* [BK][PATCH] ReiserFS CPU and memory bandwidth efficient large writes
From: Hans Reiser @ 2002-12-13 18:56 UTC (permalink / raw)
To: Linus Torvalds; +Cc: Linux Kernel Mailing List
[-- Attachment #1: Type: text/plain, Size: 1 bytes --]
[-- Attachment #2: reiserfs_file_write() for 2.5.51 --]
[-- Type: message/rfc822, Size: 66294 bytes --]
From: Oleg Drokin <green@namesys.com>
To: reiser@namesys.com
Subject: reiserfs_file_write() for 2.5.51
Date: Fri, 13 Dec 2002 20:04:50 +0300
Message-ID: <20021213200450.A8885@namesys.com>
Hello!
These three changesets implement reiserfs_file_write. Also third one
exports generic_osync_inode,block_commit_write and remove_suid
because these are now needed for reiserfs.
There was no reiserfs_file_write in the 2.4 port of reiserfs
(and Hans was very unhappy about it).
With current 'one block at a time' algorithm, writes past the end of a file
are slow because each new file block is separately added into the tree
causing shifting of other items which is CPU expensive.
With this new implementation if you write into file with big enough chunks,
it uses half as much CPU. Also this version is more SMP friendly than
the current one.
You can pull these from bk://thebsh.namesys.com/bk/reiser3-linux-2.5-own-filewrite
ChangeSet@1.847, 2002-12-07 16:56:59+03:00, green@angband.namesys.com
export generic_osync_inode,block_commit_write, remove_suid
ChangeSet@1.846, 2002-12-07 16:55:03+03:00, green@angband.namesys.com
reiserfs_file_write() implemenation. Ported from 2.4
ChangeSet@1.845, 2002-12-07 13:38:05+03:00, green@angband.namesys.com
reiserfs: Implement insertion of more than one unformatted pointer insertion at a time. This considerably speedups hole creation.
Diffstat:
fs/reiserfs/bitmap.c | 21
fs/reiserfs/do_balan.c | 105 ++--
fs/reiserfs/file.c | 1071 ++++++++++++++++++++++++++++++++++++++++-
fs/reiserfs/inode.c | 52 +
fs/reiserfs/super.c | 1
fs/reiserfs/tail_conversion.c | 5
include/linux/reiserfs_fs.h | 1
include/linux/reiserfs_fs_sb.h | 1
kernel/ksyms.c | 3
9 files changed, 1193 insertions(+), 67 deletions(-)
Plain text patch:
# This is a BitKeeper generated patch for the following project:
# Project Name: Linux kernel tree
# This patch format is intended for GNU patch command version 2.5 or higher.
# This patch includes the following deltas:
# ChangeSet 1.844 -> 1.847
# fs/reiserfs/file.c 1.18 -> 1.19
# kernel/ksyms.c 1.169 -> 1.170
# fs/reiserfs/do_balan.c 1.15 -> 1.16
# fs/reiserfs/inode.c 1.68 -> 1.70
# fs/reiserfs/super.c 1.54 -> 1.55
# include/linux/reiserfs_fs_sb.h 1.20 -> 1.21
# fs/reiserfs/bitmap.c 1.24 -> 1.25
# include/linux/reiserfs_fs.h 1.45 -> 1.46
# fs/reiserfs/tail_conversion.c 1.23 -> 1.24
#
# The following is the BitKeeper ChangeSet Log
# --------------------------------------------
# 02/12/07 green@angband.namesys.com 1.845
# reiserfs: Implement insertion of more than one unformatted pointer insertion at a time. This considerably speedups hole creation.
# --------------------------------------------
# 02/12/07 green@angband.namesys.com 1.846
# reiserfs_file_write() implemenation. Ported from 2.4
# --------------------------------------------
# 02/12/07 green@angband.namesys.com 1.847
# export generic_osync_inode,block_commit_write, remove_suid
# --------------------------------------------
#
diff -Nru a/fs/reiserfs/bitmap.c b/fs/reiserfs/bitmap.c
--- a/fs/reiserfs/bitmap.c Sat Dec 7 16:58:48 2002
+++ b/fs/reiserfs/bitmap.c Sat Dec 7 16:58:48 2002
@@ -9,6 +9,7 @@
#include <linux/errno.h>
#include <linux/buffer_head.h>
#include <linux/kernel.h>
+#include <linux/pagemap.h>
#include <linux/reiserfs_fs_sb.h>
#include <linux/reiserfs_fs_i.h>
@@ -733,7 +734,7 @@
int rest = amount_needed;
int nr_allocated;
- while (rest > 0) {
+ while (rest > 0 && start <= finish) {
nr_allocated = scan_bitmap (hint->th, &start, finish, 1,
rest + prealloc_size, !hint->formatted_node,
hint->block);
@@ -879,7 +880,9 @@
if ( !blocks )
return;
+ spin_lock(&REISERFS_SB(sb)->bitmap_lock);
REISERFS_SB(sb)->reserved_blocks += blocks;
+ spin_unlock(&REISERFS_SB(sb)->bitmap_lock);
}
/* Unreserve @blocks amount of blocks in fs pointed by @sb */
@@ -896,6 +899,22 @@
if ( !blocks )
return;
+ spin_lock(&REISERFS_SB(sb)->bitmap_lock);
REISERFS_SB(sb)->reserved_blocks -= blocks;
+ spin_unlock(&REISERFS_SB(sb)->bitmap_lock);
