# fixes:  The commits are too invasive for stable. Instead the offending patches
#         causing regressions have been reverted.
365d34540f331df57780dddf8da87235be0a6bcb mesa: correctly calculate the storage offset for i915
de0e62e1065e2d9172acf3ab7c70bba0160125c8 st/mesa: correctly calculate the storage offset

# stable: Add loader::getCapability patches. It's rather invasive infra
#         not suitable as a bugfix.
1bf703e4ea5c4f742bc7ba55d01e5afc3f4e11f9 dri_interface,egl,gallium: only expose RGBA visuals on Android
be5773fa8dfe9255d9abaf5c7d5bbbd2d922da08 Android: fix compile error for DRI2 loader getCapability
31a6750988d7dd431f72ff1ff11bfca83bde5d8c st/dri: NULL check before deref DRI loader .getCapability

# stable: The commit addresses code that did not land in the stable branch
31bb8517a194af733deefe2d821537d994d39365 radv/gfx9: fix tile swizzle handling for gfx9

# stable: Commit is not applicable when 4fab67a4415 is missing.
d496780fb2c7f2cf0e32b6a79dc528e5156dfcb3 intel/eu/validate: Look up types on demand in execution_type()

# fixes: Depend on preseding commit which adds new public GBM API
3a5e3aa5a53cff55a5e31766d713a41ffa5a93d7 egl/drm: Fix misused x and y offsets in swrast_put_image2()
fe2a6281b3b299998fe7399e7dbcc2077d773824 egl/drm: Fix misused x and y offsets in swrast_get_image()

# fixes: This commit addressed an earlier commit c7e9ebb3ab8 which did not
#        land in branch
45c5c444518b7e83d9accd9f44702fa49282a3b8 radeonsi/gfx9: proper workaround for LS/HS VGPR initialization bug

# fixes: This commit addressed earlier commits 61ad2f13 and 6dcc54b4 which did
#        not land in branch
979978ee06867a531b8d56cee252f5c83920a339 radv: Check for GFX9 for 1D arrays in image_size intrinsic.

# fixes: This commit addressed earlier commits dcf46e99 and 60878dd0 which did
#        not land in branch
8e9e339c530c7b82b5a29d4b3183e8f5a01eae28 radv: copy the number of viewports/scissors at pipeline bind time

# stable: The commit regresses a few dEQP tests. Namely:
#         dEQP-VK.api.copy_and_blit.core.buffer_to_buffer.partial
#         dEQP-VK.api.copy_and_blit.dedicated_allocation.buffer_to_buffer.partial
14555d0b7a51bd3701764fd213c2459410143431 anv: Remove unreachable cases from isl_format_for_size()

# stable: The commit addresses earlier commit a62a9793357 which is no applicable
#         for the stable branch
6c7720ed78db754d52f204cbb74897aa9e65ea7e anv/wsi: Allocate enough memory for the entire image

# stable: Commits are too invasive for 17.2.
98fdff7247b6877d028d33284f9cc63189ee204e configure.ac: factor out detection for old and buggy llvm
13a53c4f5cdd664fd155c9e78fb46a4387af006c configure.ac: rework llvm libs handling for 3.9+
a7ecf7b86f4eae59f3ceac2125e5d1725c403c07 Travis: add binutils 2.26 for a few more LLVM 3.9 builds
36d6d1e931936a80da327889862ba02942ac427b configure.ac: add llvm_add_optional_component helper
df3a43018020c16c1dfa88a76c9a84c9fb85be38 configure.ac: add missing LLVM components for OpenCL

# stable: Commit is too big for stable at this point.
4d24a7cb97641cacecd371d1968f6964785822e4 glsl: fix derived cs variables

# stable: 17.3 nomination only.
fee9d05e2136b2b7c5a1ad2be7180b99f733f539 radv: Update code pointer correctly if a variant is already created

# stable: 17.3 nomination only.
d8cefaa197f02944812ef535b1b303dd5bf26848 radv: use device name in cache creation like radeonsi.

# fixes:  This commit addressed earlier commit 35ac13ed3 which did not
#         land in branch.
11d688d9f0d2ee4d0178d1807c0075e5e8364b1d mesa/bufferobj: don't double negate the range

# extra:  Commit is not applicable when ade416d0236 is missing.
07bfdb478bf844a0ac9cf3679f51f83c4abea5a1 broadcom/vc5: Propagate vc4 aliasing fix to vc5.

# stable: This commit addressed earlier commit 8d90e28839 which did
#         not land in branch.
446c5726ecb968d06a6607e0df42be1cb74948c4 i965: fix blorp stage_prog_data->param leak

# stable: This commit addressed earlier commit 78ade659569 which did
#         not land in branch.
8fbd82f464f26a56167f7962174b2b69756a105a etnaviv: don't do resolve-in-place without valid TS

# stable: This commit addressed earlier commit 8d90e28839 which did
#         not land in branch.
7b4387519c382cffef9c62bbbbefcfe71cfde905 intel/fs: Alloc pull constants off mem_ctx

# stable: 17.3 nomination only.
3f8e3c2bd8f54ae6817f7496be47f4e1a8860d9c radeonsi: add a workaround for weird s_buffer_load_dword behavior on SI
7dae419aa7c34af820c08896acef3b65d855188e Android: move drivers' symlinks to /vendor (v2)

# fixes:  This commit has more than one Fixes tag but the commit it
#         addresses didn't land in branch.
e17e8934f9e4b008bdfb4f9abd8ed4faa604c7d9 automake: include git_sha1.h.in in release tarball

# stable: This commit is not really needed after 6ac2d169019.
e8c9e65185de3e821e1e482e77906d1d51efa3ec intel/fs: Use a pure vertical stride for large register strides

# stable: These commits addressed earlier commit 379b24a40d3 which did
#         not land in branch.
7364f080f9a272323ed3491f278a1eed3eb9b1a7 intel/nir: Add a helper for getting the NoIndirect mask
3e63cf893f096a7263eb1856d58417dd2d170d4b intel/nir: Break the linking code into a helper in brw_nir.c
951a5dc4cc29da996b54ae63eeba1915a3a65b4a intel/nir: Use the correct indirect lowering masks in link_shaders

# stable: These commits resulted in a CTS regression being addressed
#         at https://bugs.freedesktop.org/show_bug.cgi?id=103626 .
18fde36ced4279f2577097a1a7d31b55f2f5f141 intel/fs: Use the original destination region for int MUL lowering

# stable: These commits are refactorings rather than fixes.
fcd4adb9d08094520fb8d118d3448b04c6ec1fd1 intel/fs: Pass builders instead of blocks into emit_[un]zip
0d905597fe2997c89022c76cdf84dc4fba5eb055 intel/fs: Be more explicit about our placement of [un]zip
6c00240bc650805e0b66aa6e17dbe69bbe41e446 intel/fs: Don't stomp f0.1 in SIMD16 ballot

# stable: This commit addressed earlier commit ea1b97714d9b which did
#         not land in branch.
50330d7115f0d5050ec3cfe6bca2b0136222e097 r600/shader: reserve first register of vertex shader.

# stable: This commit depends on earlier commit 3735af04152b which did
#         not land in branch.
a6cc361e5fd2450249847d5ee8093d26ed7ff545 anv/cmd_buffer: Advance the address when initializing clear colors

# stable: This commit addressed earlier commit a62a97933578 which did
#         not land in branch.
a07f7b26198ce0f5c8799481a673754968ac5daf anv/cmd_buffer: Take bo_offset into account in fast clear state addresses

# stable: These commits addressed earlier commit 2c4097aff1b which did
#         not land in branch.
344252a27f8d875572bbe65641a825af8e73845d i965/bufmgr: Add a helper to mark a BO as external
0a6a137eb27129e17298cfe9dd620205588ee4f6 i965: Mark BOs as external when we export their handle

# stable: 17.3 nomination only.
6e4d65f674a70809e6df1a4f716f874828915562 broadcom/vc5: Add vc5_drm.h to the release tarball
4639cc716e89c69da41c7b54fa938457000fbd4c intel/blorp: Use mocs.tex for depth stencil
deec84fd771876b5c0755293376df11bc95b473b anv/blorp: Add a device parameter to blorp_surf_for_anv_image
bc933d0e8462871e19328f66182c35543e334013 intel/blorp: Make the MOCS setting part of blorp_address
d7a19d69ebc032ba7207fc97bc6f10d5bb35bb99 i965: Use PTE MOCS for all external buffers

# fixes:  This commit is only a typo correction on an error message.
a6932faae1074445210d392a80b94fdac147b255 glsl: Fix typo fragement -> fragment

# fixes:  This commit makes reference to 2 other commits but none have
#         made it to the 17.2 queue.
9b0223046668593deb9c0be0b557994bb5218788 egl: pass the dri2_dpy to the $plat_teardown functions

# extra:  The commit just references a proper fix that has already
#         landed.
a31d0382084c8aa860ffcef9b12592c5c44e192f Revert "intel/fs: Use a pure vertical stride for large register strides"

# stable: The commit depends on at least one other that did not land in
#          branch - 8b3a2578519.
010214b403de1b5e25a549372ba6192b89e05d06 radeonsi: allow DMABUF exports for local buffers

# stable: This commit addressed earlier commit ead0dfe31ec7 which did
#         not land in branch.
709f5bdc4a2bf31f422f5cf60797224c0463c10a swr: Fix KNOB_MAX_WORKER_THREADS thread creation override.

# stable: 17.3 nomination only.
bf0904e31fb7d9cd8932d582076c8d7beb02ba89 winsys/amdgpu: disable local BOs again due to worse performance
35c3cbad3c30ad3d40a6811dd6ca2286e013bfc5 radeonsi: don't call force_dcc_off for buffers

# fixes:  This commit addressed earlier commit d1c9f30d7ff7 which did
#         not land in branch.
1bdeac545f4ea9f7ca6947f5da7fcf4f5b3010dc radv: port merge tess info from anv

# extra:  The commit just references a fix for an additional change in its v2.
c1ff99fd70cd2ceb2cac4723e4fd5efc93834746 main: Clear shader program data whenever ProgramBinary is called

# extra:  The commit references a previous commit in which the changes
#         should have been included but, as clarified by the
#         developer, it is not needed for stable.
71e630753ebbee82e8f8709da5488296b2c070c8 r600: set DX10_CLAMP for compute shader too