public inbox for igt-dev@lists.freedesktop.org
 help / color / mirror / Atom feed
From: "Knop, Ryszard" <ryszard.knop@intel.com>
To: "Kempczynski, Zbigniew" <zbigniew.kempczynski@intel.com>,
	"igt-dev@lists.freedesktop.org" <igt-dev@lists.freedesktop.org>,
	"Heikkila, Juha-pekka" <juha-pekka.heikkila@intel.com>,
	"Karas, Krzysztof" <krzysztof.karas@intel.com>,
	"Sousa, Gustavo" <gustavo.sousa@intel.com>,
	"kamil.konieczny@linux.intel.com"
	<kamil.konieczny@linux.intel.com>
Subject: Re: [PATCH i-g-t v2 0/6] RFC: Add attachments support
Date: Mon, 30 Mar 2026 13:52:43 +0000	[thread overview]
Message-ID: <c48123c078428b7f630de7e4252a91de353d2093.camel@intel.com> (raw)
In-Reply-To: <20260330131730.jpfzopbtti4cdje5@kamilkon-DESK.igk.intel.com>

On Mon, 2026-03-30 at 15:17 +0200, Kamil Konieczny wrote:
> Hi Knop,,
> On 2026-03-30 at 09:01:39 +0000, Knop, Ryszard wrote:
> > Hey Zbigniew,
> > 
> > In general, I'm not convinced putting all attachments into the
> > results.json file is a good idea. This introduces a significant amount
> > of overhead for CI reporting processes that don't care about those
> > attachments (vis, CBL, JP Heikkila's SISO), makes those files extremely
> > large and difficult for manual inspection.
> > 
> > Can you consider specifying attachments as filenames that should be
> > present in a directory relative to the results.json instead?
> > 
> > Thanks, Ryszard
> 
> Could it be an option --attachments[=log|json] with default log?

IMO there's really no point in embedding these in the JSON. Just keep
them as separate files all the time and keep the whole thing simple. I
get it, 2+GB JSONs are now possible, but this feature is going to cause
those 2+GB files to become the standard once we start attaching stuff.

Ryszard 

> So by default files would only be saved in results/N/attachment
> (where N is a number of test in testlist) and also reported in test
> output (hence the 'log', there could be also 'none' for no report).
> 
> Regards,
> Kamil
> 
> > 
> > On Tue, 2026-03-24 at 14:12 +0100, Zbigniew Kempczyński wrote:
> > > This series introduces support for hook/test written attachments.
> > > It is somehow limited and stiff but allows to write anything test/hook
> > > wants to attachments directory and if naming contract will be fulfiled
> > > content in base64 form will be included in appropriate subtest/dynsubtest
> > > results.json part.
> > > 
> > > Series targets our need to include guc logs for failed tests especially
> > > in the point of time when it occurs.
> > > 
> > > v2: - added clearing of attachments dir in overwrite mode
> > >     - guc-log hook script is now installed to igt datadir/hooks dir
> > >     - added hook-exec-allowlist to selectively execute hook scripts
> > > 
> > > Signed-off-by: Zbigniew Kempczyński <zbigniew.kempczynski@intel.com>
> > > Cc: Kamil Konieczny <kamil.konieczny@linux.intel.com>
> > > Cc: Ryszard Knop <ryszard.knop@intel.com>
> > > Cc: Gustavo Sousa <gustavo.sousa@intel.com>
> > > Cc: Krzysztof Karas <krzysztof.karas@intel.com>
> > > 
> > > Zbigniew Kempczyński (6):
> > >   runner: Rename dirfd to avoid clash with dirfd()
> > >   runner: Create attachments directory to use by hooks
> > >   runner: Add attachments directory content in subtests results
> > >   scripts/hooks: Example guc log copy script to attachments dir
> > >   runner: Rename parsing function
> > >   runner: Add hook-exec-allowlist to execute hooks selectively
> > > 
> > >  lib/igt_hook.c                    |  4 ++
> > >  runner/executor.c                 | 88 ++++++++++++++++++++++++----
> > >  runner/executor.h                 |  2 +
> > >  runner/resultgen.c                | 97 +++++++++++++++++++++++++++++++
> > >  runner/settings.c                 | 31 +++++++---
> > >  runner/settings.h                 |  2 +
> > >  scripts/hooks/guc_copy_on_fail.sh | 24 ++++++++
> > >  scripts/meson.build               |  2 +
> > >  8 files changed, 230 insertions(+), 20 deletions(-)
> > >  create mode 100755 scripts/hooks/guc_copy_on_fail.sh

      reply	other threads:[~2026-03-30 13:52 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-24 13:12 [PATCH i-g-t v2 0/6] RFC: Add attachments support Zbigniew Kempczyński
2026-03-24 13:12 ` [PATCH i-g-t v2 1/6] runner: Rename dirfd to avoid clash with dirfd() Zbigniew Kempczyński
2026-03-24 13:12 ` [PATCH i-g-t v2 2/6] runner: Create attachments directory to use by hooks Zbigniew Kempczyński
2026-03-30  6:48   ` Krzysztof Karas
2026-04-03  5:18     ` Zbigniew Kempczyński
2026-03-30 14:20   ` Gustavo Sousa
2026-03-31 16:05     ` Zbigniew Kempczyński
2026-03-31 16:47       ` Gustavo Sousa
2026-03-24 13:12 ` [PATCH i-g-t v2 3/6] runner: Add attachments directory content in subtests results Zbigniew Kempczyński
2026-03-30  7:17   ` Krzysztof Karas
2026-03-30 15:36   ` Gustavo Sousa
2026-03-24 13:12 ` [PATCH i-g-t v2 4/6] scripts/hooks: Example guc log copy script to attachments dir Zbigniew Kempczyński
2026-03-30  7:24   ` Krzysztof Karas
2026-03-24 13:12 ` [PATCH i-g-t v2 5/6] runner: Rename parsing function Zbigniew Kempczyński
2026-03-30  8:07   ` Krzysztof Karas
2026-04-03  5:40     ` Zbigniew Kempczyński
2026-03-24 13:12 ` [PATCH i-g-t v2 6/6] runner: Add hook-exec-allowlist to execute hooks selectively Zbigniew Kempczyński
2026-03-30  8:28   ` Krzysztof Karas
2026-03-30 17:19   ` Gustavo Sousa
2026-04-03  5:55     ` Zbigniew Kempczyński
2026-03-24 19:48 ` ✓ Xe.CI.BAT: success for RFC: Add attachments support (rev2) Patchwork
2026-03-24 21:10 ` ✓ i915.CI.BAT: " Patchwork
2026-03-25  7:25 ` ✗ i915.CI.Full: failure " Patchwork
2026-03-25 10:09 ` ✓ Xe.CI.FULL: success " Patchwork
2026-03-30  9:01 ` [PATCH i-g-t v2 0/6] RFC: Add attachments support Knop, Ryszard
2026-03-30 13:17   ` Kamil Konieczny
2026-03-30 13:52     ` Knop, Ryszard [this message]

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=c48123c078428b7f630de7e4252a91de353d2093.camel@intel.com \
    --to=ryszard.knop@intel.com \
    --cc=gustavo.sousa@intel.com \
    --cc=igt-dev@lists.freedesktop.org \
    --cc=juha-pekka.heikkila@intel.com \
    --cc=kamil.konieczny@linux.intel.com \
    --cc=krzysztof.karas@intel.com \
    --cc=zbigniew.kempczynski@intel.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox