From: Jesse Brandeburg <jesse.brandeburg@intel.com>
To: Przemek Kitszel <przemyslaw.kitszel@intel.com>,
Julia Lawall <Julia.Lawall@lip6.fr>, <cocci@inria.fr>
Cc: Jacob Keller <jacob.e.keller@intel.com>
Subject: Re: [cocci] [BUG] exponentially slow spatch
Date: Tue, 21 Nov 2023 12:45:47 -0800 [thread overview]
Message-ID: <e7263df7-be7d-47a4-b708-e04393f7887a@intel.com> (raw)
In-Reply-To: <20231109161846.148524-1-przemyslaw.kitszel@intel.com>
On 11/9/2023 8:18 AM, Przemek Kitszel wrote:
> Hi Julia,
>
> we were bothered by, what we have called a "flaky spatch" issues
> in our CI servers, which have manifested as spatch "just exiting",
> without any proper exit codes.
> Jake & Jesse found out that processing large files (like 100kloc
> amalgamate builds...) or parallel processing makes it more common,
> so we have disabled those.
> Now I see that those builds were likely killed (or spatch "crashes"?).
>
> The issue hit me again, so I dig deeper, and the case is that processing
> time grows exponentially for seemingly easy case, attached.
>
> I'm on github provided sources, as of commit d235ef51f6be ("in modern versions
> of Linux, PAGE_ALIGNED looks like an ordinary macro call"), running
> Rhel 9.3 Beta.
>
> Przemek
Adding the cocci list since it's been a couple of weeks, if you want me
to bounce the original mail with the patches to the list please let me know.
> ---
> small.cocci | 8 ++++++++
> y.c | 53 +++++++++++++++++++++++++++++++++++++++++++++++++++++
> 2 files changed, 61 insertions(+)
> create mode 100644 small.cocci
> create mode 100644 y.c
>
> diff --git a/small.cocci b/small.cocci
> new file mode 100644
> index 000000000000..6a6cee97e21f
> --- /dev/null
> +++ b/small.cocci
> @@ -0,0 +1,8 @@
> +@@
> +expression list E;
> +expression hw;
> +@@
> +(
> +- ice_info(hw,E);
> ++ dev_info(ice_hw_to_dev(hw),E);
> +)
> diff --git a/y.c b/y.c
> new file mode 100644
> index 000000000000..560ddc26430b
> --- /dev/null
> +++ b/y.c
> @@ -0,0 +1,53 @@
> +/**
> + * ice_dump_common_caps - print struct ice_hw_common_caps fields
> + * @hw: pointer to the ice_hw instance
> + * @caps: pointer to common caps instance
> + * @prefix: string to prefix when printing
> + */
> +static void
> +ice_dump_common_caps(struct ice_hw *hw, struct ice_hw_common_caps *caps,
> + char const *prefix)
> +{
> + ice_info(hw, "%s: switching_mode = %d\n", prefix, caps->switching_mode);
> + ice_info(hw, "%s: mgmt_mode = %d\n", prefix, caps->mgmt_mode);
> + ice_info(hw, "%s: mgmt_protocols_mctp = %d\n", prefix,
> + caps->mgmt_protocols_mctp);
> + ice_info(hw, "%s: os2bmc = %d\n", prefix, caps->os2bmc);
> + ice_info(hw, "%s: valid_functions (bitmap) = %d\n", prefix,
> + caps->valid_functions);
> + ice_info(hw, "%s: sr_iov_1_1 = %d\n", prefix, caps->sr_iov_1_1);
> + ice_info(hw, "%s: vmdq = %d\n", prefix, caps->vmdq);
> + ice_info(hw, "%s: evb_802_1_qbg = %d\n", prefix, caps->evb_802_1_qbg);
> + ice_info(hw, "%s: evb_802_1_qbh = %d\n", prefix, caps->evb_802_1_qbh);
> + ice_info(hw, "%s: dcb = %d\n", prefix, caps->dcb);
> + ice_info(hw, "%s: active_tc_bitmap = %d\n", prefix,
> + caps->active_tc_bitmap);
> + ice_info(hw, "%s: maxtc = %d\n", prefix, caps->maxtc);
> + ice_info(hw, "%s: iscsi = %d\n", prefix, caps->iscsi);
> + ice_info(hw, "%s: rss_table_size = %d\n", prefix, caps->rss_table_size);
> + ice_info(hw, "%s: rss_table_entry_width = %d\n", prefix,
> + caps->rss_table_entry_width);
> + ice_info(hw, "%s: num_rxq = %d\n", prefix, caps->num_rxq);
> + ice_info(hw, "%s: rxq_first_id = %d\n", prefix, caps->rxq_first_id);
> + ice_info(hw, "%s: num_txq = %d\n", prefix, caps->num_txq);
> + ice_info(hw, "%s: txq_first_id = %d\n", prefix, caps->txq_first_id);
> + ice_info(hw, "%s: num_msix_vectors = %d\n", prefix,
> + caps->num_msix_vectors);
> + ice_info(hw, "%s: msix_vector_first_id = %d\n", prefix,
> + caps->msix_vector_first_id);
> + ice_info(hw, "%s: ieee_1588 = %d\n", prefix, caps->ieee_1588);
> + ice_info(hw, "%s: mgmt_cem = %d\n", prefix, caps->mgmt_cem);
> + ice_info(hw, "%s: iwarp = %d\n", prefix, caps->iwarp);
> + ice_info(hw, "%s: roce_lag = %d\n", prefix, caps->roce_lag);
> +} /* This takes 27 seconds as of now, doubles with each ice_info() call added
> +
> + ice_info(hw, "%s: wr_csr_prot = 0x%llX\n", prefix,
> + (unsigned long long)caps->wr_csr_prot);
> + ice_info(hw, "%s: num_wol_proxy_fltr = %d\n", prefix,
> + caps->num_wol_proxy_fltr);
> + ice_info(hw, "%s: wol_proxy_vsi_seid = %d\n", prefix,
> + caps->wol_proxy_vsi_seid);
> + ice_info(hw, "%s: max_mtu = %d\n", prefix, caps->max_mtu);
> + ice_print_led_caps(hw, caps, prefix, false);
> + ice_print_sdp_caps(hw, caps, prefix, false);
> +*/
next parent reply other threads:[~2023-11-21 20:46 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20231109161846.148524-1-przemyslaw.kitszel@intel.com>
2023-11-21 20:45 ` Jesse Brandeburg [this message]
2023-11-21 20:58 ` [cocci] [BUG] exponentially slow spatch Julia Lawall
2023-11-21 22:34 ` Julia Lawall
2023-11-21 22:57 ` Julia Lawall
2023-11-22 8:20 ` Markus Elfring
2023-11-27 19:48 ` Keller, Jacob E
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=e7263df7-be7d-47a4-b708-e04393f7887a@intel.com \
--to=jesse.brandeburg@intel.com \
--cc=Julia.Lawall@lip6.fr \
--cc=cocci@inria.fr \
--cc=jacob.e.keller@intel.com \
--cc=przemyslaw.kitszel@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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).