Few small fixes to allow envoy to compile on ARM32 by costinm · Pull Request #1781 · envoyproxy/envoy

costinm · 2017-09-29T16:40:44Z

Tested compilation for rasbperry pi and android (the 'endian' change is
required for android NDK - recent androids can compile with 64 bit).

Signed-off-by: Costin Manolache costin@gmail.com

costinm · 2017-09-29T16:41:34Z

Note that I compiled/tested using a custom cmake, not bazel.

htuch

Looks alright to me.

htuch · 2017-09-29T16:44:01Z

source/server/options_impl.cc


  TCLAP::CmdLine cmd("envoy", ' ', VersionInfo::version());
-  TCLAP::ValueArg<uint64_t> base_id(
+  TCLAP::ValueArg<uint32_t> base_id(


Why are these necessary?

I get this error with the pi cross-compiler. Not sure if there is some other fix, not an expert in c++...

In file included from /ws/istio-master/src/tclap/include/tclap/Arg.h:54:0,
from /ws/istio-master/src/tclap/include/tclap/SwitchArg.h:30,
from /ws/istio-master/src/tclap/include/tclap/CmdLine.h:27,
from /ws/istio-master/envoy/source/server/options_impl.cc:13:
/ws/istio-master/src/tclap/include/tclap/ArgTraits.h: In instantiation of 'struct TCLAP::ArgTraits':
/ws/istio-master/src/tclap/include/tclap/ValueArg.h:403:66: required from 'void TCLAP::ValueArg::_extractValue(const string&) [with T = long long unsigned int; std::string = std::basic_string]'
/ws/istio-master/src/tclap/include/tclap/ValueArg.h:363:29: required from 'bool TCLAP::ValueArg::processArg(int*, std::vector<std::basic_string >&) [with T = long long unsigned int]'
/ws/istio-master/envoy/source/server/options_impl.cc:120:1: required from here
/ws/istio-master/src/tclap/include/tclap/ArgTraits.h:80:39: error: 'long long unsigned int' is not a class, struct, or union type
typedef typename T::ValueCategory ValueCategory;

Also - I don't think any of the options would need 64 bit parameters.

That's with arm-rpi-4.9.3-linux-gnueabihf

Might be an issue in converting from the literal to the type. I think in general, for clean 32/64-bit support, we should have some macros for writing fixed width literals, e.g. CONST64(0), since we're going to hit this elswhere. I do agree that these are fine as uint32_t though.

Seems specific to tclap macros - so far everything else seems to work on 32 bit arm, I'll do more testing and try to get it into a continuous build.

mattklein123 · 2017-09-30T01:57:25Z

In general this seems fine to me but please fix DCO and format.

dnoe · 2017-10-02T13:32:52Z

include/envoy/http/codec.h

  static const uint32_t DEFAULT_HPACK_TABLE_SIZE = (1 << 12);
  // no maximum from HTTP/2 spec, use unsigned 32-bit maximum
-  static const uint32_t MAX_HPACK_TABLE_SIZE = (1UL << 32) - 1;
+  static const uint32_t MAX_HPACK_TABLE_SIZE = 0xFFFFFFFF;


You should be able to use the C++ numeric_limits for this:

#include <limits> static const uint32_t MAX_HPACK_TABLE_SIZE = std::numeric_limits<uint32_t>::max();

Signed-off-by: Matt Klein <mklein@lyft.com>

Signed-off-by: Stephan Zuercher <stephan@turbinelabs.io>

) The v2 API supports using the downstream (source) IP address to route HTTP requests to upstream hosts. This PR implements that functionality, as specified in envoyproxy#1296. Signed-off-by: Alex Konradi <akonradi@google.com>

…1753) Generalizes existing zone support. Fixes envoyproxy#1750. Signed-off-by: Harvey Tuch htuch@google.com

…yproxy#1757) Signed-off-by: Matt Klein <mklein@lyft.com>

…ion (envoyproxy#1733) resolves envoyproxy#1711 Signed-off-by: Gary Brown <gary@brownuk.com>

Signed-off-by: Daniel Hochman <dhochman@lyft.com>

Fixes envoyproxy#1422 Signed-off-by: Alex Konradi <akonradi@google.com>

Signed-off-by: Matt Klein <mklein@lyft.com>

Signed-off-by: Gabriel Rosenhouse <grosenhouse@pivotal.io>

Signed-off-by: Matt Rice <mattrice@google.com>

There is no minimal language and due to travis-ci/travis-ci#4895, it will fallback to ruby. shell should serve our usecase Signed-off-by: bndw <benjamindwoodward@gmail.com>

Signed-off-by: bndw <benjamindwoodward@gmail.com>

Signed-off-by: Daniel Hochman <dhochman@lyft.com>

Signed-off-by: Matt Klein <mklein@lyft.com>

Signed-off-by: bndw <benjamindwoodward@gmail.com>

Signed-off-by: Stephan Zuercher <stephan@turbinelabs.io>

Tested compilation for rasbperry pi and android (the 'endian' change is required for android NDK - recent androids can compile with 64 bit).

Signed-off-by: Costin Manolache <costin@google.com>

This PR adds per-host counters to support UpstreamLocalityStats. At load report time, these will be aggregated on a per-locality basis. Signed-off-by: Harvey Tuch htuch@google.com

Signed-off-by: Piotr Sikora <piotrsikora@google.com>

…nvoyproxy#1776) Signed-off-by: Greg Greenway <ggreenway@apple.com>

Signed-off-by: Stephan Zuercher <stephan@turbinelabs.io>

…yproxy#1786) This is helpful for Google import and shouldn't affect anybody else. Signed-off-by: Piotr Sikora <piotrsikora@google.com>

See ci/README.md for instructions. Signed-off-by: David Mackey tdmackey@booleanhaiku.com

…envoyproxy#1790) Signed-off-by: Harvey Tuch <htuch@google.com>

This makes it appopriate to use the OsSysCalls abstraction in other areas of the code-base, such as hot-restart. Signed-off-by: Greg Greenway <ggreenway@apple.com>

Tested compilation for rasbperry pi and android (the 'endian' change is required for android NDK - recent androids can compile with 64 bit). Signed-off-by: Costin Manolache <costin@google.com>

costinm · 2017-10-02T22:21:47Z

Git mixups attempting to edit commits to add DCO - will upload a clean PR

Description: Move tsan workflow to Engflow's remote execution Same as envoyproxy/envoy-mobile#1745 but using a couple of tsan flags instead of asan. Using a separate config for modularity. Risk Level: Low Testing: See tsan workflow Docs Changes: N/A Release Notes: N/A Signed-off-by: Luis Fernando Pino Duque <luis@engflow.com> Signed-off-by: JP Simard <jp@jpsim.com>

Signed-off-by: fbalicchia <fbalicchia@cuebiq.com>

htuch reviewed Sep 29, 2017

View reviewed changes

dnoe reviewed Oct 2, 2017

View reviewed changes

mattklein123 and others added 26 commits October 2, 2017 15:11

ci: fix coverage and alpine image (envoyproxy#1749)

df11fbd

Signed-off-by: Matt Klein <mklein@lyft.com>

mac-ci: add a CI workflow for Mac OS builds (envoyproxy#1748)

fe00e48

Signed-off-by: Stephan Zuercher <stephan@turbinelabs.io>

host: plumb envoy::api::v2::Locality to HostDescription. (envoyproxy#…

ce0d755

…1753) Generalizes existing zone support. Fixes envoyproxy#1750. Signed-off-by: Harvey Tuch htuch@google.com

ci: disable OSX due to needing to sort out Circle config issues (envo…

f8fd341

…yproxy#1757) Signed-off-by: Matt Klein <mklein@lyft.com>

tracing: Update docs to clarify situation with trace context propagat…

dfd99ad

…ion (envoyproxy#1733) resolves envoyproxy#1711 Signed-off-by: Gary Brown <gary@brownuk.com>

NVLOG to prevent unused var (envoyproxy#1758)

6babcab

Signed-off-by: Daniel Hochman <dhochman@lyft.com>

Support multiple hash policies for HTTP routing (envoyproxy#1756)

21611cc

Fixes envoyproxy#1422 Signed-off-by: Alex Konradi <akonradi@google.com>

Update README.md (envoyproxy#1760)

4b3048c

Signed-off-by: Matt Klein <mklein@lyft.com>

docs: fix up language in websocket arch overview (envoyproxy#1765)

1afb867

Signed-off-by: Gabriel Rosenhouse <grosenhouse@pivotal.io>

Fix up doc comments in hot_restart.h (envoyproxy#1763)

36c3497

Signed-off-by: Gabriel Rosenhouse <grosenhouse@pivotal.io>

whitespace addition to kick new build (envoyproxy#1770)

ba9af7d

Signed-off-by: Matt Rice <mattrice@google.com>

Fix bazel build against release 0.6.0 for Ubuntu. (envoyproxy#1769)

3de1399

Signed-off-by: Matt Rice <mattrice@google.com>

Change .travis.yml's language: minimal to shell. (envoyproxy#1771)

2204810

There is no minimal language and due to travis-ci/travis-ci#4895, it will fallback to ruby. shell should serve our usecase Signed-off-by: bndw <benjamindwoodward@gmail.com>

CI: Publish tags to Dockerhub (envoyproxy#1768)

6be9a13

Signed-off-by: bndw <benjamindwoodward@gmail.com>

redis: fix pipeline timeout logic (envoyproxy#1773)

7d90c20

Signed-off-by: Daniel Hochman <dhochman@lyft.com>

optional: make coverity happy (envoyproxy#1767)

e12831d

Signed-off-by: Matt Klein <mklein@lyft.com>

Fixes to Dockerhub release tagging (envoyproxy#1777)

5e6b5b0

Signed-off-by: bndw <benjamindwoodward@gmail.com>

build: fix macOS build (requires bazel 0.6.0) (envoyproxy#1779)

20ac63e

Signed-off-by: Stephan Zuercher <stephan@turbinelabs.io>

Few small fixes to allow envoy to compile on ARM32

4e5aeab

Tested compilation for rasbperry pi and android (the 'endian' change is required for android NDK - recent androids can compile with 64 bit).

Fix header order

a28d5d3

Signed-off-by: Costin Manolache <costin@google.com>

host: track per-host UpstreamLocalityStats. (envoyproxy#1755)

09305c1

This PR adds per-host counters to support UpstreamLocalityStats. At load report time, these will be aggregated on a per-locality basis. Signed-off-by: Harvey Tuch htuch@google.com

router: initialize RuntimeData fields. (envoyproxy#1787)

26410e5

Signed-off-by: Piotr Sikora <piotrsikora@google.com>

Fix race condition in TEST(FileSystemImpl, flushToLogFileOnDemand) (e…

8ccabbe

…nvoyproxy#1776) Signed-off-by: Greg Greenway <ggreenway@apple.com>

macos: tweak boringssl build config for xcode 9 (envoyproxy#1785)

cfbedb4

Signed-off-by: Stephan Zuercher <stephan@turbinelabs.io>

config: few changes to help with protobuf string type mismatch. (envo…

027ac58

…yproxy#1786) This is helpful for Google import and shouldn't affect anybody else. Signed-off-by: Piotr Sikora <piotrsikora@google.com>

tdmackey and others added 4 commits October 2, 2017 15:11

ci: add local Coverity Scan build option (envoyproxy#1775)

b624efb

See ci/README.md for instructions. Signed-off-by: David Mackey tdmackey@booleanhaiku.com

test: convert //test/integration:ratelimit_integration_test to proto. (…

267931f

…envoyproxy#1790) Signed-off-by: Harvey Tuch <htuch@google.com>

Move Filesystem::OsSysCalls to Api::OsSysCalls. (envoyproxy#1791)

fddac6f

This makes it appopriate to use the OsSysCalls abstraction in other areas of the code-base, such as hot-restart. Signed-off-by: Greg Greenway <ggreenway@apple.com>

Few small fixes to allow envoy to compile on ARM32

bc5706b

Tested compilation for rasbperry pi and android (the 'endian' change is required for android NDK - recent androids can compile with 64 bit). Signed-off-by: Costin Manolache <costin@google.com>

costinm force-pushed the istio32 branch from f185c81 to bc5706b Compare October 2, 2017 22:18

Merge branch 'master' of https://github.com/envoyproxy/envoy into fix

065942d

costinm closed this Oct 2, 2017

costinm mentioned this pull request Oct 2, 2017

Few small fixes to allow envoy to compile on ARM32 #1795

Merged

gizmochief7 mentioned this pull request Oct 16, 2017

Provide an official ARM docker image and/or provide documentation for building on ARM #1861

Closed

mathetake pushed a commit that referenced this pull request Mar 3, 2026

test: add token usage in testupstream fake response (#1781)

a3a21f6

Signed-off-by: fbalicchia <fbalicchia@cuebiq.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Few small fixes to allow envoy to compile on ARM32#1781

Few small fixes to allow envoy to compile on ARM32#1781
costinm wants to merge 31 commits intoenvoyproxy:masterfrom
costinm:istio32

costinm commented Sep 29, 2017 •

edited

Loading

Uh oh!

costinm commented Sep 29, 2017

Uh oh!

htuch left a comment

Uh oh!

htuch Sep 29, 2017

Uh oh!

costinm Sep 29, 2017

Uh oh!

costinm Sep 29, 2017

Uh oh!

costinm Sep 29, 2017

Uh oh!

htuch Sep 29, 2017

Uh oh!

costinm Oct 2, 2017

Uh oh!

mattklein123 commented Sep 30, 2017

Uh oh!

dnoe Oct 2, 2017

Uh oh!

costinm Oct 2, 2017

Uh oh!

costinm commented Oct 2, 2017

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

14 participants

Conversation

costinm commented Sep 29, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

costinm commented Sep 29, 2017

Uh oh!

htuch left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mattklein123 commented Sep 30, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

costinm commented Oct 2, 2017

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

14 participants

costinm commented Sep 29, 2017 •

edited

Loading