RFALSE( REISERFS_SB(sb)->reserved_blocks < 0, "amount of blocks reserved became zero?");
+}
+
+/* This function estimates how much pages we will be able to write to FS
+ used for reiserfs_file_write() purposes for now. */
+int reiserfs_can_fit_pages ( struct super_block *sb /* superblock of filesystem
+ to estimate space */ )
+{
+ unsigned long space;
+
+ spin_lock(&REISERFS_SB(sb)->bitmap_lock);
+ space = (SB_FREE_BLOCKS(sb) - REISERFS_SB(sb)->reserved_blocks) / ( PAGE_CACHE_SIZE/sb->s_blocksize);
+ spin_unlock(&REISERFS_SB(sb)->bitmap_lock);
+
+ return space;
}
diff -Nru a/fs/reiserfs/do_balan.c b/fs/reiserfs/do_balan.c
--- a/fs/reiserfs/do_balan.c Sat Dec 7 16:58:48 2002
+++ b/fs/reiserfs/do_balan.c Sat Dec 7 16:58:48 2002
@@ -319,8 +319,6 @@
int new_item_len;
int version;
- RFALSE (!is_direct_le_ih (ih),
- "PAP-12075: only direct inserted item can be broken. %h", ih);
ret_val = leaf_shift_left (tb, tb->lnum[0]-1, -1);
/* Calculate item length to insert to S[0] */
@@ -343,7 +341,7 @@
version = ih_version (ih);
/* Calculate key component, item length and body to insert into S[0] */
- set_le_ih_k_offset( ih, le_ih_k_offset( ih ) + tb->lbytes );
+ set_le_ih_k_offset( ih, le_ih_k_offset( ih ) + tb->lbytes * (is_indirect_le_ih(ih)?tb->tb_sb->s_blocksize/UNFM_P_SIZE:1) );
put_ih_item_len( ih, new_item_len );
if ( tb->lbytes > zeros_num ) {
@@ -452,23 +450,28 @@
ih_item_len( B_N_PITEM_HEAD(tb->L[0],n+item_pos-ret_val)),
l_n,body, zeros_num > l_n ? l_n : zeros_num
);
-
- RFALSE( l_n &&
- is_indirect_le_ih(B_N_PITEM_HEAD
- (tb->L[0],
- n + item_pos - ret_val)),
- "PAP-12110: pasting more than 1 unformatted node pointer into indirect item");
-
/* 0-th item in S0 can be only of DIRECT type when l_n != 0*/
{
- int version;
-
- version = ih_version (B_N_PITEM_HEAD (tbS0, 0));
- set_le_key_k_offset (version, B_N_PKEY (tbS0, 0),
- le_key_k_offset (version, B_N_PKEY (tbS0, 0)) + l_n);
- version = ih_version (B_N_PITEM_HEAD(tb->CFL[0],tb->lkey[0]));
- set_le_key_k_offset (version, B_N_PDELIM_KEY(tb->CFL[0],tb->lkey[0]),
- le_key_k_offset (version, B_N_PDELIM_KEY(tb->CFL[0],tb->lkey[0])) + l_n);
+ int version;
+ int temp_l = l_n;
+
+ RFALSE (ih_item_len (B_N_PITEM_HEAD (tbS0, 0)),
+ "PAP-12106: item length must be 0");
+ RFALSE (comp_short_le_keys (B_N_PKEY (tbS0, 0),
+ B_N_PKEY (tb->L[0],
+ n + item_pos - ret_val)),
+ "PAP-12107: items must be of the same file");
+ if (is_indirect_le_ih(B_N_PITEM_HEAD (tb->L[0],
+ n + item_pos - ret_val))) {
+ temp_l = (l_n / UNFM_P_SIZE) * tb->tb_sb->s_blocksize;
+ }
+ /* update key of first item in S0 */
+ version = ih_version (B_N_PITEM_HEAD (tbS0, 0));
+ set_le_key_k_offset (version, B_N_PKEY (tbS0, 0),
+ le_key_k_offset (version, B_N_PKEY (tbS0, 0)) + temp_l);
+ /* update left delimiting key */
+ set_le_key_k_offset (version, B_N_PDELIM_KEY(tb->CFL[0],tb->lkey[0]),
+ le_key_k_offset (version, B_N_PDELIM_KEY(tb->CFL[0],tb->lkey[0])) + temp_l);
}
/* Calculate new body, position in item and insert_size[0] */
@@ -537,7 +540,7 @@
);
/* if appended item is indirect item, put unformatted node into un list */
if (is_indirect_le_ih (pasted))
- set_ih_free_space (pasted, ((struct unfm_nodeinfo*)body)->unfm_freespace);
+ set_ih_free_space (pasted, 0);
tb->insert_size[0] = 0;
zeros_num = 0;
}
@@ -565,15 +568,11 @@
{ /* new item or its part falls to R[0] */
if ( item_pos == n - tb->rnum[0] + 1 && tb->rbytes != -1 )
{ /* part of new item falls into R[0] */
- int old_key_comp, old_len, r_zeros_number;
+ loff_t old_key_comp, old_len, r_zeros_number;
const char * r_body;
int version;
loff_t offset;
- RFALSE( !is_direct_le_ih (ih),
- "PAP-12135: only direct item can be split. (%h)",
- ih);
-
leaf_shift_right(tb,tb->rnum[0]-1,-1);
version = ih_version(ih);
@@ -582,7 +581,7 @@
old_len = ih_item_len(ih);
/* Calculate key component and item length to insert into R[0] */
- offset = le_ih_k_offset( ih ) + (old_len - tb->rbytes );
+ offset = le_ih_k_offset( ih ) + (old_len - tb->rbytes )*(is_indirect_le_ih(ih)?tb->tb_sb->s_blocksize/UNFM_P_SIZE:1);
set_le_ih_k_offset( ih, offset );
put_ih_item_len( ih, tb->rbytes);
/* Insert part of the item into R[0] */
@@ -590,13 +589,13 @@
bi.bi_bh = tb->R[0];
bi.bi_parent = tb->FR[0];
bi.bi_position = get_right_neighbor_position (tb, 0);
- if ( offset - old_key_comp > zeros_num ) {
+ if ( (old_len - tb->rbytes) > zeros_num ) {
