From: Martin Jansa <martin.jansa@gmail.com>
To: yocto-patches@lists.yoctoproject.org
Cc: Martin Jansa <martin.jansa@gmail.com>
Subject: [meta-tensorflow][mickledore][PATCH 1/3] tensorflow: drop Upstream-Status inside third_party .patch files
Date: Tue, 11 Feb 2025 17:14:57 +0100 [thread overview]
Message-ID: <20250211161500.3077588-1-martin.jansa@gmail.com> (raw)
* these patches create:
third_party/0001-XNNPACK-support-32-bit-x86.patch
third_party/0001-cpuinfo-support-32-bit-x86.patch
third_party/0001-hardcode-go-SDKS.patch
third_party/eigen3/0001-workaround-ice-failure-while-building-from-source-wi.patch
which aren't applied by do_patch bitbake task but by bazel:
tensorflow $ grep patch_file tensorflow/workspace2.bzl | head -n 2
tensorflow/workspace2.bzl: patch_file = ["//third_party:0001-XNNPACK-support-32-bit-x86.patch"],
tensorflow/workspace2.bzl: patch_file = ["//third_party:0001-cpuinfo-support-32-bit-x86.patch"],
tensorflow/workspace2.bzl: patch_file = ["//third_party:0001-hardcode-go-SDKS.patch"],
third_party/eigen3/workspace.bzl: patch_file = ["//third_party/eigen3:0001-workaround-ice-failure-while-building-from-source-wi.patch"],
so they don't really need Upstream-Status and including it there
might cause false possitive e.g. in
0001-fix-XNNPACK-build-failure-for-qemuarm.patch
ERROR: tensorflow-lite-2.10.0-r0 do_patch: QA Issue: Malformed Upstream-Status in patch
recipes-framework/tensorflow/tensorflow-lite/0001-fix-XNNPACK-build-failure-for-qemuarm.patch
Please correct according to https://docs.yoctoproject.org/contributor-guide/recipe-style-guide.html#patch-upstream-status :
Upstream-Status: Inappropriate [oe specific] [patch-status]
ERROR: tensorflow-lite-2.10.0-r0 do_patch: Fatal QA errors were found, failing task.
caused by extra space before Upstream-Status
Maybe we should drop 0001-fix-XNNPACK-build-failure-for-qemuarm.patch completely
as it was already dropped from tensorflow in:
https://git.yoctoproject.org/meta-tensorflow/commit/?h=mickledore&id=baae9cd6d92f69aa16b874958a332e1d3a362fea
and from tensorflow-lite in scarthgap branch:
https://git.yoctoproject.org/meta-tensorflow/commit/?h=scarthgap&id=29509c97e9009ed0ae99ce6c2554e463575dc10c
and is now used only for tensorflow-lite, refresh it to apply
It was originally introduced for both in:
https://git.yoctoproject.org/meta-tensorflow/commit/?h=mickledore&id=e0efac349f2aac3d65ccbeee20475dbb1ea3dd19
and refreshed in:
https://git.yoctoproject.org/meta-tensorflow/commit/?h=mickledore&id=30fbec74ea37b646473b08d0ee8d3992ccd56245
Signed-off-by: Martin Jansa <martin.jansa@gmail.com>
---
.../tensorflow/files/0001-hardcode-go-SDKS.patch | 4 +---
.../0001-support-32-bit-x64-and-arm-for-yocto.patch | 7 ++-----
...y-eigen_archive-workaround-ice-failure-whi.patch | 4 +---
...019-fix-XNNPACK-build-failure-for-qemuarm.patch} | 13 ++++---------
.../tensorflow/tensorflow-lite_2.12.0.bb | 2 +-
5 files changed, 9 insertions(+), 21 deletions(-)
rename recipes-framework/tensorflow/files/{0001-fix-XNNPACK-build-failure-for-qemuarm.patch => 0019-fix-XNNPACK-build-failure-for-qemuarm.patch} (97%)
diff --git a/recipes-framework/tensorflow/files/0001-hardcode-go-SDKS.patch b/recipes-framework/tensorflow/files/0001-hardcode-go-SDKS.patch
index b705ec7..e365c27 100644
--- a/recipes-framework/tensorflow/files/0001-hardcode-go-SDKS.patch
+++ b/recipes-framework/tensorflow/files/0001-hardcode-go-SDKS.patch
@@ -38,7 +38,7 @@ new file mode 100644
