SH4: SCI improvement

I've add the flash support at 0xA0000000
and have qemu refusing to boot from it
qemu: fatal: Trying to execute code outside RAM or ROM at 0xa000000A

I've also try to boot u-boot as a kernel and have the output working but not the input. Impossible to type anything

U-Boot 2008.10-00237-g6d56cf3 (Nov 23 2008 - 21:08:17)

CPU: SH4
BOARD: Renesas Solutions R2D Plus
DRAM: 64MB
Timer with period zero, disabling
## Unknown FLASH on Bank 1 - Size = 0x00000000 = 0 MB
FLASH: 0MB
*** Warning - bad CRC, using default environment

In:    serial
Out:   serial
Err:   serial
Net:   No ethernet found.

=>

Best Regards,
J.

Hi, Jean-san!

It seems a good idea to make u-boot run on SH4-QEMU, for brushing up SH4 emulation.

Quoted 6 lines · Jean-Christophe PLAGNIOL-VILLARD

I've add the flash support at 0xA0000000
and have qemu refusing to boot from it
qemu: fatal: Trying to execute code outside RAM or ROM at 0xa000000A

I've also try to boot u-boot as a kernel and have the output working but not the input. Impossible to type anything

I'm afraid that I have no idea about these two points now.
I just guess MMU emulation and SCIF emulation might not be enough to run u-boot.

To investigate them more closely, I downloaded current u-boot git repository, build if for r2d, and run u-boot.bin with -kernel option. But u-boot couldn't boot with assertion failures. Following two parts caused them.

- MMUCR initialization with 16 bit word write causes failure. QEMU only implements 32 bit write for it, following SH7750 document.
- BSC's PCR access cause failure. QEMU does not implement it yet.

Have you handled them?

From you

From 54e5318dbc1beb1eb37843dc94331ce2d6d48f60 Mon Sep 17 00:00:00 2001

Quoted 18 lines · Shin-ichiro KAWASAKI

Hi, Jean-san!

It seems a good idea to make u-boot run on SH4-QEMU, for brushing up SH4 emulation.

Quoted 5 lines · Jean-Christophe PLAGNIOL-VILLARD

I've add the flash support at 0xA0000000
and have qemu refusing to boot from it
qemu: fatal: Trying to execute code outside RAM or ROM at 0xa000000A I've also try to boot u-boot as a kernel and have the output working but not the input. Impossible to type anything

I'm afraid that I have no idea about these two points now.
I just guess MMU emulation and SCIF emulation might not be enough to run u-boot.

To investigate them more closely, I downloaded current u-boot git repository, build if for r2d, and run u-boot.bin with -kernel option. But u-boot couldn't boot with assertion failures. Following two parts caused them.

Quoted 6 lines

- MMUCR initialization with 16 bit word write causes failure. QEMU only implements 32 bit write for it, following SH7750 document.
- BSC's PCR access cause failure. QEMU does not implement it yet.

Have you handled them?

you need to apply 1 patch on u-boot

r2dplus fix register access
I've send to the U-Boot ML

and 2 patchs to qemu

Quoted 1 line

From you

sh4: pci and mmu support
and from I
SH7750/51: add register BCR3, NCR4, PCR, RTCOR, RTCNT, RTCSR, SDMR2, SDMR3 and fix BCR2 support

Quoted 30 lines · Jean-Christophe PLAGNIOL-VILLARD

Quoted 18 lines · Shin-ichiro KAWASAKI

Hi, Jean-san!

It seems a good idea to make u-boot run on SH4-QEMU, for brushing up SH4 emulation.

Quoted 5 lines · Jean-Christophe PLAGNIOL-VILLARD

I've add the flash support at 0xA0000000
and have qemu refusing to boot from it
qemu: fatal: Trying to execute code outside RAM or ROM at 0xa000000A I've also try to boot u-boot as a kernel and have the output working but not the input. Impossible to type anything

I'm afraid that I have no idea about these two points now.
I just guess MMU emulation and SCIF emulation might not be enough to run u-boot.

To investigate them more closely, I downloaded current u-boot git repository, build if for r2d, and run u-boot.bin with -kernel option. But u-boot couldn't boot with assertion failures. Following two parts caused them.

