commit 345fc4161967f15fb80848cd7dc6a63100f8c12d
Author: Eric Anholt <eric at anholt.net>
Date:   Wed Oct 12 17:05:20 2011 -0700

    swrast: Convert color glReadPixels slow path to using MapRenderbuffer.

    This may be a bit slower than before because we're switching from
    per-format compiled loops in GetRow to
    _mesa_unpack_rgba_block_unpack's loop around a callback to unpack a
    pixel.  The solution there would be to make _mesa_unpack_rgba_block
    fold the span loop into the format handlers.

    (On the other hand, function call overhead will hardly matter if
    MapRenderbuffer means the driver gets the data into cacheable memory
    instead of uncached).

    The adjust_colors code should no longer be required, since the unpack
    function does the 565 to float conversion in a single pass instead of
    converting it (poorly) through 8888 as apparently happened in the
    past.

    Reviewed-by: Brian Paul <brianp at vmware.com>

commit 7d91ecf7a3a08c01a704f2d427444f7a97991680
Author: Dave Airlie <airlied at redhat.com>
Date:   Mon Dec 5 19:15:04 2011 +0000

    radeon/r200: add draw/stencil buffer detiling

    This moves the detiling to the fbo mapping, r200 depth is always tiled,
    and we can't detile it with the blitter.

    Signed-off-by: Dave Airlie <airlied at redhat.com>

-- 
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.