lwip2: lighter implementation #3362

d-a-v · 2017-06-19T12:37:11Z

This PR is replaced by #3783

full sources are separate and available as a submodule
included in this PR: include/ and liblwip2.a + Makefile for pull&rebuild

I guess this is the way @igrr imports his cloned axtls, so arduino/esp8266 is lighter.

aside from this lwip2 implementation gets more genericity for other esp8266 projects.

devyte · 2017-09-09T05:54:57Z

@igrr are we waiting until after the next release to merge this, or do we want it in the next release?

igrr · 2017-09-12T13:54:40Z

Yes, we're waiting until after the next release.

@earlephilhower

Includes two PRs: - igrr/axtls-8266#46 by @earlephilhower: Move debug strings from RAM to Flash - igrr/axtls-8266#50: Fix memory leak in ssl_ext_host_name

Every assert() includes a __FILE__ constant string to RODATA which can be quite large as it includes the complete path as well as the filename. Move that string into PMEM, and update the postmortem to work with either PMEM or RAM strings for dumping abort/assert/exception information.

The ax_port_malloc, ax_port_calloc, ax_port_realloc, and ax_port_free functions in WiFiClientSecure are not actually used by the AXTLS library. It's directly using the library routines, and these function are never used. Remove these dead bits of code to make the axtls operation clearer.

Custom manifest for @platformio which instructs PIO Build System to do not archive library's object files. See - esp8266#3321 (comment) - http://docs.platformio.org/en/latest/librarymanager/config.html#libarchive

UMM debugging strings are normally placed in RODATA, which uses up scarse memory. Move them to PROGMEM and use macros to replace printf with a version that can handle ROM strings.

The extension -> MIME type routine uses lots of constant strings which end up in the RODATA segment of RAM. Refactor the comparison to use a table of strings stored in PMMEM instead, freeing ~370 bytes for the heap.

…8266#3398)

The Arduino IDE requires the use of a tab separator between the name and identifier. Without this tab the keyword is not highlighted. Reference: https://github.com/arduino/Arduino/wiki/Arduino-IDE-1.5:-Library-specification#keywords

The RST spec does not appear to believe in intra-cell line continuation as appeared to be expected by the original author. Widen columns and unbreak words to fix output.

realloc() is called with newSize > 0 (at least 16), so newbuffer==0 means the old memory was not deallocated. Therefore, the pointer should still point to the old buffer. This change should resolve issue esp8266#3516.

as per https://tools.ietf.org/html/rfc2046#section-5.1.1

esp8266#3592

d-a-v · 2017-10-25T21:25:06Z

ram + flash footprint is now better than with lwip1.4.
Below the footprint variation with one of my sketches: The two last columns show the latest commits benefits.
mainly: I had forgotten to disable lwip internal asserts :-/

                1.4             v2              -assert         optsync+dns
sketch          373912          380616          367684          367236
sketch%         74              76              73              73
ram             40396           47940           40932           40020
ram%            49              58              49              48
heap            33240           25360           32872           33280

…t (get it from arduino's sdk lwip1.4)

…266#3701

SoftwareSerial: update to 3.3.1

…266#2557)

espressif/ESP8266_NONOS_SDK@33f234f

d-a-v · 2017-11-01T13:01:20Z

@igrr @devyte @davisonja At this point, and

because people are beginning to be aware that lwip2 can solve some old issues and some of them already use it,
because this PR cannot not be merged as-is (gitk shows real tube map due to repaired mistakes and also to my inability to stay clean in the long run with git and the multiple branches I'm dealing with - I lost liblwip2.a the other day, and I regularly lose myself in suddenly appearing head-detached branch :),
because this PR is really small (it's one makefile, one line in the main ldscript, the include directory+lib which are however auto-generated, and a submodule ref),
because memory footprint seems equivalent to lwip1.4's,
because lwip2 needs more serious testing
because I don't want to and will not advertise my own repo

What do you think if

I include in here the regenerated boards.txt from boards.txt generator #3722 so it is tested too (and has lwip2+debug for all known boards)
lwip2 is advertised in issues where relevant with an easy step-by-step setup doc
and of course I keep up syncing with master so testers can always have master+lwip2

igrr · 2017-11-01T13:03:31Z

That sounds very good to me @d-a-v. Let's add it to master to make things easier in terms of testing and updates.

…

On Wed, Nov 1, 2017, 21:01 david gauchard ***@***.***> wrote: @igrr <https://github.com/igrr> @devyte <https://github.com/devyte> @davisonja <https://github.com/davisonja> At this point, and - because people are beginning to be aware that lwip2 can solve some old issues and some of them already use it, - because this PR cannot not be merged as-is (gitk shows real tube map due to repaired mistakes and also to my inability to stay clean in the long run with git and the multiple branches I'm dealing with - I lost liblwip2.a the other day, and I regularly lose myself in suddenly appearing head-detached branch :), - because this PR is really small (it's one makefile, one line in the main ldscript, the include directory+lib which are however auto-generated, and a submodule ref), - because memory footprint seems equivalent to lwip1.4's, - because lwip2 needs more serious testing - because I don't want to and will not advertise my own repo What do you think if - I include in here the regenerated boards.txt from #3722 <#3722> so it is tested too (and has lwip2+debug for all known boards) - lwip2 is advertised in issues where relevant with an easy step-by-step setup doc - and of course I keep up syncing with master so testers can always have master+lwip2 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3362 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEJcepUlCDfMH_fUCnb5cFNH1lvYb5HAks5syGupgaJpZM4N-Kxb> .

d-a-v · 2017-11-01T13:11:36Z

To master, I was not asking for that much :) Did I understand well ?
The boards.txt issue must be solved first so lwip2 can be available in menus, but still along with the former one that we must have around. @davisonja what about #3722 and ldscripts ?
Or should we provide a generated boards.txt at first ?
Or simply keep the modified master's boards.txt from this PR ?

davisonja · 2017-11-01T23:06:58Z

Because the ldscripts need to match I would vote to include a generated boards.txt, though in practice since it should look the same as the modified version in this pr its probably just as simple to leave the current one. I've got some queries about the current script, but been tied up with new child, hopefully can post them tonight 😊

devyte · 2017-11-02T04:44:43Z

@d-a-v if this PR is superseded, should it be closed?
Similar question for PR #3206

d-a-v · 2017-11-02T08:37:57Z

closing in favour of #3783

igrr · 2017-11-02T11:47:11Z

@d-a-v I did mean master. Since lwip v2 is optional, it shouldn't be any harm, and hopefully will help folks who run into lwip related issues.

d-a-v added 5 commits June 17, 2017 03:21

add lwip2 as a submodule

aeab73a

makefile for lwip2 submodule

6c4f5cc

fix lwip2 builder makefile

c289363

+ precompiled lwip2 library with includes

48b7fe2

upstream fixes

602c1b5

d-a-v mentioned this pull request Jul 15, 2017

LWIP memory leak? #3419

Closed

d-a-v mentioned this pull request Aug 30, 2017

open source lwip2 #3206

Closed

d-a-v mentioned this pull request Sep 12, 2017

New Simple Easily Reproducable Catastrophic TCP Stack Failure Results in Low Heap Crash shown with WiFiServer, Latency > 100ms, and Larger Send Bytes #2925

Closed

igrr and others added 20 commits September 29, 2017 11:31

Update axTLS to 5de79d71.

73f4307

Includes two PRs: - igrr/axtls-8266#46 by @earlephilhower: Move debug strings from RAM to Flash - igrr/axtls-8266#50: Fix memory leak in ssl_ext_host_name

Fixed: urlDecode done before parsing args (esp8266#2956)

a377aba

Replace r_rand with os_random for LWIP_RAND() (esp8266#3499)

bbd6f75

Manifest for @platformio

a6cca4f

Custom manifest for @platformio which instructs PIO Build System to do not archive library's object files. See - esp8266#3321 (comment) - http://docs.platformio.org/en/latest/librarymanager/config.html#libarchive

Integration with @platformio Build System

157aa31

Move debug constant strings to PROGMEM (esp8266#3478)

71495e2

UMM debugging strings are normally placed in RODATA, which uses up scarse memory. Move them to PROGMEM and use macros to replace printf with a version that can handle ROM strings.

Move MIME type table into PROGMEM to save RAM (esp8266#3475)

56cd198

The extension -> MIME type routine uses lots of constant strings which end up in the RODATA segment of RAM. Refactor the comparison to use a table of strings stored in PMMEM instead, freeing ~370 bytes for the heap.

