Skip to contentSkip to conversation

arm: ramdisk uimage support for arm

PATCH In review Michele Dionisio 2 participants 4 messages Dec 16, 2015 – Dec 17, 2015

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 arm: ramdisk uimage support for arm

Hi, I have add support for uboot ramdisk support in the patch attached.

the new features is implemented in:
kexec/arch/arm/kexec-uImage-arm.c
kexec/arch/arm/kexec-zImage-arm.c

but there is also a fix in kexec/kexec-uImage.c where there is a endianes bug.

If someone can review the patch I think that it can be usefull to be added in the next version.

Regards.

kexec/arch/arm/kexec-uImage-arm.c +8 -2
index 03c2f4d..20d71a2 100644
@@ -17,6 +17,12 @@ int uImage_arm_probe(const char *buf, off_t len)
 int uImage_arm_load(int argc, char **argv, const char *buf, off_t len,
 	struct kexec_info *info)
 {
-	return zImage_arm_load(argc, argv, buf + sizeof(struct image_header),
-	                       len - sizeof(struct image_header), info);
+    struct Image_info img;
+	int ret;
+
+	ret = uImage_load(buf, len, &img);
+	if (ret)
+		return ret;
+    
+	return zImage_arm_load(argc, argv, img.buf, img.len, info);
 }
kexec/arch/arm/kexec-zImage-arm.c +33 -1
index d85ab9b..fda30f8 100644
@@ -14,6 +14,8 @@
 #include <getopt.h>
 #include <unistd.h>
 #include <libfdt.h>
+#include <image.h>
+#include <kexec-uImage.h>
 #include <arch/options.h>
 #include "../../kexec.h"
 #include "../../kexec-syscall.h"
@@ -318,6 +320,36 @@ static int setup_dtb_prop(char **bufp, off_t *sizep, const char *node_name,
 	return 0;
 }
 
+/*
+ * Load the ramdisk into buffer.
+ *  If the supplied image is in uImage format use
+ *  uImage_load() to read the payload from the image.
+ */
+char *slurp_ramdisk_arm(const char *filename, off_t *r_size)
+{
+	struct Image_info img;
+	off_t size;
+	const unsigned char *buf = slurp_file(filename, &size);
+	int rc;
+
+	/* Check if this is a uImage RAMDisk */
+	if (!buf)
+		return buf;
+	rc = uImage_probe_ramdisk(buf, size, IH_ARCH_ARM); 
+	if (rc < 0)
+		die("uImage: Corrupted ramdisk file %s\n", filename);
+	else if (rc == 0) {
+		if (uImage_load(buf, size, &img) != 0)
+			die("uImage: Reading %ld bytes from %s failed\n",
+				size, filename);
+		buf = img.buf;
+		size = img.len;
+	}
+
+	*r_size = size;
+	return buf;
+}
+
 int zImage_arm_load(int argc, char **argv, const char *buf, off_t len,
 	struct kexec_info *info)
 {
@@ -410,7 +442,7 @@ int zImage_arm_load(int argc, char **argv, const char *buf, off_t len,
 			command_line_len = COMMAND_LINE_SIZE;
 	}
 	if (ramdisk)
-		ramdisk_buf = slurp_file(ramdisk, &initrd_size);
+		ramdisk_buf = slurp_ramdisk_arm(ramdisk, &initrd_size);
 
 	if (dtb_file)
 		dtb_buf = slurp_file(dtb_file, &dtb_length);
kexec/kexec-uImage.c +1 -1
index 9df601b..5b85e86 100644
@@ -235,7 +235,7 @@ int uImage_load(const unsigned char *buf, off_t len, struct Image_info *image)
 {
 	const struct image_header *header = (const struct image_header *)buf;
 	const unsigned char *img_buf = buf + sizeof(struct image_header);
-	off_t img_len = header->ih_size;
+	off_t img_len =  cpu_to_be32(header->ih_size);
 
 	/*
 	 * Prevent loading a modified image.

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#1 lore raw

Hi Michele,

kexec/arch/arm/kexec-zImage-arm.c static int setup_dtb_prop(char **bufp, off_t *sizep, const char *node_name, · line 328

Quoted 6 lines · Michele Dionisio
kexec/arch/arm/kexec-zImage-arm.c +6 -0
+/*
+ * Load the ramdisk into buffer.
+ *  If the supplied image is in uImage format use
+ *  uImage_load() to read the payload from the image.
+ */
+char *slurp_ramdisk_arm(const char *filename, off_t *r_size)

This function is exactly as slurp_ramdisk_ppc(). So, why not we have a common slurp_ramdisk() function usable for both arm and ppc.

~Pratyush

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#2 lore raw

yes it is true.

I'm not in the position to test ppc. So my only way to be sure to avoid to introduce regration is to avoid to refactor the code.

But it is true. If you are able to test ppc I can try to post the refactoring.

Il 17/12/2015 05:23, Pratyush Anand ha scritto:

Quoted 12 lines

Hi Michele,

kexec/arch/arm/kexec-zImage-arm.c static int setup_dtb_prop(char **bufp, off_t *sizep, const char *node_name, · line 328

Quoted 6 lines · Michele Dionisio
kexec/arch/arm/kexec-zImage-arm.c +6 -0
+/*
+ * Load the ramdisk into buffer.
+ *  If the supplied image is in uImage format use
+ *  uImage_load() to read the payload from the image.
+ */
+char *slurp_ramdisk_arm(const char *filename, off_t *r_size)

This function is exactly as slurp_ramdisk_ppc(). So, why not we have a common slurp_ramdisk() function usable for both arm and ppc.

Quoted 3 lines

~Pratyush
.

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec

#3 lore raw
Quoted 7 lines · Michele Dionisio

yes it is true.

I'm not in the position to test ppc. So my only way to be sure to avoid to introduce regration is to avoid to refactor the code.

But it is true. If you are able to test ppc I can try to post the refactoring.

I also do not be able to test it on ppc, but I believe refactoring would be trivial and would not have any functionality difference. So, better to go with that rather than duplicating the code. You may keep ppc developer in CC and ask for ACK from them.

~Pratyush

_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec