�I was looking at the SAVE_ALL code for kernel 2.6.
I am confused why SAVE_ALL restores __USER_DS to ds and es ?
#define SAVE_ALL \
� � � �cld; \
� � � �pushl %fs; \
� � � �pushl %es; \
� � � �pushl %ds; \
� � � �pushl %eax; \
� � � �pushl %ebp; \
� � � �pushl %edi; \
� � � �pushl %esi; \
� � � �pushl %edx; \
� � � �pushl %ecx; \
� � � �pushl %ebx; \
� � � �movl $(__USER_DS), %edx; \ � <---- why __USER_DS ?
� � � �movl %edx, %ds; \
� � � �movl %edx, %es; \
� � � �movl $(__KERNEL_PDA), %edx; \
� � � �movl %edx, %fs
If you look at the linux 2.4, it restores __KERNEL_DS in SAVE_ALL.
#define SAVE_ALL \
� � � �cld; \
� � � �pushl %es; \
� � � �pushl %ds; \
� � � �pushl %eax; \
� � � �pushl %ebp; \
� � � �pushl %edi; \
� � � �pushl %esi; \
� � � �pushl %edx; \
� � � �pushl %ecx; \
� � � �pushl %ebx; \
� � � �movl $(__KERNEL_DS),%edx; \ � � <-- __KERNEL_DS is restored here.
� � � �movl %edx,%ds; \
� � � �movl %edx,%es;
Does anyone know why do we restore __USER_DS on context switch from user to kernel instead of __KERNEL_DS ?
since Linux doesn't use segmentation , all the segment descriptor values are the same. So I think it doesn't matter what selector value you load in DS, ES etc. Though I don't know the real reason for why __USER_DS �is used instead of __KERNEL_DS
take care,
-Joel
