From: "H. Peter Anvin" <hpa@zytor.com>
To: Alain Knaff <alain@knaff.lu>
Cc: linux-kernel@vger.kernel.org
Subject: Re: update8 [PATCH 2/5] init: bzip2 or lzma -compressed kernels and initrds
Date: Sun, 04 Jan 2009 15:08:12 -0800 [thread overview]
Message-ID: <4961415C.1050708@zytor.com> (raw)
In-Reply-To: <200901042146.n04LkHgP005837@hitchhiker.hitchhiker.org.lu.hitchhiker.org.lu>
[-- Attachment #1: Type: text/plain, Size: 1000 bytes --]
Hi Alain,
Looking pretty good now. The only issues that I see so far are:
- Putting the kernel compression selection in init/Kconfig isn't going
to work too well, since it affects all architectures, and not all
architectures even do their own decompression (for quite a few
architectures it's the boot loader's responsibility.)
I see two options here: either have this be parameterized by the arch
Kconfig files, or simply move this chunk into arch/*/Kconfig as part
of the arch-enablement patches. The latter is probably the simplest,
even if it means some replicated code.
Let me know what you think -- I can do this pretty easily while
importing, so you don't need to submit a full new patchset.
- I did a followon patch (attached) to change the ramdisk compression
search to table driven. Since I already did the followon patch, don't
worry about it.
-hpa
--
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel. I don't speak on their behalf.
[-- Attachment #2: 0003-init-use-a-table-to-search-for-initramfs-compressio.patch --]
[-- Type: text/x-patch, Size: 3804 bytes --]
>From 832edddc59b5f8d8243aca91a20fb55ed64d00fb Mon Sep 17 00:00:00 2001
From: H. Peter Anvin <hpa@zytor.com>
Date: Sun, 4 Jan 2009 14:42:52 -0800
Subject: [PATCH] init: use a table to search for initramfs compression formats
Impact: Code simplification
Instead of open-coding testing for initramfs compression formats, use
a table.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
---
init/do_mounts_rd.c | 82 +++++++++++++++++++++-----------------------------
1 files changed, 35 insertions(+), 47 deletions(-)
diff --git a/init/do_mounts_rd.c b/init/do_mounts_rd.c
index dcaeb1f..9c9d7db 100644
--- a/init/do_mounts_rd.c
+++ b/init/do_mounts_rd.c
@@ -43,13 +43,31 @@ static int __init crd_load(int in_fd, int out_fd, decompress_fn deco);
* numbers could not be found.
*
* We currently check for the following magic numbers:
- * minix
- * ext2
+ * minix
+ * ext2
* romfs
* cramfs
- * gzip
+ * gzip
*/
-static int __init
+static const struct compress_format {
+ unsigned char magic[2];
+ const char *name;
+ decompress_fn decompressor;
+} compressed_formats[] = {
+#ifdef CONFIG_RD_GZIP
+ { {037, 0213}, "gzip", gunzip },
+ { {037, 0236}, "gzip", gunzip },
+#endif
+#ifdef CONFIG_RD_BZIP2
+ { {0x42, 0x5a}, "bzip2", bunzip2 },
+#endif
+#ifdef CONFIG_RD_LZMA
+ { {0x5d, 0x00}, "lzma", unlzma },
+#endif
+ { {0, 0}, NULL, NULL }
+};
+
+static int __init
identify_ramdisk_image(int fd, int start_block, decompress_fn *decompressor)
{
const int size = 512;
@@ -59,6 +77,7 @@ identify_ramdisk_image(int fd, int start_block, decompress_fn *decompressor)
struct cramfs_super *cramfsb;
int nblocks = -1;
unsigned char *buf;
+ const struct compress_format *cf;
buf = kmalloc(size, GFP_KERNEL);
if (!buf)
@@ -71,52 +90,21 @@ identify_ramdisk_image(int fd, int start_block, decompress_fn *decompressor)
memset(buf, 0xe5, size);
/*
- * Read block 0 to test for gzipped kernel
+ * Read block 0 to test for compressed kernel
*/
sys_lseek(fd, start_block * BLOCK_SIZE, 0);
sys_read(fd, buf, size);
-#ifdef CONFIG_RD_GZIP
- /*
- * If it matches the gzip magic numbers, return 0
- */
- if (buf[0] == 037 && ((buf[1] == 0213) || (buf[1] == 0236))) {
- printk(KERN_NOTICE
- "RAMDISK: Compressed image found at block %d\n",
- start_block);
- *decompressor = gunzip;
- nblocks = 0;
- goto done;
- }
-#endif
-
-#ifdef CONFIG_RD_BZIP2
- /*
