Wireshark-dev: Re: [Wireshark-dev] Failed pipeline for nvmeof_getlog_page | wireshark | 3a8e09e
Date Index · Thread Index · Other Months · All Mailing Lists
Date Prev · Date Next · Thread Prev · Thread Next
From: ronnie sahlberg <ronniesahlberg@xxxxxxxxx>
Date: Wed, 31 Mar 2021 17:41:03 +1000
It seems like you already know what is wrong and how to solve this.
Why not just change your patch so it does not trigger this compiler error/warning?


Maybe the compiler is overly sensitive here? Who knows.
Why is this a problem for wireshark developers?

Do this:
* fix your code so it does not trigger the warnings.  Wireshark developers are not supposed to
keep track of what are false warnings that are bogus or what are real problems.
* if you have a problem with the compiler, bring it up with them. we are not compiler developers or maintainers.
* work on your people skills. You are rubbing people that could help you in seriosly the wrong way.


On Wed, Mar 31, 2021 at 5:21 PM Constantine Gavrilov <CONSTG@xxxxxxxxxx> wrote:
I have a question regarding for a special form of automatic builds that I do not understand.

Occasionally, I get an email for additional pipeline build that is different from default pipeline linked to the ticket. It is a wider set of compilers and distributions.

Several questions:
  1. What is the significance of this and when it is triggered?
  2. Why it is not linked to MR (meaning I cannot see this failure in MR)?
  3. Shall I fix these failures?
  4. How do I know that the issue is fixed, since such builds are not linked to MR?
I have looked at this particular one, and it is a a bug in compiler:

guint64 off =...; /* take from command context, now looking at reply */
if (off < 40)
        proto_add_item(...., .... 40-off,.....);

So, the error (in CLANG-11) is (40-off) is 64-bit and passing it as 32-bit parameter "looses high-order bits".

First, the compiler shall see that no loss of value takes place because of the "IF" statement here. Second, since when passing 64-bit value as a 32-bit parameter shall be a compiler error in C language?

I can easily fix this (check the value in saved context, and if it is above possible payload length return, then declare off as 32-bit), but I need to know if CLANG-11 (with draconian compile options) is a MUST to pass.


-
----------------------------------------
Constantine Gavrilov
Storage Architect
Master Inventor
Tel-Aviv Storage Lab IDT Lead
Tel-Aviv IBM Storage Lab
1 Azrieli Center, Tel-Aviv
----------------------------------------
----- Forwarded by Constantine Gavrilov/Israel/IBM on 03/31/2021 10:07 AM -----

From:        GitLab <gitlab@xxxxxxxxxxxxx>
To:        
Date:        03/30/2021 05:37 PM
Subject:        [EXTERNAL] Failed pipeline for nvmeof_getlog_page | wireshark | 3a8e09ef
Sent by:        gitlab@xxxxxxxxxxxxx


Pipeline #278885577 has failed!   Project Constantine Gavrilov / wireshark Branch nvmeof_getlog_page Commit ‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍ZjQcmQRYFpfptBannerStart
This Message Is From an External Sender
This message came from outside your organization.
ZjQcmQRYFpfptBannerEnd
Pipeline #278885577 has failed!

 
Project
Branch

Commit
NVMe: Get LogPage: Commands Supported and Effects
Commit Author


 
Pipeline #278885577 triggered by

had 1 failed build.
Failed builds
build





You're receiving this email because of your account on gitlab.com. Manage all notifications · Help



___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:wireshark-dev-request@xxxxxxxxxxxxx?subject=unsubscribe