From: Simon Marchi <simon.marchi@polymtl.ca>
To: Yao Qi <qiyaoltc@gmail.com>
Cc: gdb-patches@sourceware.org
Subject: Re: [PATCH 06/25] Generate c for feature instead of tdesc
Date: Mon, 26 Jun 2017 21:38:00 -0000 [thread overview]
Message-ID: <3308812a9a2209d76f83656eeccb39bc@polymtl.ca> (raw)
In-Reply-To: <1497256916-4958-7-git-send-email-yao.qi@linaro.org>
On 2017-06-12 10:41, Yao Qi wrote:
> diff --git a/gdb/features/Makefile b/gdb/features/Makefile
> index 3bc8b5a..28a7e20 100644
> --- a/gdb/features/Makefile
> +++ b/gdb/features/Makefile
> @@ -252,12 +252,39 @@ $(outdir)/%.dat: %.xml number-regs.xsl
> sort-regs.xsl gdbserver-regs.xsl
> $(XSLTPROC) gdbserver-regs.xsl - >> $(outdir)/$*.tmp
> sh ../../move-if-change $(outdir)/$*.tmp $(outdir)/$*.dat
>
> -cfiles: $(CFILES)
> -%.c: %.xml
> +FEATURE_XMLFILES = i386/32bit-core.xml \
> + i386/32bit-sse.xml \
> + i386/32bit-linux.xml \
> + i386/32bit-avx.xml \
> + i386/32bit-mpx.xml \
> + i386/32bit-avx512.xml \
> + i386/32bit-pkeys.xml
> +
> +FEATURE_CFILES = $(patsubst %.xml,%.c,$(FEATURE_XMLFILES))
> +
> +cfiles: $(CFILES) $(FEATURE_CFILES)
If we are going to have different kinds of CFILES, it would be nice to
rename CFILES (to TDESC_CFILES I suppose) to clarify the purpose of each
variable.
> +
> +$(CFILES): %.c: %.xml
> $(GDB) -nx -q -batch \
> -ex "set tdesc filename $<" -ex 'maint print c-tdesc' > $@.tmp
> sh ../../move-if-change $@.tmp $@
>
> +$(FEATURE_CFILES): %.c: %.xml.tmp
> + $(GDB) -nx -q -batch \
> + -ex 'maint print c-tdesc $<' > $@.tmp
> + sh ../../move-if-change $@.tmp $@
> + rm $<
> +
> +%.xml.tmp: %.xml
> + echo "<?xml version=\"1.0\"?>" > $@
> + echo "<!DOCTYPE target SYSTEM \"gdb-target.dtd\">" >> $@
> + echo "<target>" >> $@
> + echo " <architecture>" >> $@
> + if test $(findstring i386/32bit-,$@); then echo "i386" >> $@ ; fi;
Should this be
test -n "$(findstring ...)"
Quotes would be important here, if findstring returns en empty string,
> + echo " </architecture>" >> $@
> + echo " <xi:include href=\"$(notdir $<)\"/>" >> $@
> + echo "</target>" >> $@
> +
Instead of creating this temporary file, wouldn't it be simpler to make
"maint print c-tdesc" (or a new "maint print c-feature") take directly a
path to a feature XML, and generate the C from that?
> @@ -2033,38 +2040,123 @@ public:
> printf_unfiltered ("%d, \"%s\");\n", reg->bitsize, reg->type);
> }
>
> +protected:
> + std::string m_filename_after_features;
> +
> private:
> char *m_function;
> - std::string m_filename_after_features;
> bool m_printed_field_type = false;
> bool m_printed_type = false;
> };
>
> +class print_c_feature : public print_c_tdesc
I am really not convinced that such an inheritance is appropriate here.
To make print_c_feature inherit from print_c_tdesc, we should be able to
say that a print_c_feature object _is_ a print_c_tdesc with further
specialization. I don't think it's the case here. If they can share
some code, I think it would be appropriate for them to have a common
base class though.
