[v2] gdb/remote: Remove negative tid/pid handling in wite_ptid
Commit Message
Actually thread and process ID's are positive values. Accorting to
http://man7.org/linux/man-pages/man7/pthreads.7.html
threads are creating using "clone" syscall, so the ID generation mechanism
is similar for threads and processes. According to Linux source code
there is a function call tree, which allocates PID[TID]:
clone
|->_do_fork
|->copy_process
|->alloc_pid
|->idr_alloc_cyclic
|->idr_alloc_u32(idr, ptr, &id, max, gfp);
And in idr_alloc_u32() "id" is u32 value, which means positiveness.
Also according to:
https://elixir.bootlin.com/linux/latest/source/kernel/pid.c#L177
PID cannot be less than 1.
In Zephyr RTOS the k_thread_create function returns
thread ID which is actually pointer to k_thread structure.
If the memory addressing starts from 0x80000000, passing such
big values to write_ptid() leads to overflow of "int tid" variable
and thread ID becomes negative.
So lets remove the code, which handles negative tid/pid values.
gdb/ChangeLog:
2019-11-06 Evgeniy Didin <didin@synopsys.com>
* remote.c (remote_target::write_ptid): Remove handling
negative tid,pid. Change "int" to "unsigned int" for pid/tid.
Signed-off-by: Evgeniy Didin <didin@synopsys.com>
Cc: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Claudiu Zissulescu <claziss@synopsys.com>
Cc: Cupertino Miranda <cmiranda@synopsys.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Shahab Vahedi <shahab@synopsys.com>
Cc: linux-snps-arc@lists.infradead.org
Signed-off-by: Evgeniy Didin <didin@synopsys.com>
---
Changes v1-v2:
-make no change of tid/pid bitness,
use generic "unsigned int" instead of "uint32_t"
gdb/ChangeLog | 6 ++++++
gdb/remote.c | 12 +++---------
2 files changed, 9 insertions(+), 9 deletions(-)
Comments
On 11/6/19 1:27 PM, Evgeniy Didin wrote:
> Actually thread and process ID's are positive values. Accorting to
> http://man7.org/linux/man-pages/man7/pthreads.7.html
> threads are creating using "clone" syscall, so the ID generation mechanism
> is similar for threads and processes. According to Linux source code
> there is a function call tree, which allocates PID[TID]:
> clone
> |->_do_fork
> |->copy_process
> |->alloc_pid
> |->idr_alloc_cyclic
> |->idr_alloc_u32(idr, ptr, &id, max, gfp);
> And in idr_alloc_u32() "id" is u32 value, which means positiveness.
> Also according to:
> https://elixir.bootlin.com/linux/latest/source/kernel/pid.c#L177
> PID cannot be less than 1.
Sure for Linux. But negative numbers have meaning in the remote protocol:
https://sourceware.org/gdb/current/onlinedocs/gdb/Packets.html#thread_002did-syntax
Thanks,
Pedro Alves
@@ -1,3 +1,9 @@
+2019-11-06 Evgeniy Didin <didin@synopsys.com>
+
+ * remote.c (remote_target::write_ptid): Remove handling
+ negative tid,pid. Change "int" to "unsigned int" for pid/tid.
+
+
2019-11-05 Tom Tromey <tom@tromey.com>
* tui/tui-disasm.c (struct tui_asm_line) <addr_size>: New member.
@@ -2909,22 +2909,16 @@ static int remote_newthread_step (threadref *ref, void *context);
char *
remote_target::write_ptid (char *buf, const char *endbuf, ptid_t ptid)
{
- int pid, tid;
+ unsigned int pid, tid;
struct remote_state *rs = get_remote_state ();
if (remote_multi_process_p (rs))
{
pid = ptid.pid ();
- if (pid < 0)
- buf += xsnprintf (buf, endbuf - buf, "p-%x.", -pid);
- else
- buf += xsnprintf (buf, endbuf - buf, "p%x.", pid);
+ buf += xsnprintf (buf, endbuf - buf, "p%x.", pid);
}
tid = ptid.lwp ();
- if (tid < 0)
- buf += xsnprintf (buf, endbuf - buf, "-%x", -tid);
- else
- buf += xsnprintf (buf, endbuf - buf, "%x", tid);
+ buf += xsnprintf (buf, endbuf - buf, "%x", tid);
return buf;
}