r_zeros_number = 0;
- r_body = body + offset - old_key_comp - zeros_num;
+ r_body = body + (old_len - tb->rbytes) - zeros_num;
}
else {
r_body = body;
- r_zeros_number = zeros_num - (offset - old_key_comp);
+ r_zeros_number = zeros_num - (old_len - tb->rbytes);
zeros_num -= r_zeros_number;
}
@@ -707,12 +706,17 @@
{
int version;
+ unsigned long temp_rem = n_rem;
version = ih_version (B_N_PITEM_HEAD (tb->R[0],0));
+ if (is_indirect_le_key(version,B_N_PKEY(tb->R[0],0))){
+ temp_rem = (n_rem / UNFM_P_SIZE) *
+ tb->tb_sb->s_blocksize;
+ }
set_le_key_k_offset (version, B_N_PKEY(tb->R[0],0),
- le_key_k_offset (version, B_N_PKEY(tb->R[0],0)) + n_rem);
+ le_key_k_offset (version, B_N_PKEY(tb->R[0],0)) + temp_rem);
set_le_key_k_offset (version, B_N_PDELIM_KEY(tb->CFR[0],tb->rkey[0]),
- le_key_k_offset (version, B_N_PDELIM_KEY(tb->CFR[0],tb->rkey[0])) + n_rem);
+ le_key_k_offset (version, B_N_PDELIM_KEY(tb->CFR[0],tb->rkey[0])) + temp_rem);
}
/* k_offset (B_N_PKEY(tb->R[0],0)) += n_rem;
k_offset (B_N_PDELIM_KEY(tb->CFR[0],tb->rkey[0])) += n_rem;*/
@@ -736,13 +740,12 @@
leaf_paste_in_buffer(&bi, 0, n_shift, tb->insert_size[0] - n_rem, r_body, r_zeros_number);
if (is_indirect_le_ih (B_N_PITEM_HEAD(tb->R[0],0))) {
-
+#if 0
RFALSE( n_rem,
"PAP-12160: paste more than one unformatted node pointer");
-
- set_ih_free_space (B_N_PITEM_HEAD(tb->R[0],0), ((struct unfm_nodeinfo*)body)->unfm_freespace);
+#endif
+ set_ih_free_space (B_N_PITEM_HEAD(tb->R[0],0), 0);
}
-
tb->insert_size[0] = n_rem;
if ( ! n_rem )
pos_in_item ++;
@@ -781,7 +784,7 @@
}
if (is_indirect_le_ih (pasted))
- set_ih_free_space (pasted, ((struct unfm_nodeinfo*)body)->unfm_freespace);
+ set_ih_free_space (pasted, 0);
zeros_num = tb->insert_size[0] = 0;
}
}
@@ -858,12 +861,6 @@
const char * r_body;
int version;
- RFALSE( !is_direct_le_ih(ih),
- /* The items which can be inserted are:
- Stat_data item, direct item, indirect item and directory item which consist of only two entries "." and "..".
- These items must not be broken except for a direct one. */
- "PAP-12205: non-direct item can not be broken when inserting");
-
/* Move snum[i]-1 items from S[0] to S_new[i] */
leaf_move_items (LEAF_FROM_S_TO_SNEW, tb, snum[i] - 1, -1, S_new[i]);
/* Remember key component and item length */
@@ -873,7 +870,7 @@
/* Calculate key component and item length to insert into S_new[i] */
set_le_ih_k_offset( ih,
- le_ih_k_offset(ih) + (old_len - sbytes[i] ) );
+ le_ih_k_offset(ih) + (old_len - sbytes[i] )*(is_indirect_le_ih(ih)?tb->tb_sb->s_blocksize/UNFM_P_SIZE:1) );
put_ih_item_len( ih, sbytes[i] );
@@ -883,13 +880,13 @@
bi.bi_parent = 0;
bi.bi_position = 0;
- if ( le_ih_k_offset (ih) - old_key_comp > zeros_num ) {
+ if ( (old_len - sbytes[i]) > zeros_num ) {
r_zeros_number = 0;
- r_body = body + (le_ih_k_offset(ih) - old_key_comp) - zeros_num;
+ r_body = body + (old_len - sbytes[i]) - zeros_num;
}
else {
r_body = body;
- r_zeros_number = zeros_num - (le_ih_k_offset (ih) - old_key_comp);
+ r_zeros_number = zeros_num - (old_len - sbytes[i]);
zeros_num -= r_zeros_number;
}
@@ -1010,11 +1007,14 @@
tmp = B_N_PITEM_HEAD(S_new[i],0);
if (is_indirect_le_ih (tmp)) {
- if (n_rem)
- reiserfs_panic (tb->tb_sb, "PAP-12230: balance_leaf: invalid action with indirect item");
- set_ih_free_space (tmp, ((struct unfm_nodeinfo*)body)->unfm_freespace);
+ set_ih_free_space (tmp, 0);
+ set_le_ih_k_offset( tmp, le_ih_k_offset(tmp) +
+ (n_rem / UNFM_P_SIZE) *
+ tb->tb_sb->s_blocksize);
+ } else {
+ set_le_ih_k_offset( tmp, le_ih_k_offset(tmp) +
+ n_rem );
}
- set_le_ih_k_offset( tmp, le_ih_k_offset(tmp) + n_rem );
}
tb->insert_size[0] = n_rem;
@@ -1060,7 +1060,7 @@
/* if we paste to indirect item update ih_free_space */
if (is_indirect_le_ih (pasted))
- set_ih_free_space (pasted, ((struct unfm_nodeinfo*)body)->unfm_freespace);
+ set_ih_free_space (pasted, 0);
zeros_num = tb->insert_size[0] = 0;
}
}
@@ -1152,11 +1152,12 @@
leaf_paste_in_buffer (&bi, item_pos, pos_in_item, tb->insert_size[0], body, zeros_num);
if (is_indirect_le_ih (pasted)) {
-
+#if 0
RFALSE( tb->insert_size[0] != UNFM_P_SIZE,
"PAP-12280: insert_size for indirect item must be %d, not %d",
UNFM_P_SIZE, tb->insert_size[0]);
- set_ih_free_space (pasted, ((struct unfm_nodeinfo*)body)->unfm_freespace);
+#endif
+ set_ih_free_space (pasted, 0);
}
tb->insert_size[0] = 0;