Simon
next prev parent reply other threads:[~2017-06-26 21:38 UTC|newest]
Thread overview: 82+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-06-12 8:42 [PATCH 00/25 V2] Make GDB builtin target descriptions more flexible Yao Qi
2017-06-12 8:42 ` [PATCH 13/25] Dynamically create tdesc in GDBserver Yao Qi
2017-06-12 8:42 ` [PATCH 21/25] Lazily and dynamically create amd64-linux target descriptions Yao Qi
2017-06-12 8:42 ` [PATCH 16/25] Dynamically composite xml in reply to GDB Yao Qi
2017-06-12 8:42 ` [PATCH 24/25] [GDBserver] Use pre-generated amd64-linux tdesc as test Yao Qi
2017-06-12 8:42 ` [PATCH 22/25] Regenerate two regformats/i386/.dat files Yao Qi
2017-06-22 12:43 ` Yao Qi
2017-06-12 8:42 ` [PATCH 09/25] Use target_desc fields expedite_regs and xmltarget ifndef IN_PROCESS_AGENT Yao Qi
2017-06-28 16:16 ` Pedro Alves
2017-06-28 17:42 ` Pedro Alves
2017-06-28 17:45 ` Pedro Alves
2017-06-29 11:45 ` Yao Qi
2017-06-12 8:42 ` [PATCH 07/25] Lazily and dynamically create i386-linux target descriptions Yao Qi
2017-06-20 11:01 ` Pedro Alves
2017-06-20 14:07 ` Yao Qi
2017-06-28 15:30 ` Pedro Alves
2017-06-12 8:42 ` [PATCH 19/25] GDBserver: remove srv_i386_linux_xmlfiles Yao Qi
2017-06-12 8:42 ` [PATCH 03/25] Class-fy tdesc_reg tdesc_type and tdesc_feature Yao Qi
2017-06-19 20:55 ` Simon Marchi
2017-06-19 21:30 ` Simon Marchi
2017-06-20 10:31 ` Yao Qi
2017-06-12 8:42 ` [PATCH 10/25] Adjust code generated by regformats/regdat.sh Yao Qi
2017-06-20 11:09 ` Pedro Alves
2017-06-21 14:28 ` Yao Qi
2017-06-12 8:42 ` [PATCH 01/25] Move initialize_tdesc_mips* calls from mips-linux-nat.c to mips-linux-tdep.c Yao Qi
2017-06-12 15:25 ` Maciej W. Rozycki
2017-06-13 8:07 ` Yao Qi
2017-06-12 8:42 ` [PATCH 20/25] Centralize amd64-linux target descriptions Yao Qi
2017-06-12 8:42 ` [PATCH 14/25] [RFC] GDBserver self test Yao Qi
2017-06-28 17:09 ` Pedro Alves
2017-06-29 9:08 ` Yao Qi
2017-06-12 8:42 ` [PATCH 02/25] Adjust the order of 32bit-linux.xml and 32bit-sse.xml in i386/i386-linux.xml Yao Qi
2017-06-19 20:22 ` Simon Marchi
2017-06-19 21:24 ` Pedro Alves
2017-06-19 21:48 ` Simon Marchi
2017-06-19 21:56 ` Pedro Alves
2017-06-20 9:20 ` Yao Qi
2017-06-20 10:12 ` Pedro Alves
2017-06-20 11:09 ` Yao Qi
2017-06-12 8:42 ` [PATCH 15/25] [RFC] GDBserver unit test to i386_tdesc Yao Qi
2017-06-28 17:22 ` Pedro Alves
2017-06-29 9:27 ` Yao Qi
2017-06-12 8:42 ` [PATCH 17/25] Remove features/i386/i386-*linux.c Yao Qi
2017-06-12 8:42 ` [PATCH 23/25] [GDBserver] Convert amd64-linux target descriptions Yao Qi
2017-06-28 19:00 ` Pedro Alves
2017-06-12 8:42 ` [PATCH 04/25] Centralize i386 linux " Yao Qi
2017-06-19 21:27 ` Simon Marchi
2017-06-12 8:42 ` [PATCH 25/25] Remove features/i386/amd64-*linux.c and features/i386/x32-*linux.c Yao Qi
2017-06-12 8:42 ` [PATCH 11/25] Use VEC for target_desc.reg_defs Yao Qi
2017-06-28 19:01 ` Pedro Alves
2017-06-29 11:05 ` Yao Qi
2017-06-29 11:31 ` Pedro Alves
2017-06-29 13:24 ` Yao Qi
2017-06-12 8:42 ` [PATCH 08/25] Add "maint check xml-descriptions" to test builtin xml target descriptions Yao Qi
2017-06-28 16:13 ` Pedro Alves
2017-06-12 8:42 ` [PATCH 05/25] Use visitor pattern for "maint print c-tdesc" Yao Qi
2017-06-20 23:37 ` Simon Marchi
2017-06-12 8:42 ` [PATCH 18/25] [GDBserver] Use pre-generated tdesc as test Yao Qi
2017-06-12 8:42 ` [PATCH 12/25] [GDBserver] Centralize tdesc for i386-linux Yao Qi
2017-06-12 8:42 ` [PATCH 06/25] Generate c for feature instead of tdesc Yao Qi
2017-06-12 14:48 ` Eli Zaretskii
2017-06-13 12:07 ` Yao Qi
2017-06-13 14:49 ` Eli Zaretskii
2017-06-13 15:31 ` Yao Qi
2017-06-13 15:41 ` Eli Zaretskii
2017-06-14 16:21 ` Yao Qi
2017-06-14 16:32 ` Eli Zaretskii
2017-06-15 13:19 ` Yao Qi
2017-06-15 14:45 ` Eli Zaretskii
2017-06-20 10:59 ` Pedro Alves
2017-06-22 14:49 ` Yao Qi
2017-06-22 15:36 ` Pedro Alves
2017-06-22 15:58 ` Yao Qi
2017-06-26 21:38 ` Simon Marchi [this message]
2017-06-29 15:24 ` Yao Qi
2017-06-19 19:59 ` [PATCH 00/25 V2] Make GDB builtin target descriptions more flexible Simon Marchi
2017-06-20 11:02 ` Yao Qi
2017-06-26 14:45 ` Tedeschi, Walfred
2017-06-27 13:49 ` Alan Hayward
2017-06-28 8:28 ` Yao Qi
2017-06-28 8:06 ` Yao Qi
2017-06-28 19:06 ` Pedro Alves
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=3308812a9a2209d76f83656eeccb39bc@polymtl.ca \
--to=simon.marchi@polymtl.ca \
--cc=gdb-patches@sourceware.org \
--cc=qiyaoltc@gmail.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).