index 00000000000..b51d7262f4c
--- /dev/null
+++ b/third_party/0001-hardcode-go-SDKS.patch
-@@ -0,0 +1,49 @@
+@@ -0,0 +1,47 @@
+From 34242e73f76eff09ee33152c58c3be650331b7ff Mon Sep 17 00:00:00 2001
+From: Hongxu Jia <hongxu.jia@eng.windriver.com>
+Date: Thu, 1 Dec 2022 15:07:10 +0000
@@ -53,8 +53,6 @@ index 00000000000..b51d7262f4c
+downloaded archive, so hardcode go SDKS list files rather than
+fetch from internet
+
-+Upstream-Status: Inappropriate [Yocto specific]
-+
+Signed-off-by: Hongxu Jia <hongxu.jia@eng.windriver.com>
+---
+ go/private/sdk.bzl | 7 +++++++
diff --git a/recipes-framework/tensorflow/files/0001-support-32-bit-x64-and-arm-for-yocto.patch b/recipes-framework/tensorflow/files/0001-support-32-bit-x64-and-arm-for-yocto.patch
index a855b30..22bd6c1 100644
--- a/recipes-framework/tensorflow/files/0001-support-32-bit-x64-and-arm-for-yocto.patch
+++ b/recipes-framework/tensorflow/files/0001-support-32-bit-x64-and-arm-for-yocto.patch
@@ -56,7 +56,7 @@ new file mode 100644
index 00000000000..2141c37236b
--- /dev/null
+++ b/third_party/0001-XNNPACK-support-32-bit-x86.patch
-@@ -0,0 +1,29 @@
+@@ -0,0 +1,27 @@
+From 2345756776fa09fd95e29f96f781e15e3b18a61b Mon Sep 17 00:00:00 2001
+From: Hongxu Jia <hongxu.jia@windriver.com>
+Date: Tue, 17 Aug 2021 01:09:53 -0700
@@ -64,8 +64,6 @@ index 00000000000..2141c37236b
+
+Use android_x86 as a workaround to support 32 bit x86
+
-+Upstream-Status: Inappropriate [oe specific]
-+
+Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
+---
+ BUILD.bazel | 1 -
@@ -91,7 +89,7 @@ new file mode 100644
index 00000000000..0c83cf01675
--- /dev/null
+++ b/third_party/0001-cpuinfo-support-32-bit-x86.patch
-@@ -0,0 +1,28 @@
+@@ -0,0 +1,27 @@
+From f5e52b2db5516954dab9026e4f386982c4c3c3a4 Mon Sep 17 00:00:00 2001
+From: Hongxu Jia <hongxu.jia@windriver.com>
+Date: Mon, 19 Sep 2022 00:10:18 -0700
@@ -99,7 +97,6 @@ index 00000000000..0c83cf01675
+
+Use android_x86 as a workaround to support 32 bit x86
+
-+Upstream-Status: Inappropriate [oe specific]
+Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
+---
+ BUILD.bazel | 1 -
diff --git a/recipes-framework/tensorflow/files/0001-third_party-eigen_archive-workaround-ice-failure-whi.patch b/recipes-framework/tensorflow/files/0001-third_party-eigen_archive-workaround-ice-failure-whi.patch
index b1571ab..2056c20 100644
--- a/recipes-framework/tensorflow/files/0001-third_party-eigen_archive-workaround-ice-failure-whi.patch
+++ b/recipes-framework/tensorflow/files/0001-third_party-eigen_archive-workaround-ice-failure-whi.patch
@@ -24,7 +24,7 @@ new file mode 100644
index 00000000000..436f35fca13
--- /dev/null
+++ b/third_party/eigen3/0001-workaround-ice-failure-while-building-from-source-wi.patch
-@@ -0,0 +1,69 @@
+@@ -0,0 +1,67 @@
+From 035fd96026e18d61c19da038ab2b0f7a61e97799 Mon Sep 17 00:00:00 2001
+From: Hongxu Jia <hongxu.jia@windriver.com>
+Date: Mon, 19 Sep 2022 01:08:18 -0700
@@ -37,8 +37,6 @@ index 00000000000..436f35fca13
+But we could not patch on native gcc,so try to workaround,
+Refer: https://github.com/tensorflow/tensorflow/issues/25323
+
-+Upstream-Status: Inappropriate [wr-installer specific]
-+
+Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
+---
+ unsupported/Eigen/CXX11/src/Tensor/TensorImagePatch.h | 3 ++-
diff --git a/recipes-framework/tensorflow/files/0001-fix-XNNPACK-build-failure-for-qemuarm.patch b/recipes-framework/tensorflow/files/0019-fix-XNNPACK-build-failure-for-qemuarm.patch
similarity index 97%
rename from recipes-framework/tensorflow/files/0001-fix-XNNPACK-build-failure-for-qemuarm.patch
rename to recipes-framework/tensorflow/files/0019-fix-XNNPACK-build-failure-for-qemuarm.patch
index dea74d9..1a792ee 100644
--- a/recipes-framework/tensorflow/files/0001-fix-XNNPACK-build-failure-for-qemuarm.patch
+++ b/recipes-framework/tensorflow/files/0019-fix-XNNPACK-build-failure-for-qemuarm.patch
@@ -1,4 +1,4 @@
-From c15caa4d5f133307a23114c41a9c497e5ec9da2c Mon Sep 17 00:00:00 2001
+From 2ace7a16a8033169b7ddb71a79956416644ffa3e Mon Sep 17 00:00:00 2001
From: Hongxu Jia <hongxu.jia@windriver.com>
Date: Thu, 23 Mar 2023 10:28:04 +0800
Subject: [PATCH] fix XNNPACK build failure for qemuarm
@@ -13,10 +13,10 @@ Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
1 file changed, 209 insertions(+), 7 deletions(-)
diff --git a/third_party/0001-XNNPACK-support-32-bit-x86.patch b/third_party/0001-XNNPACK-support-32-bit-x86.patch
-index 2141c37236b..ca6a38545b1 100644
+index e2c053e6f8c..89ba615a9fe 100644
--- a/third_party/0001-XNNPACK-support-32-bit-x86.patch
+++ b/third_party/0001-XNNPACK-support-32-bit-x86.patch
-@@ -1,22 +1,36 @@
+@@ -1,20 +1,34 @@
-From 2345756776fa09fd95e29f96f781e15e3b18a61b Mon Sep 17 00:00:00 2001
+From 5ba6883121e91a334cc97ee98eea6ef0afb7a314 Mon Sep 17 00:00:00 2001
From: Hongxu Jia <hongxu.jia@windriver.com>
@@ -28,8 +28,6 @@ index 2141c37236b..ca6a38545b1 100644
+Does not support 32 bit ARM with Neon DotProduct
+
- Upstream-Status: Inappropriate [oe specific]
-
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
---
- BUILD.bazel | 1 -
@@ -59,7 +57,7 @@ index 2141c37236b..ca6a38545b1 100644
config_setting(
name = "android_x86",
values = {
-@@ -24,6 +38,194 @@ index 27ab07c9..f223d2ab 100644
+@@ -22,6 +36,194 @@ index 27ab07c9..f223d2ab 100644
"cpu": "x86",
},
)
@@ -255,6 +253,3 @@ index 2141c37236b..ca6a38545b1 100644
-2.30.2
+2.27.0
---
-2.27.0
-
diff --git a/recipes-framework/tensorflow/tensorflow-lite_2.12.0.bb b/recipes-framework/tensorflow/tensorflow-lite_2.12.0.bb
index 83fd12d..e76a09b 100644
--- a/recipes-framework/tensorflow/tensorflow-lite_2.12.0.bb
+++ b/recipes-framework/tensorflow/tensorflow-lite_2.12.0.bb
@@ -15,7 +15,7 @@ SRC_URI += " \
file://yocto_compiler_configure.bzl \
"
SRC_URI:append:arm = " \
- file://0001-fix-XNNPACK-build-failure-for-qemuarm.patch \
+ file://0019-fix-XNNPACK-build-failure-for-qemuarm.patch \
"
SRC_URI += "https://storage.googleapis.com/download.tensorflow.org/models/inception_v3_2016_08_28_frozen.pb.tar.gz;name=model-inv3"
--
2.48.1
next reply other threads:[~2025-02-11 16:15 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-02-11 16:14 Martin Jansa [this message]
2025-02-11 16:14 ` [meta-tensorflow][mickledore][PATCH 2/3] tensorflow: refresh and rename the .patch files Martin Jansa
2025-02-11 16:14 ` [meta-tensorflow][mickledore][PATCH 3/3] tensorflow: add 0001-fix-python3-include-directory-not-found.patch everywhere Martin Jansa
2025-02-14 3:20 ` [yocto-patches] [meta-tensorflow][mickledore][PATCH 1/3] tensorflow: drop Upstream-Status inside third_party .patch files Hongxu Jia
2025-02-14 7:25 ` Martin Jansa
2025-02-19 2:51 ` Hongxu Jia
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250211161500.3077588-1-martin.jansa@gmail.com \
--to=martin.jansa@gmail.com \
--cc=yocto-patches@lists.yoctoproject.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.