}
diff -Nru a/fs/reiserfs/file.c b/fs/reiserfs/file.c
--- a/fs/reiserfs/file.c Sat Dec 7 16:58:48 2002
+++ b/fs/reiserfs/file.c Sat Dec 7 16:58:48 2002
@@ -6,6 +6,8 @@
#include <linux/time.h>
#include <linux/reiserfs_fs.h>
#include <linux/smp_lock.h>
+#include <asm/uaccess.h>
+#include <linux/pagemap.h>
/*
** We pack the tails of files on file close, not at the time they are written.
@@ -140,9 +142,1076 @@
return error ;
}
+/* this function from inode.c would be used here, too */
+extern void restart_transaction(struct reiserfs_transaction_handle *th,
+ struct inode *inode, struct path *path);
+
+/* I really do not want to play with memory shortage right now, so
+ to simplify the code, we are not going to write more than this much pages at
+ a time. This still should considerably improve performance compared to 4k
+ at a time case. */
+#define REISERFS_WRITE_PAGES_AT_A_TIME 32
+
+/* Allocates blocks for a file to fulfil write request.
+ Maps all unmapped but prepared pages from the list.
+ Updates metadata with newly allocated blocknumbers as needed */
+int reiserfs_allocate_blocks_for_region(
+ struct inode *inode, /* Inode we work with */
+ loff_t pos, /* Writing position */
+ int num_pages, /* number of pages write going
+ to touch */
+ int write_bytes, /* amount of bytes to write */
+ struct page **prepared_pages, /* array of
+ prepared pages
+ */
+ int blocks_to_allocate /* Amount of blocks we
+ need to allocate to
+ fit the data into file
+ */
+ )
+{
+ struct cpu_key key; // cpu key of item that we are going to deal with
+ struct item_head *ih; // pointer to item head that we are going to deal with
+ struct buffer_head *bh; // Buffer head that contains items that we are going to deal with
+ struct reiserfs_transaction_handle th; // transaction handle for transaction we are going to create.
+ __u32 * item; // pointer to item we are going to deal with
+ INITIALIZE_PATH(path); // path to item, that we are going to deal with.
+ b_blocknr_t allocated_blocks[blocks_to_allocate]; // Pointer to a place where allocated blocknumbers would be stored. Right now statically allocated, later that will change.
+ reiserfs_blocknr_hint_t hint; // hint structure for block allocator.
+ size_t res; // return value of various functions that we call.
+ int curr_block; // current block used to keep track of unmapped blocks.
+ int i; // loop counter
+ int itempos; // position in item
+ unsigned int from = (pos & (PAGE_CACHE_SIZE - 1)); // writing position in
+ // first page
+ unsigned int to = ((pos + write_bytes - 1) & (PAGE_CACHE_SIZE - 1)) + 1; /* last modified byte offset in last page */
+ __u64 hole_size ; // amount of blocks for a file hole, if it needed to be created.
+ int modifying_this_item = 0; // Flag for items traversal code to keep track
+ // of the fact that we already prepared
+ // current block for journal
+
+
+ RFALSE(!blocks_to_allocate, "green-9004: tried to allocate zero blocks?");
+
+ /* First we compose a key to point at the writing position, we want to do
+ that outside of any locking region. */
+ make_cpu_key (&key, inode, pos+1, TYPE_ANY, 3/*key length*/);
+
+ /* If we came here, it means we absolutely need to open a transaction,
+ since we need to allocate some blocks */
+ reiserfs_write_lock(inode->i_sb); // Journaling stuff and we need that.
+ journal_begin(&th, inode->i_sb, JOURNAL_PER_BALANCE_CNT * 3 + 1); // Wish I know if this number enough
+ reiserfs_update_inode_transaction(inode) ;
+
+ /* Look for the in-tree position of our write, need path for block allocator */
+ res = search_for_position_by_key(inode->i_sb, &key, &path);
+ if ( res == IO_ERROR ) {
+ res = -EIO;
+ goto error_exit;
+ }
+
+ /* Allocate blocks */
+ /* First fill in "hint" structure for block allocator */
+ hint.th = &th; // transaction handle.
+ hint.path = &path; // Path, so that block allocator can determine packing locality or whatever it needs to determine.
+ hint.inode = inode; // Inode is needed by block allocator too.
+ hint.search_start = 0; // We have no hint on where to search free blocks for block allocator.
+ hint.key = key.on_disk_key; // on disk key of file.
+ hint.block = inode->i_blocks/(inode->i_sb->s_blocksize/512); // Number of disk blocks this file occupies already.
+ hint.formatted_node = 0; // We are allocating blocks for unformatted node.
+ hint.preallocate = 0; // We do not do any preallocation for now.
+
+ /* Call block allocator to allocate blocks */
+ res = reiserfs_allocate_blocknrs(&hint, allocated_blocks, blocks_to_allocate, blocks_to_allocate);
+ if ( res != CARRY_ON ) {
+ if ( res == NO_DISK_SPACE ) {
+ /* We flush the transaction in case of no space. This way some
+ blocks might become free */
+ SB_JOURNAL(inode->i_sb)->j_must_wait = 1;
+ restart_transaction(&th, inode, &path);
+
+ /* We might have scheduled, so search again */
+ res = search_for_position_by_key(inode->i_sb, &key, &path);
+ if ( res == IO_ERROR ) {
+ res = -EIO;
+ goto error_exit;
+ }
+
+ /* update changed info for hint structure. */
+ res = reiserfs_allocate_blocknrs(&hint, allocated_blocks, blocks_to_allocate, blocks_to_allocate);
+ if ( res != CARRY_ON ) {
+ res = -ENOSPC;
+ pathrelse(&path);
+ goto error_exit;
+ }
+ } else {
+ res = -ENOSPC;
+ pathrelse(&path);
+ goto error_exit;
+ }
+ }
+
+#ifdef __BIG_ENDIAN
+ // Too bad, I have not found any way to convert a given region from
+ // cpu format to little endian format
+ {
+ int i;
+ for ( i = 0; i < blocks_to_allocate ; i++)
+ allocated_blocks[i]=cpu_to_le32(allocated_blocks[i]);
+ }
+#endif
+
+ /* Blocks allocating well might have scheduled and tree might have changed,
+ let's search the tree again */
+ /* find where in the tree our write should go */
+ res = search_for_position_by_key(inode->i_sb, &key, &path);
+ if ( res == IO_ERROR ) {
+ res = -EIO;
+ goto error_exit_free_blocks;
+ }
+
+ bh = get_last_bh( &path ); // Get a bufferhead for last element in path.
+ ih = get_ih( &path ); // Get a pointer to last item head in path.
+ item = get_item( &path ); // Get a pointer to last item in path
+
+ /* Let's see what we have found */
+ if ( res != POSITION_FOUND ) { /* position not found, this means that we
+ might need to append file with holes
+ first */
+ // Since we are writing past the file's end, we need to find out if
+ // there is a hole that needs to be inserted before our writing
+ // position, and how many blocks it is going to cover (we need to
+ // populate pointers to file blocks representing the hole with zeros)
+
+ hole_size = (pos + 1 - (le_key_k_offset( get_inode_item_key_version(inode), &(ih->ih_key))+op_bytes_number(ih, inode->i_sb->s_blocksize))) >> inode->i_sb->s_blocksize_bits;
+
+ if ( hole_size > 0 ) {
+ int to_paste = min_t(__u64, hole_size, MAX_ITEM_LEN(inode->i_sb->s_blocksize)/UNFM_P_SIZE ); // How much data to insert first time.
+ /* area filled with zeroes, to supply as list of zero blocknumbers
+ We allocate it outside of loop just in case loop would spin for
+ several iterations. */
+ char *zeros = kmalloc(to_paste*UNFM_P_SIZE, GFP_ATOMIC); // We cannot insert more than MAX_ITEM_LEN bytes anyway.
+ if ( !zeros ) {
+ res = -ENOMEM;
+ goto error_exit_free_blocks;
+ }
+ memset ( zeros, 0, to_paste*UNFM_P_SIZE);
+ do {
+ to_paste = min_t(__u64, hole_size, MAX_ITEM_LEN(inode->i_sb->s_blocksize)/UNFM_P_SIZE );
+ if ( is_indirect_le_ih(ih) ) {
+ /* Ok, there is existing indirect item already. Need to append it */
+ /* Calculate position past inserted item */
+ make_cpu_key( &key, inode, le_key_k_offset( get_inode_item_key_version(inode), &(ih->ih_key)) + op_bytes_number(ih, inode->i_sb->s_blocksize), TYPE_INDIRECT, 3);
+ res = reiserfs_paste_into_item( &th, &path, &key, (char *)zeros, UNFM_P_SIZE*to_paste);
+ if ( res ) {
+ kfree(zeros);
+ goto error_exit_free_blocks;
+ }
+ } else if ( is_statdata_le_ih(ih) ) {
+ /* No existing item, create it */
+ /* item head for new item */
+ struct item_head ins_ih;
+
+ /* create a key for our new item */
+ make_cpu_key( &key, inode, 1, TYPE_INDIRECT, 3);
+
+ /* Create new item head for our new item */
+ make_le_item_head (&ins_ih, &key, key.version, 1,
+ TYPE_INDIRECT, to_paste*UNFM_P_SIZE,
+ 0 /* free space */);
+
+ /* Find where such item should live in the tree */
+ res = search_item (inode->i_sb, &key, &path);
+ if ( res != ITEM_NOT_FOUND ) {
+ /* item should not exist, otherwise we have error */
+ if ( res != -ENOSPC ) {
+ reiserfs_warning ("green-9008: search_by_key (%K) returned %d\n",
+ &key, res);
+ }
+ res = -EIO;
+ kfree(zeros);
+ goto error_exit_free_blocks;
+ }
+ res = reiserfs_insert_item( &th, &path, &key, &ins_ih, (char *)zeros);
+ } else {
+ reiserfs_panic(inode->i_sb, "green-9011: Unexpected key type %K\n", &key);
+ }
+ if ( res ) {
+ kfree(zeros);
+ goto error_exit_free_blocks;
+ }
+ /* Now we want to check if transaction is too full, and if it is
+ we restart it. This will also free the path. */
+ if (journal_transaction_should_end(&th, th.t_blocks_allocated))
+ restart_transaction(&th, inode, &path);
+
+ /* Well, need to recalculate path and stuff */
+ set_cpu_key_k_offset( &key, cpu_key_k_offset(&key) + to_paste * inode->i_sb->s_blocksize );
+ res = search_for_position_by_key(inode->i_sb, &key, &path);
+ if ( res == IO_ERROR ) {
+ res = -EIO;
+ kfree(zeros);
+ goto error_exit_free_blocks;
+ }
+ bh=get_last_bh(&path);
+ ih=get_ih(&path);
+ item = get_item(&path);
+ hole_size -= to_paste;
+ } while ( hole_size );
+ kfree(zeros);
+ }
+ }
+
+ // Go through existing indirect items first
+ // replace all zeroes with blocknumbers from list
+ // Note that if no corresponding item was found, by previous search,
+ // it means there are no existing in-tree representation for file area
+ // we are going to overwrite, so there is nothing to scan through for holes.
