From mboxrd@z Thu Jan 1 00:00:00 1970 From: Graeme Russ Date: Wed, 04 Jan 2012 22:14:48 +1100 Subject: [U-Boot] [PATCH 07/17] x86: Use fs for global data In-Reply-To: References: <1325477374-6417-1-git-send-email-graeme.russ@gmail.com> <1325477374-6417-8-git-send-email-graeme.russ@gmail.com> Message-ID: <4F0434A8.8030707@gmail.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Hi Simon, On 04/01/12 16:36, Simon Glass wrote: > Hi Graeme, > > On Sun, Jan 1, 2012 at 8:09 PM, Graeme Russ wrote: >> Use the base address of the 'F' segment as a pointer to the global data >> structure. By adding the linear address (i.e. the 'D' segment address) >> as the first word of the global data structure, the address of the >> global data relative to the 'D' segment can be found simply, for >> example, by: >> >> fs movl 0, %eax >> >> This makes the gd 'pointer' writable prior to relocation (by reloading >> the GDT) which brings x86 into line with all other arches > > What is the GDT? Global Descriptor Table - It's a kind of lookup table which the x86 CPU uses to calculate physical addresses relative to 'segments' - By placing the start of the 'F' segment to the physical address of the global data structure, and adding a self-referencing physical address to the global data structure (as the first member), reading the first word of the 'F' segment provides the physical address of the global data >> >> NOTE: Writing to the gd 'pointer' is expensive (but we only do it >> twice) but using it to access global data members (read and write) is >> still fairly cheap >> >> Signed-off-by: Graeme Russ >> --- >> arch/x86/cpu/cpu.c | 53 ++++++++++++++++++++-------------- >> arch/x86/cpu/start.S | 8 ++++- >> arch/x86/include/asm/global_data.h | 21 +++++++++---- >> arch/x86/include/asm/processor.h | 6 +++- >> arch/x86/include/asm/u-boot-x86.h | 2 + >> arch/x86/lib/board.c | 56 +++++++++++++++++++++++++----------- >> 6 files changed, 98 insertions(+), 48 deletions(-) >> >> diff --git a/arch/x86/cpu/cpu.c b/arch/x86/cpu/cpu.c >> index bf55c26..e7a5fc0 100644 >> --- a/arch/x86/cpu/cpu.c >> +++ b/arch/x86/cpu/cpu.c >> @@ -90,6 +90,37 @@ static void load_gdt(const u64 *boot_gdt, u16 num_entries) >> asm volatile("lgdtl %0\n" : : "m" (gdt)); >> } >> >> +void init_gd(gd_t *id, u64 *gdt_addr) >> +{ >> + id->gd_addr = (ulong)id; >> + setup_gdt(id, gdt_addr); >> +} >> + >> +void setup_gdt(gd_t *id, u64 *gdt_addr) > > I will probably never understand this function but a comment might be nice. Hmm, I guess I really don't know how to comment this to explain it clearly - The GDT is a bit of a magical beast which you kind of have to learn by osmosis ;) > >> +{ >> + /* CS: code, read/execute, 4 GB, base 0 */ >> + gdt_addr[GDT_ENTRY_32BIT_CS] = GDT_ENTRY(0xc09b, 0, 0xfffff); >> + >> + /* DS: data, read/write, 4 GB, base 0 */ >> + gdt_addr[GDT_ENTRY_32BIT_DS] = GDT_ENTRY(0xc093, 0, 0xfffff); >> + >> + /* FS: data, read/write, 4 GB, base (Global Data Pointer) */ >> + gdt_addr[GDT_ENTRY_32BIT_FS] = GDT_ENTRY(0xc093, (ulong)id, 0xfffff); >> + >> + /* 16-bit CS: code, read/execute, 64 kB, base 0 */ > > Extra indent? Removed [snip] >> diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h >> index 203c63a..07897f9 100644 >> --- a/arch/x86/include/asm/processor.h >> +++ b/arch/x86/include/asm/processor.h >> @@ -24,13 +24,17 @@ >> #ifndef __ASM_PROCESSOR_H_ >> #define __ASM_PROCESSOR_H_ 1 >> >> +#define GDT_ENTRY_SIZE 8 >> + >> #define GDT_ENTRY_NULL 0 >> #define GDT_ENTRY_UNUSED (GDT_ENTRY_NULL + 1) >> #define GDT_ENTRY_32BIT_CS (GDT_ENTRY_UNUSED + 1) >> #define GDT_ENTRY_32BIT_DS (GDT_ENTRY_32BIT_CS + 1) >> -#define GDT_ENTRY_16BIT_CS (GDT_ENTRY_32BIT_DS + 1) >> +#define GDT_ENTRY_32BIT_FS (GDT_ENTRY_32BIT_DS + 1) >> +#define GDT_ENTRY_16BIT_CS (GDT_ENTRY_32BIT_FS + 1) >> #define GDT_ENTRY_16BIT_DS (GDT_ENTRY_16BIT_CS + 1) >> >> #define GDT_NUM_ENTRIES (GDT_ENTRY_16BIT_DS + 1) >> +#define GDT_SIZE (GDT_NUM_ENTRIES * GDT_ENTRY_SIZE) > > Wasn't this already done in an earlier patch? I've moved this change into patch #3 Regards, Graeme