Fix 32bit builds #50
Fix 32bit builds #50
Conversation
…he offset
It appears that the previous code cannot possibly work correctly, but I
could not produce any failure without this patch.
However, when building for arm32 bit, the following error shows up:
run_system_call.c: In function 'syscall_splice':
run_system_call.c:3250:31: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
3250 | result = splice(fd_in_live, (loff_t *) off_in, fd_out_live,
| ^
run_system_call.c:3251:5: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
3251 | (loff_t *) off_out, len, flags);
| ^
cc1: all warnings being treated as errors
make: *** [<builtin>: run_system_call.o] Error 1
The man pages for splice says that splice expects a pointer to the
offset. This is done for ifc.splice. Doing the same for splice.
Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
When building for arm32 the following error shows up:
run_system_call.c: In function 'cmsg_expect_eq':
run_system_call.c:860:38: error: format '%lu' expects argument of type 'long unsigned int', but argument 4 has type 'size_t' {aka 'unsigned in
t'} [-Werror=format=]
860 | "Bad len in cmsg %d: expected=%lu actual=%lu",
| ~~^
| |
| long unsigned int
| %u
861 | i, ecm->cmsg_len, acm->cmsg_len);
| ~~~~~~~~~~~~~
run_system_call.c:3129:6: error: format '%lu' expects argument of type 'long unsigned int', but argument 4 has type 'uint64_t' {aka 'long long
unsigned int'} [-Werror=format=]
3129 | "epoll_event->data does not match script: "
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
......
3133 | event_live->data.u64);
| ~~~~~~~~~~~~~~~~~~~~
To fix these use %zu instead of %lu for size_t, and for the uint64_t use
%PRIu64 instad of %lu.
Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
|
Thanks or the patches! Regarding the patch: Regarding the second patch, "net-test: packetdrill: run_system_call: fix 32 bit builds"; in the commit message: (1) please document the compile error you received on this line, that motivated changing that one: (2) please document that you tested your change by also compiling the resulting code on a 64-bit x86 Linux machine as well, and verifying that this had no warnings or errors. Thanks! |
I'm sorry but its not clear to me what the code should do.
Will update "net-test: packetdrill: run_system_call: fix 32 bit builds" according to your suggestions. Cheers |
These patches solves the 32-bit builds that was reported a few days back and it fixes a bug.