From: Tom de Vries <tdevries@suse.de>
To: Simon Marchi <simon.marchi@efficios.com>, gdb-patches@sourceware.org
Subject: Re: [PATCH] gdb/testsuite: allow "require" callbacks to provide a reason
Date: Tue, 28 Mar 2023 17:44:05 +0200 [thread overview]
Message-ID: <0e0254fc-5fa8-71b2-01b8-cc5742909727@suse.de> (raw)
In-Reply-To: <20230328122305.6108-1-simon.marchi@efficios.com>
On 3/28/23 14:23, Simon Marchi via Gdb-patches wrote:
> When an allow_* proc returns false, it can be a bit difficult what check
> failed exactly, if the procedure does multiple checks. To make
> investigation easier, I propose to allow the "require" callbacks to be
> able to return a list of two elements: the zero/non-zero value, and a
> reason string.
>
Hi,
I like the idea.
The only question I had was what the desired behaviour is for:
...
return {1 "bla"}
...
AFAICT, the current implementation just ignores "bla", and I wondered if
perhaps we should error out to avoid the impression that something will
be done with "bla".
That however will make this fail if $res == 1:
...
set res [try_foo]
return {$res "foo"}
...
so probably that's just too restrictive.
LGTM as-is.
Thanks,
- Tom
> Use the new feature in allow_hipcc_tests to demonstrate it (it's also
> where I hit actually hit this inconvenience). On my computer (where GDB
> is built with amd-dbgapi support but where I don't have a suitable GPU
> target), I get:
>
> UNSUPPORTED: gdb.rocm/simple.exp: require failed: allow_hipcc_tests (no suitable amdgpu targets found)
>
> vs before:
>
> UNSUPPORTED: gdb.rocm/simple.exp: require failed: allow_hipcc_tests
>
> Change-Id: Id1966535b87acfcbe9eac99f49dc1196398c6578
> ---
> gdb/testsuite/lib/gdb.exp | 37 +++++++++++++++++++++++++++++--------
> gdb/testsuite/lib/rocm.exp | 8 ++++----
> 2 files changed, 33 insertions(+), 12 deletions(-)
>
> diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
> index 46aa1441d6d0..7fb4f1cbdab4 100644
> --- a/gdb/testsuite/lib/gdb.exp
> +++ b/gdb/testsuite/lib/gdb.exp
> @@ -9251,22 +9251,43 @@ gdb_caching_proc have_avx {} {
> #
> # ARG can either be a name, or of the form !NAME.
> #
> -# Each name is a proc to evaluate in the caller's context. It returns
> -# a boolean, and a "!" means to invert the result. If this is
> -# nonzero, all is well. If it is zero, an "untested" is emitted and
> -# this proc causes the caller to return.
> +# Each name is a proc to evaluate in the caller's context. It can return a
> +# boolean or a two element list with a boolean and a reason string.
> +# A "!" means to invert the result. If this is true, all is well. If it is
> +# false, an "unsupported" is emitted and this proc causes the caller to return.
> +#
> +# The reason string is used to provide some context about a require failure,
> +# and is included in the "unsupported" message.
>
> proc require { args } {
> foreach arg $args {
> if {[string index $arg 0] == "!"} {
> - set ok 0
> + set required_val 0
> set fn [string range $arg 1 end]
> } else {
> - set ok 1
> + set required_val 1
> set fn $arg
> }
> - if {$ok != !![uplevel 1 $fn]} {
> - unsupported "require failed: $arg"
> +
> + set result [uplevel 1 $fn]
> + set len [llength $result]
> + if { $len == 2 } {
> + set actual_val [lindex $result 0]
> + set msg [lindex $result 1]
> + } elseif { $len == 1 } {
> + set actual_val $result
> + set msg ""
> + } else {
> + error "proc $fn returned a list of unexpected length $len"
> + }
> +
> + if {$required_val != !!$actual_val} {
> + if { [string length $msg] > 0 } {
> + unsupported "require failed: $arg ($msg)"
> + } else {
> + unsupported "require failed: $arg"
> + }
> +
> return -code return 0
> }
> }
> diff --git a/gdb/testsuite/lib/rocm.exp b/gdb/testsuite/lib/rocm.exp
> index b3e435339dbc..389d73bcaa5f 100644
> --- a/gdb/testsuite/lib/rocm.exp
> +++ b/gdb/testsuite/lib/rocm.exp
> @@ -64,19 +64,19 @@ gdb_caching_proc allow_hipcc_tests {} {
> # testing against GDBserver, there's no point in running the ROCm
> # tests.
> if {[target_info gdb_protocol] != ""} {
> - return 0
> + return {0 "remote debugging"}
> }
>
> # Ensure that GDB is built with amd-dbgapi support.
> set output [remote_exec host $::GDB "$::INTERNAL_GDBFLAGS --configuration"]
> if { [string first "--with-amd-dbgapi" $output] == -1 } {
> - return 0
> + return {0 "amd-dbgapi not supported"}
> }
>
> # Check we have a working hipcc compiler available.
> set targets [hcc_amdgpu_targets]
> if { [llength $targets] == 0} {
> - return 0
> + return {0 "no suitable amdgpu targets found"}
> }
>
> set flags [list hip additional_flags=--offload-arch=[join $targets ","]]
> @@ -93,7 +93,7 @@ gdb_caching_proc allow_hipcc_tests {} {
> return 0;
> }
> } executable $flags]} {
> - return 0
> + return {0 "failed to compile hip program"}
> }
>
> return 1
next prev parent reply other threads:[~2023-03-28 15:44 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-28 12:23 Simon Marchi
2023-03-28 15:44 ` Tom de Vries [this message]
2023-03-28 15:51 ` Simon Marchi
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=0e0254fc-5fa8-71b2-01b8-cc5742909727@suse.de \
--to=tdevries@suse.de \
--cc=gdb-patches@sourceware.org \
--cc=simon.marchi@efficios.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).