Quoted 6 lines

- MMUCR initialization with 16 bit word write causes failure. QEMU only implements 32 bit write for it, following SH7750 document.
- BSC's PCR access cause failure. QEMU does not implement it yet.

Have you handled them?

you need to apply 1 patch on u-boot

r2dplus fix register access
I've send to the U-Boot ML

Quoted 7 lines

and 2 patchs to qemu

Quoted 1 line

From you

sh4: pci and mmu support
and from I
SH7750/51: add register BCR3, NCR4, PCR, RTCOR, RTCNT, RTCSR, SDMR2, SDMR3 and fix BCR2 support

Thanks.

One more simple modification seems necessary to avoid assertion error, as attached to this mail. With these patches, I saw that the U-boot for r2dplus boots cleanly when it is passed with -kernel option to qemu. Does it help you?
# Please be sure to give two '-serial' options in qemu command line.

I'm interested in the flash support, which will make it more easy to handle linux kernel boot conditions for qemu-sh users.

This patch simply implement one register of SH4's SCI := Serial Communication Interface.

R2D evaluation board uses SCI for SPI connection. So, Linux kernel for R2D with default configuration causes a QEMU assertion failure when it initializes SPI driver. This patch avoids it and reduces the kernel config modification work for QEMU.

Completing SCI implementation task is left. Other board support is desirable to confirm this task, which uses SCI for a serial terminal.

Regards,
Shin-ichiro KAWASAKI

Signed-off-by Shin-ichiro KAWASAKI <[email protected]>

Index: trunk/hw/sh_serial.c
===================================================================

hw/sh_serial.c +6 -6

@@ -168,19 +168,19 @@
         }
     }
     else {
+        switch(offs) {
 #if 0
-        switch(offs) {
         case 0x0c:
             ret = s->dr;
             break;
         case 0x10:
             ret = 0;
             break;
+#endif
         case 0x1c:
-            ret = s->sptr;
-            break;
+            s->sptr = val & 0x8f;
+            return;
         }
-#endif
     }
 
     fprintf(stderr, "sh_serial: unsupported write to 0x%02x\n", offs);

@@ -260,8 +260,8 @@
         }
     }
     else {
+        switch(offs) {
 #if 0
-        switch(offs) {
         case 0x0c:
             ret = s->dr;
             break;

@@ -271,11 +271,11 @@
         case 0x14:
             ret = s->rx_fifo[0];
             break;
+#endif
         case 0x1c:
             ret = s->sptr;
             break;
         }
-#endif
     }
 #ifdef DEBUG_SERIAL
     printf("sh_serial: read base=0x%08lx offs=0x%02x val=0x%x\n",

This patch simply implement one register of SH4's SCI := Serial Communication Interface.

R2D evaluation board uses SCI for SPI connection. So, Linux kernel for R2D with default configuration causes a QEMU assertion failure when it initializes SPI driver. This patch avoids it and reduces the kernel config modification work for QEMU.

Completing SCI implementation task is left. Other board support is desirable to confirm this task, which uses SCI for a serial terminal.

Regards,
Shin-ichiro KAWASAKI

Signed-off-by Shin-ichiro KAWASAKI <[email protected]>

Quoted 9 lines · Shin-ichiro KAWASAKI

This patch simply implement one register of SH4's SCI := Serial Communication Interface.

R2D evaluation board uses SCI for SPI connection. So, Linux kernel for R2D with default configuration causes a QEMU assertion failure when it initializes SPI driver. This patch avoids it and reduces the kernel config modification work for QEMU.

Completing SCI implementation task is left. Other board support is desirable to confirm this task, which uses SCI for a serial terminal.

Quoted 6 lines

Regards,
Shin-ichiro KAWASAKI

Signed-off-by Shin-ichiro KAWASAKI <[email protected]>

After this, are you able to use the standard defconfig for R2D?

If so, I
can kill off the qemu defconfig in-tree. If not, please send updates for the qemu defconfig so we can make sure that the kernel and the qemu tree are in sync with regards to feature completeness.