Removed literal '\n' in ESP8266HTTPUpdateServer response (esp8266#3421)

1d79311

Fix to parse parameters in the URL of a POST with empty content. (esp…

5764446

…8266#3398)

patch for esp8266#3462

d68c484

lwip2v2: merge upstream

6499933

Functional Interrupts initial

0a5d445

Use correct separator in keywords.txt

d8f9956

The Arduino IDE requires the use of a tab separator between the name and identifier. Without this tab the keyword is not highlighted. Reference: https://github.com/arduino/Arduino/wiki/Arduino-IDE-1.5:-Library-specification#keywords

Fix exception_causes.rst formatting

92978a7

The RST spec does not appear to believe in intra-cell line continuation as appeared to be expected by the original author. Widen columns and unbreak words to fix output.

Update WString.cpp

47baf29

realloc() is called with newSize > 0 (at least 16), so newbuffer==0 means the old memory was not deallocated. Therefore, the pointer should still point to the old buffer. This change should resolve issue esp8266#3516.

Update spiffs 0.3.7 to and mkspiffs 0.1.3

6e883e2

Allow for double quotes in boundary (esp8266#3455)

e0e3a77

as per https://tools.ietf.org/html/rfc2046#section-5.1.1

link airkiss library

36f09dc

esp8266#3592

devyte mentioned this pull request Oct 29, 2017

DHCP Transaction Id not random #3773

Closed

d-a-v and others added 13 commits October 30, 2017 22:13

update lwip2 include files (now stable-2.0.3) + fix typedef lwip_err_…

f2d573f

…t (get it from arduino's sdk lwip1.4)

Minimal file with a few ESP8266-specific keywords - github issue esp8…

7023a65

…266#3701

Update SoftwareSerial to 3.3.0 (esp8266#3727)

8d5280e

SoftwareSerial: update to 3.3.1

Increase version for PlatformIO (esp8266#3742)

4ab9766

Put WiFi into station mode in examples (esp8266#3731)

0b3340f

Make DNSServer debug output like other ESP libs

bcb534c

Make ESP8266mDNS debug output like other ESP libs

48f0381

allocate HTTPUpload struct on demand (esp8266#2557)

da872a9

minimize number of exit paths in ESP8266WebServer::handleClient (esp8…

2b5a977

…266#2557)

sdk: update to v2.1.0-14-g33f234f

7fa1286

espressif/ESP8266_NONOS_SDK@33f234f

Merge branch 'master' into lwip2v2

fb52539

Merge branch 'master' of https://github.com/esp8266/Arduino into lwip2v2

928abeb

restore boards.txt

d3139b6

dragondaud mentioned this pull request Oct 31, 2017

configTime ignores timezone in rc2 #3774

Closed

d-a-v added 3 commits October 31, 2017 21:24

restore liblwip2.a

286b04b

update lib: upstream fix timezone handling

3362f4b

sync refs with upstream/linklayer

448867c

d-a-v closed this Nov 2, 2017

d-a-v deleted the lwip2v2 branch November 3, 2017 08:26

s0170071 mentioned this pull request Jan 13, 2018

preparations for lwip2 letscontrolit/ESPEasy#707

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

lwip2: lighter implementation #3362

lwip2: lighter implementation #3362

d-a-v commented Jun 19, 2017 •

edited

Loading

devyte commented Sep 9, 2017

igrr commented Sep 12, 2017

d-a-v commented Oct 25, 2017 •

edited

Loading

d-a-v commented Nov 1, 2017

igrr commented Nov 1, 2017 via email

d-a-v commented Nov 1, 2017

davisonja commented Nov 1, 2017

devyte commented Nov 2, 2017

d-a-v commented Nov 2, 2017

igrr commented Nov 2, 2017

lwip2: lighter implementation #3362

lwip2: lighter implementation #3362

Conversation

d-a-v commented Jun 19, 2017 • edited Loading

This PR is replaced by #3783

devyte commented Sep 9, 2017

igrr commented Sep 12, 2017

d-a-v commented Oct 25, 2017 • edited Loading

d-a-v commented Nov 1, 2017

igrr commented Nov 1, 2017 via email

d-a-v commented Nov 1, 2017

davisonja commented Nov 1, 2017

devyte commented Nov 2, 2017

d-a-v commented Nov 2, 2017

igrr commented Nov 2, 2017

d-a-v commented Jun 19, 2017 •

edited

Loading

d-a-v commented Oct 25, 2017 •

edited

Loading