- * If it matches the bzip2 magic numbers, return -1
- */
- if (buf[0] == 0x42 && (buf[1] == 0x5a)) {
- printk(KERN_NOTICE
- "RAMDISK: Bzipped image found at block %d\n",
- start_block);
- *decompressor = bunzip2;
- nblocks = 0;
- goto done;
- }
-#endif
-
-#ifdef CONFIG_RD_LZMA
- /*
- * If it matches the lzma magic numbers, return -1
- */
- if (buf[0] == 0x5d && (buf[1] == 0x00)) {
- printk(KERN_NOTICE
- "RAMDISK: Lzma image found at block %d\n",
- start_block);
- *decompressor = unlzma;
- nblocks = 0;
- goto done;
+ for (cf = compressed_formats; cf->decompressor; cf++) {
+ if (buf[0] == cf->magic[0] && buf[1] == cf->magic[1]) {
+ printk(KERN_NOTICE
+ "RAMDISK: %s image found at block %d\n",
+ cf->name, start_block);
+ *decompressor = cf->decompressor;
+ nblocks = 0;
+ goto done;
+ }
}
-#endif
/* romfs is at block zero too */
if (romfsb->word0 == ROMSB_WORD0 &&
@@ -165,7 +153,7 @@ identify_ramdisk_image(int fd, int start_block, decompress_fn *decompressor)
printk(KERN_NOTICE
"RAMDISK: Couldn't find valid RAM disk image starting at %d.\n",
start_block);
-
+
done:
sys_lseek(fd, start_block * BLOCK_SIZE, 0);
kfree(buf);
@@ -224,7 +212,7 @@ int __init rd_load_image(char *from)
nblocks, rd_blocks);
goto done;
}
-
+
/*
* OK, time to copy in the data
*/
--
1.5.6.6
next prev parent reply other threads:[~2009-01-04 23:08 UTC|newest]
Thread overview: 50+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-01-04 21:46 update8 [PATCH 2/5] init: bzip2 or lzma -compressed kernels and initrds Alain Knaff
2009-01-04 23:08 ` H. Peter Anvin [this message]
2009-01-04 23:12 ` Alain Knaff
2009-01-04 23:14 ` H. Peter Anvin
2009-01-04 23:21 ` Alain Knaff
2009-01-04 23:58 ` tip: bzip2/lzma now in tip:x86/setup-lzma H. Peter Anvin
2009-01-05 3:03 ` Sam Ravnborg
2009-01-05 5:09 ` H. Peter Anvin
2009-01-05 5:42 ` Sam Ravnborg
[not found] ` <49615136.9080900@knaff.lu>
[not found] ` <4961580A.1020301@zytor.com>
[not found] ` <4961A816.40302@knaff.lu>
[not found] ` <4961A997.10108@zytor.com>
[not found] ` <4961ADC5.6030108@knaff.lu>
[not found] ` <49622DE9.2010200@zytor.com>
[not found] ` <496240DF.2010102@knaff.lu>
[not found] ` <49624F6C.8010103@zytor.com>
[not found] ` <4962522F.20804@knaff.lu>
[not found] ` <496255B0.1050208@zytor.com>
2009-01-05 18:57 ` Alain Knaff
2009-01-05 19:36 ` H. Peter Anvin
2009-01-05 22:07 ` Alain Knaff
2009-01-05 22:11 ` H. Peter Anvin
2009-01-05 22:12 ` Alain Knaff
2009-01-05 22:59 ` H. Peter Anvin
2009-01-06 7:09 ` Alain Knaff
2009-01-06 7:21 ` Willy Tarreau
2009-01-06 7:22 ` H. Peter Anvin
2009-01-06 7:30 ` Alain Knaff
2009-01-06 21:57 ` [bzip2/lzma] fix for built-in initramfs issue Alain Knaff
2009-01-06 22:48 ` H. Peter Anvin
2009-01-06 22:50 ` Alain Knaff
2009-01-06 22:58 ` H. Peter Anvin
2009-01-06 22:58 ` Alain Knaff
2009-01-06 7:18 ` tip: bzip2/lzma now in tip:x86/setup-lzma Jaswinder Singh Rajput
2009-01-06 7:24 ` H. Peter Anvin
2009-01-06 7:53 ` Jaswinder Singh Rajput
2009-01-06 8:27 ` H. Peter Anvin
2009-02-17 21:03 ` Jan Engelhardt
2009-02-17 21:05 ` H. Peter Anvin
2009-02-17 22:08 ` Ingo Molnar
2009-02-17 23:37 ` Ingo Molnar
2009-02-18 0:52 ` H. Peter Anvin
2009-02-18 7:48 ` Alain Knaff
2009-02-18 9:20 ` Jan Engelhardt
2009-02-18 9:40 ` Alain Knaff
2009-02-18 10:29 ` Jan Engelhardt
2009-02-18 19:53 ` H. Peter Anvin
2009-02-19 6:14 ` Alain Knaff
2009-02-19 14:46 ` H. Peter Anvin
2009-02-19 15:41 ` Alain Knaff
2009-02-19 18:03 ` H. Peter Anvin
2009-02-18 19:52 ` H. Peter Anvin
2009-02-18 21:09 ` Willy Tarreau
2009-02-19 20:11 ` Alain Knaff
2009-03-01 13:16 ` Alain Knaff
2009-03-01 19:27 ` H. Peter Anvin
2009-03-02 9:53 ` Ingo Molnar
2009-03-02 9:54 ` Alain Knaff
2009-03-02 10:22 ` Ingo Molnar
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4961415C.1050708@zytor.com \
--to=hpa@zytor.com \
--cc=alain@knaff.lu \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox