- Update to 11.0.2
* This is a emergency release which covers the final dEQP
regressions introduced in the 11.0 development cycle.

- update upstream status:
* remove n_i965-Remove-early-release-of-DRI2-miptree.patch
* add U_i965-Remove-early-release-of-DRI2-miptree.patch

- Update to 11.0.1
* i965/vec4: Fix saturation errors when coalescing registers
* i965/vec4_nir: Load constants as integers
* meta: Abort meta pbo path if TexSubImage need signed unsigned
conversion
* docs: add sha256 checksums for 11.0.0
* Update version to 11.0.1
* docs: add release notes for 11.0.1
* mesa: Fix GL_FRAMEBUFFER_ATTACHMENT_OBJECT_TYPE for default
framebuffer.
* t_dd_dmatmp: Make "count" actually be the count
* t_dd_dmatmp: Clean up improper code formatting from previous
patch
* t_dd_dmatmp: Use '& 3' instead of '% 4' everywhere
* t_dd_dmatmp: Pull out common 'count -= count & 3' code
* t_dd_dmatmp: Use addition instead of subtraction in loop
bounds
* st/mesa: avoid integer overflows with buffers >= 512MB
* nv50, nvc0: fix max texture buffer size to 128M elements
* freedreno/a3xx: fix blending of L8 format
* nv50,nvc0: detect underlying resource changes and update tic
* nv50,nvc0: flush texture cache in presence of coherent bufs
* radeonsi: load fmask ptr relative to the resources array
* nir: Fix a bunch of ralloc parenting errors
* i965/vec4: Don't reswizzle hardware registers
* configure.ac: Add support to enable read-only text segment on
x86.
* gbm: convert gbm bo format to fourcc format on dma-buf import
* mesa: fix errors when reading depth with glReadPixels
* i965: fix textureGrad for cubemaps
* mesa: Fix texture compression on big-endian systems