Intel PT decoder switch case fallthrough cases reported by gcc 7
Arnaldo Carvalho de Melo
maintainerHi,
I've updated the container with Fedora Rawhide I use to buildtools/perf/ and samples/bcc/ and it now comes with gcc 7, where I get things like:
CC /tmp/build/perf/tests/code-reading.outil/intel-pt-decoder/intel-pt-decoder.c: In function 'intel_pt_walk_psb':util/intel-pt-decoder/intel-pt-decoder.c:1748:31: error: this statement may fall through [-Werror=implicit-fallthrough=]
decoder->continuous_period = false;
^util/intel-pt-decoder/intel-pt-decoder.c:1749:3: note: here
case INTEL_PT_TIP_PGE:
^~~~util/intel-pt-decoder/intel-pt-decoder.c:1801:4: error: this statement may fall through [-Werror=implicit-fallthrough=]
intel_pt_clear_tx_flags(decoder);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~util/intel-pt-decoder/intel-pt-decoder.c:1802:3: note: here
case INTEL_PT_TNT:
^~~~util/intel-pt-decoder/intel-pt-decoder.c: In function 'intel_pt_walk_to_ip':util/intel-pt-decoder/intel-pt-decoder.c:1841:31: error: this statement may fall through [-Werror=implicit-fallthrough=]
decoder->continuous_period = false;
^util/intel-pt-decoder/intel-pt-decoder.c:1842:3: note: here
case INTEL_PT_TIP_PGE:
^~~~
MKDIR /tmp/build/perf/util/scripting-engines/
This gets solved with a new attribute, that you have to add where in the past we added:
/* Fall through */
To indicate that the fall through to the next case statement block is intentional, so now I have a __fallthrough and I am addressing all the cases, please check if the ones below are the ones intended for the Intel PT parts, please Ack or advise, only one seemed like a bug, i.e. a break should be used, but I'm not sure.
- Arnaldo
tools/perf/util/intel-pt-decoder/intel-pt-decoder.c +4 -0
@@ -1746,6 +1746,7 @@ static int intel_pt_walk_psb(struct intel_pt_decoder *decoder)switch (decoder->packet.type) {case INTEL_PT_TIP_PGD : decoder->continuous_period = false; + __fallthrough;case INTEL_PT_TIP_PGE :case INTEL_PT_TIP :intel_pt_log ("ERROR: Unexpected packet\n" );
@@ -1799,6 +1800,8 @@ static int intel_pt_walk_psb(struct intel_pt_decoder *decoder) decoder->pge = false; decoder->continuous_period = false;intel_pt_clear_tx_flags (decoder); +break ; +case INTEL_PT_TNT : decoder->have_tma = false;intel_pt_log ("ERROR: Unexpected packet\n" );
@@ -1839,6 +1842,7 @@ static int intel_pt_walk_to_ip(struct intel_pt_decoder *decoder)switch (decoder->packet.type) {case INTEL_PT_TIP_PGD : decoder->continuous_period = false; + __fallthrough;case INTEL_PT_TIP_PGE :case INTEL_PT_TIP : decoder->pge = decoder->packet.type !=INTEL_PT_TIP_PGD ;
tools/perf/util/intel-pt-decoder/intel-pt-pkt-decoder.c +2 -0
@@ -17,6 +17,7 @@#include <string.h> #include <endian.h> #include <byteswap.h> +#include <linux/compiler.h> #include "intel-pt-pkt-decoder.h"
@@ -498,6 +499,7 @@ int intel_pt_pkt_desc(const struct intel_pt_pkt *packet, char *buf,case INTEL_PT_FUP :if (!(packet->count))return snprintf (buf, buf_len,"%s no ip" , name); + __fallthrough;case INTEL_PT_CYC :case INTEL_PT_VMCS :case INTEL_PT_MTC :
Quoted 5 lines · Arnaldo Carvalho de Melo
Hi,
I've updated the container with Fedora Rawhide I use to build
tools/perf/andsamples/bcc/and it now comes with gcc 7, where I get things like:
FWIW, but it just shows that you should never ship software with -Werror enabled. New compiler releases add new warnings, and you just randomly break the build for users who use newer compilers.
It's ok to use in your own builds, but should never be default.
The kernel gets it right, but perf is wrong here.
-Andi
Arnaldo Carvalho de Melo
maintainerEm Thu, Feb 09, 2017 at 10:25:19AM -0800, Andi Kleen escreveu:
Quoted 7 lines · Arnaldo Carvalho de Melo
Quoted 5 lines
Hi,
I've updated the container with Fedora Rawhide I use to build
tools/perf/andsamples/bcc/and it now comes with gcc 7, where I get things like:
Quoted 7 lines
FWIW, but it just shows that you should never ship software with -Werror enabled. New compiler releases add new warnings, and you just randomly break the build for users who use newer compilers.
It's ok to use in your own builds, but should never be default.
The kernel gets it right, but perf is wrong here.
But what about my question? Do you think the changes are ok? I actually made all be fallthrough, i.e. considered that the existing code was ok.
- Arnaldo
Quoted 2 lines
But what about my question? Do you think the changes are ok? I actually made all be fallthrough, i.e. considered that the existing code was ok.
Yes the changes are fine.
-Andi
Arnaldo Carvalho de Melo
maintainerEm Thu, Feb 09, 2017 at 11:13:16AM -0800, Andi Kleen escreveu:
Quoted 4 lines
Quoted 2 lines
But what about my question? Do you think the changes are ok? I actually made all be fallthrough, i.e. considered that the existing code was ok.
Yes the changes are fine.
Thanks!
- Arnaldo