Keil C51 right shift for 16 bit int

[hithesh123]

When I try to right shift 16bit data, the MSbit gets loaded with 1.

mydata=0x8000;
mydata>>=1;

Now, mydata becomes 0xC000;

Does the 16bit shift work differently in keil C51.
My Mcu is C8051F120.

 

[jayanth.devarayanadurga]

0x8000 >> 1 = 0x4000
0x4000 >> 1 = 0x2000
0x2000 >> 1 = 0x1000
0x1000 >> 1 = 0x0800
0x0800 >> 1 = 0x0400
0x0400 >> 1 = 0x0200
0x0200 >> 1 = 0x0100
0x0100 >> 1 = 0x0080
0x0080 >> 1 = 0x0040
0x0040 >> 1 = 0x0020
0x0020 >> 1 = 0x0010
0x0010 >> 1 = 0x0008
0x0008 >> 1 = 0x0004
0x0004 >> 1 = 0x0002
0x0002 >> 1 = 0x0001

 

[hithesh123]

0x8000 >> 1 = 0x4000
0x4000 >> 1 = 0x2000
0x2000 >> 1 = 0x1000
0x1000 >> 1 = 0x0800
0x0800 >> 1 = 0x0400
0x0400 >> 1 = 0x0200
0x0200 >> 1 = 0x0100
0x0100 >> 1 = 0x0080
0x0080 >> 1 = 0x0040
0x0040 >> 1 = 0x0020
0x0020 >> 1 = 0x0010
0x0010 >> 1 = 0x0008
0x0008 >> 1 = 0x0004
0x0004 >> 1 = 0x0002
0x0002 >> 1 = 0x0001

Jayanth,

My point was - when you expect 0x8000 to become 0x4000, in keil C51 I get 0xc000.

 

[jayanth.devarayanadurga]

Zip and post your Keil project files. I will debug and see. Did you use unsigned int?

 

[embpic]

i think you are doing again ORing with last data that is 0x8000 >> 1 = 0x4000
and after ORing
0x8000 | 0x4000 = 0xc000.

 

[FvM]

mydata=0x8000;
mydata>>=1;

Now, mydata becomes 0xC000;

That's the expectable result if mydata is signed int (arithmetic shift).
You may want to use an unsigned int data type.