+ for ( curr_block = 0, itempos = path.pos_in_item ; curr_block < blocks_to_allocate && res == POSITION_FOUND ; ) {
+
+ if ( itempos >= ih_item_len(ih)/UNFM_P_SIZE ) {
+ /* We run out of data in this indirect item, let's look for another
+ one. */
+ /* First if we are already modifying current item, log it */
+ if ( modifying_this_item ) {
+ journal_mark_dirty (&th, inode->i_sb, bh);
+ modifying_this_item = 0;
+ }
+ /* Then set the key to look for a new indirect item (offset of old
+ item is added to old item length */
+ set_cpu_key_k_offset( &key, le_key_k_offset( get_inode_item_key_version(inode), &(ih->ih_key)) + op_bytes_number(ih, inode->i_sb->s_blocksize));
+ /* Search ofor position of new key in the tree. */
+ res = search_for_position_by_key(inode->i_sb, &key, &path);
+ if ( res == IO_ERROR) {
+ res = -EIO;
+ goto error_exit_free_blocks;
+ }
+ bh=get_last_bh(&path);
+ ih=get_ih(&path);
+ item = get_item(&path);
+ itempos = path.pos_in_item;
+ continue; // loop to check all kinds of conditions and so on.
+ }
+ /* Ok, we have correct position in item now, so let's see if it is
+ representing file hole (blocknumber is zero) and fill it if needed */
+ if ( !item[itempos] ) {
+ /* Ok, a hole. Now we need to check if we already prepared this
+ block to be journaled */
+ while ( !modifying_this_item ) { // loop until succeed
+ /* Well, this item is not journaled yet, so we must prepare
+ it for journal first, before we can change it */
+ struct item_head tmp_ih; // We copy item head of found item,
+ // here to detect if fs changed under
+ // us while we were preparing for
+ // journal.
+ int fs_gen; // We store fs generation here to find if someone
+ // changes fs under our feet
+
+ copy_item_head (&tmp_ih, ih); // Remember itemhead
+ fs_gen = get_generation (inode->i_sb); // remember fs generation
+ reiserfs_prepare_for_journal(inode->i_sb, bh, 1); // Prepare a buffer within which indirect item is stored for changing.
+ if (fs_changed (fs_gen, inode->i_sb) && item_moved (&tmp_ih, &path)) {
+ // Sigh, fs was changed under us, we need to look for new
+ // location of item we are working with
+
+ /* unmark prepaerd area as journaled and search for it's
+ new position */
+ reiserfs_restore_prepared_buffer(inode->i_sb, bh);
+ res = search_for_position_by_key(inode->i_sb, &key, &path);
+ if ( res == IO_ERROR) {
+ res = -EIO;
+ goto error_exit_free_blocks;
+ }
+ bh=get_last_bh(&path);
+ ih=get_ih(&path);
+ item = get_item(&path);
+ // Itempos is still the same
+ continue;
+ }
+ modifying_this_item = 1;
+ }
+ item[itempos] = allocated_blocks[curr_block]; // Assign new block
+ curr_block++;
+ }
+ itempos++;
+ }
+
+ if ( modifying_this_item ) { // We need to log last-accessed block, if it
+ // was modified, but not logged yet.
+ journal_mark_dirty (&th, inode->i_sb, bh);
+ }
+
+ if ( curr_block < blocks_to_allocate ) {
+ // Oh, well need to append to indirect item, or to create indirect item
+ // if there weren't any
+ if ( is_indirect_le_ih(ih) ) {
+ // Existing indirect item - append. First calculate key for append
+ // position. We do not need to recalculate path as it should
+ // already point to correct place.
+ make_cpu_key( &key, inode, le_key_k_offset( get_inode_item_key_version(inode), &(ih->ih_key)) + op_bytes_number(ih, inode->i_sb->s_blocksize), TYPE_INDIRECT, 3);
+ res = reiserfs_paste_into_item( &th, &path, &key, (char *)(allocated_blocks+curr_block), UNFM_P_SIZE*(blocks_to_allocate-curr_block));
+ if ( res ) {
+ goto error_exit_free_blocks;
+ }
+ } else if (is_statdata_le_ih(ih) ) {
+ // Last found item was statdata. That means we need to create indirect item.
+ struct item_head ins_ih; /* itemhead for new item */
+
+ /* create a key for our new item */
+ make_cpu_key( &key, inode, 1, TYPE_INDIRECT, 3); // Position one,
+ // because that's
+ // where first
+ // indirect item
+ // begins
+ /* Create new item head for our new item */
+ make_le_item_head (&ins_ih, &key, key.version, 1, TYPE_INDIRECT,
+ (blocks_to_allocate-curr_block)*UNFM_P_SIZE,
+ 0 /* free space */);
+ /* Find where such item should live in the tree */
+ res = search_item (inode->i_sb, &key, &path);
+ if ( res != ITEM_NOT_FOUND ) {
+ /* Well, if we have found such item already, or some error
+ occured, we need to warn user and return error */
+ if ( res != -ENOSPC ) {
+ reiserfs_warning ("green-9009: search_by_key (%K) returned %d\n",
+ &key, res);
+ }
+ res = -EIO;
+ goto error_exit_free_blocks;
+ }
+ /* Insert item into the tree with the data as its body */
+ res = reiserfs_insert_item( &th, &path, &key, &ins_ih, (char *)(allocated_blocks+curr_block));
+ } else {
+ reiserfs_panic(inode->i_sb, "green-9010: unexpected item type for key %K\n",&key);
+ }
+ }
+
+ /* Now the final thing, if we have grew the file, we must update it's size*/
+ if ( pos + write_bytes > inode->i_size) {
+ inode->i_size = pos + write_bytes; // Set new size
+ }
+
+ /* Amount of on-disk blocks used by file have changed, update it */
+ inode->i_blocks += blocks_to_allocate * (inode->i_sb->s_blocksize / 512);
+ reiserfs_update_sd(&th, inode); // And update on-disk metadata
+ // finish all journal stuff now, We are not going to play with metadata
+ // anymore.
+ pathrelse(&path);
+ journal_end(&th, inode->i_sb, JOURNAL_PER_BALANCE_CNT * 3 + 1);
+ reiserfs_write_unlock(inode->i_sb);
+
+ // go through all the pages/buffers and map the buffers to newly allocated
+ // blocks (so that system knows where to write these pages later).
+ curr_block = 0;
+ for ( i = 0; i < num_pages ; i++ ) {
+ struct page *page=prepared_pages[i]; //current page
+ struct buffer_head *head = page_buffers(page);// first buffer for a page
+ int block_start, block_end; // in-page offsets for buffers.
+
+ if (!page_buffers(page))
+ reiserfs_panic(inode->i_sb, "green-9005: No buffers for prepared page???");
+
+ /* For each buffer in page */
+ for(bh = head, block_start = 0; bh != head || !block_start;
+ block_start=block_end, bh = bh->b_this_page) {
+ if (!bh)
+ reiserfs_panic(inode->i_sb, "green-9006: Allocated but absent buffer for a page?");
+ block_end = block_start+inode->i_sb->s_blocksize;
+ if (i == 0 && block_end <= from )
+ /* if this buffer is before requested data to map, skip it */
+ continue;
+ if (i == num_pages - 1 && block_start >= to)
+ /* If this buffer is after requested data to map, abort
+ processing of current page */
+ break;
+
+ if ( !buffer_mapped(bh) ) { // Ok, unmapped buffer, need to map it
+ map_bh( bh, inode->i_sb, le32_to_cpu(allocated_blocks[curr_block]));
+ curr_block++;
+ }
+ }
+ }
+
+ RFALSE( curr_block > blocks_to_allocate, "green-9007: Used too many blocks? weird");
+
+ return 0;
+
+// Need to deal with transaction here.
+error_exit_free_blocks:
+ pathrelse(&path);
+ // free blocks
+ for( i = 0; i < blocks_to_allocate; i++ )
+ reiserfs_free_block( &th, le32_to_cpu(allocated_blocks[i]));
+
+error_exit:
+ journal_end(&th, inode->i_sb, JOURNAL_PER_BALANCE_CNT * 3 + 1);
+ reiserfs_write_unlock(inode->i_sb);
+
+ return res;
+}
+
+/* Unlock pages prepared by reiserfs_prepare_file_region_for_write */
+void reiserfs_unprepare_pages(struct page **prepared_pages, /* list of locked pages */
+ int num_pages /* amount of pages */) {
+ int i; // loop counter
+
+ for (i=0; i < num_pages ; i++) {
+ struct page *page = prepared_pages[i];
+
+ try_to_free_buffers(page);
+ kunmap(page);
+ unlock_page(page);
+ page_cache_release(page);
+ }
+}
+
+/* This function will copy data from userspace to specified pages within
+ supplied byte range */
+int reiserfs_copy_from_user_to_file_region(
+ loff_t pos, /* In-file position */
+ int num_pages, /* Number of pages affected */
+ int write_bytes, /* Amount of bytes to write */
+ struct page **prepared_pages, /* pointer to
+ array to
+ prepared pages
+ */
+ const char *buf /* Pointer to user-supplied
+ data*/
+ )
+{
+ long page_fault=0; // status of copy_from_user.
+ int i; // loop counter.
+ int offset; // offset in page
+
+ for ( i = 0, offset = (pos & (PAGE_CACHE_SIZE-1)); i < num_pages ; i++,offset=0) {
+ int count = min_t(int,PAGE_CACHE_SIZE-offset,write_bytes); // How much of bytes to write to this page
+ struct page *page=prepared_pages[i]; // Current page we process.
+
+ /* Bring in the user page. We need to do it to fight some deadlocks
+ with copying the page to itself and page being inaccessible at
+ the same time.*/
+ { volatile unsigned char dummy;
+ __get_user(dummy, buf);
+ __get_user(dummy, buf+count-1);
+ /* We do getuser for beginning and ending of the region just because
+ userdata may be not page aligned, and therefore data that will
+ go into one page of file may be splitted onto two actual pages
+ in userspace */
+ }
+
+ /* Copy data from userspace to the current page */
+ kmap(page);
+ page_fault = __copy_from_user(page_address(page)+offset, buf, count); // Copy the data.
+ /* Flush processor's dcache for this page */
+ flush_dcache_page(page);
+ kunmap(page);
+ buf+=count;
+ write_bytes-=count;
+
+ if (page_fault)
+ break; // Was there a fault? abort.
+ }
+
+ return page_fault?-EFAULT:0;
+}
+
+
+
+/* Submit pages for write. This was separated from actual file copying
+ because we might want to allocate block numbers in-between.
+ This function assumes that caller will adjust file size to correct value. */
+int reiserfs_submit_file_region_for_write(
+ loff_t pos, /* Writing position offset */
+ int num_pages, /* Number of pages to write */
+ int write_bytes, /* number of bytes to write */
+ struct page **prepared_pages /* list of pages */
+ )
+{
+ int status; // return status of block_commit_write.
+ int retval = 0; // Return value we are going to return.
+ int i; // loop counter
+ int offset; // Writing offset in page.
+
+ for ( i = 0, offset = (pos & (PAGE_CACHE_SIZE-1)); i < num_pages ; i++,offset=0) {
+ int count = min_t(int,PAGE_CACHE_SIZE-offset,write_bytes); // How much of bytes to write to this page
+ struct page *page=prepared_pages[i]; // Current page we process.
+
+ status = block_commit_write(page, offset, offset+count);
+ if ( status )
+ retval = status; // To not overcomplicate matters We are going to
+ // submit all the pages even if there was error.
+ // we only remember error status to report it on
+ // exit.
+ write_bytes-=count;
+ SetPageReferenced(page);
+ unlock_page(page); // We unlock the page as it was locked by earlier call
+ // to grab_cache_page
+ page_cache_release(page);
+ }
+ return retval;
+}
+
+/* Look if passed writing region is going to touch file's tail
+ (if it is present). And if it is, convert the tail to unformatted node */
+int reiserfs_check_for_tail_and_convert( struct inode *inode, /* inode to deal with */
+ loff_t pos, /* Writing position */
+ int write_bytes /* amount of bytes to write */
+ )
+{
+ INITIALIZE_PATH(path); // needed for search_for_position
+ struct cpu_key key; // Key that would represent last touched writing byte.
+ struct item_head *ih; // item header of found block;
+ int res; // Return value of various functions we call.
+ int cont_expand_offset; // We will put offset for generic_cont_expand here
+ // This can be int just because tails are created
+ // only for small files.
+
+/* this embodies a dependency on a particular tail policy */
+ if ( inode->i_size >= inode->i_sb->s_blocksize*4 ) {
+ /* such a big files do not have tails, so we won't bother ourselves
+ to look for tails, simply return */
+ return 0;
+ }
+
+ reiserfs_write_lock(inode->i_sb);
+ /* find the item containing the last byte to be written, or if
+ * writing past the end of the file then the last item of the
+ * file (and then we check its type). */
+ make_cpu_key (&key, inode, pos+write_bytes+1, TYPE_ANY, 3/*key length*/);
+ res = search_for_position_by_key(inode->i_sb, &key, &path);
+ if ( res == IO_ERROR ) {
+ reiserfs_write_unlock(inode->i_sb);
+ return -EIO;
+ }
+ ih = get_ih(&path);
+ res = 0;
+ if ( is_direct_le_ih(ih) ) {
+ /* Ok, closest item is file tail (tails are stored in "direct"
+ * items), so we need to unpack it. */
+ /* To not overcomplicate matters, we just call generic_cont_expand
+ which will in turn call other stuff and finally will boil down to
+ reiserfs_get_block() that would do necessary conversion. */
+ cont_expand_offset = le_key_k_offset(get_inode_item_key_version(inode), &(ih->ih_key));
+ pathrelse(&path);
+ res = generic_cont_expand( inode, cont_expand_offset);
+ } else
+ pathrelse(&path);
+
+ reiserfs_write_unlock(inode->i_sb);
+ return res;
+}
+
+/* This function locks pages starting from @pos for @inode.
+ @num_pages pages are locked and stored in
+ @prepared_pages array. Also buffers are allocated for these pages.
+ First and last page of the region is read if it is overwritten only
+ partially. If last page did not exist before write (file hole or file
+ append), it is zeroed, then.
+ Returns number of unallocated blocks that should be allocated to cover
+ new file data.*/
+int reiserfs_prepare_file_region_for_write(
+ struct inode *inode /* Inode of the file */,
+ loff_t pos, /* position in the file */
+ int num_pages, /* number of pages to
+ prepare */
+ int write_bytes, /* Amount of bytes to be
+ overwritten from
+ @pos */
+ struct page **prepared_pages /* pointer to array
+ where to store
+ prepared pages */
+ )
+{
+ int res=0; // Return values of different functions we call.
+ unsigned long index = pos >> PAGE_CACHE_SHIFT; // Offset in file in pages.
+ int from = (pos & (PAGE_CACHE_SIZE - 1)); // Writing offset in first page
+ int to = ((pos + write_bytes - 1) & (PAGE_CACHE_SIZE - 1)) + 1;
+ /* offset of last modified byte in last
+ page */
+ struct address_space *mapping = inode->i_mapping; // Pages are mapped here.
+ int i; // Simple counter
+ int blocks = 0; /* Return value (blocks that should be allocated) */
+ struct buffer_head *bh, *head; // Current bufferhead and first bufferhead
+ // of a page.
+ unsigned block_start, block_end; // Starting and ending offsets of current
+ // buffer in the page.
+ struct buffer_head *wait[2], **wait_bh=wait; // Buffers for page, if
+ // Page appeared to be not up
+ // to date. Note how we have
+ // at most 2 buffers, this is
+ // because we at most may
+ // partially overwrite two
+ // buffers for one page. One at // the beginning of write area
+ // and one at the end.
+ // Everything inthe middle gets // overwritten totally.
+
+ struct cpu_key key; // cpu key of item that we are going to deal with
+ struct item_head *ih = NULL; // pointer to item head that we are going to deal with
+ struct buffer_head *itembuf=NULL; // Buffer head that contains items that we are going to deal with
+ INITIALIZE_PATH(path); // path to item, that we are going to deal with.
+ __u32 * item=0; // pointer to item we are going to deal with
+
+
+ if ( num_pages < 1 ) {
+ reiserfs_warning("green-9001: reiserfs_prepare_file_region_for_write called with zero number of pages to process\n");
+ return -EFAULT;
+ }
+
+ /* We have 2 loops for pages. In first loop we grab and lock the pages, so
+ that nobody would touch these until we release the pages. Then
+ we'd start to deal with mapping buffers to blocks. */
+ for ( i = 0; i < num_pages; i++) {
+ prepared_pages[i] = grab_cache_page(mapping, index + i); // locks the page
+ if ( !prepared_pages[i]) {
+ res = -ENOMEM;
+ goto failed_page_grabbing;
+ }
+ if (!page_has_buffers(prepared_pages[i]))
+ create_empty_buffers(prepared_pages[i], inode->i_sb->s_blocksize, 0);
+ }
+
+ /* Let's count amount of blocks for a case where all the blocks
+ overwritten are new (we will substract already allocated blocks later)*/
+ if ( num_pages > 2 )
+ /* These are full-overwritten pages so we count all the blocks in
+ these pages are counted as needed to be allocated */
+ blocks = (PAGE_CACHE_SIZE/inode->i_sb->s_blocksize) * (num_pages - 2);
+
+ /* count blocks needed for first page (possibly partially written) */
+ blocks += ((PAGE_CACHE_SIZE - from) >> inode->i_sb->s_blocksize_bits) +
+ !!(from & (inode->i_sb->s_blocksize-1)); /* roundup */
+
+ /* Now we account for last page. If last page == first page (we
+ overwrite only one page), we substract all the blocks past the
+ last writing position in a page out of already calculated number
+ of blocks */
+ blocks += (PAGE_CACHE_SIZE/inode->i_sb->s_blocksize) * (num_pages > 1) -
+ ((PAGE_CACHE_SIZE - to) >> inode->i_sb->s_blocksize_bits);
+ /* Note how we do not roundup here since partial blocks still
+ should be allocated */
+
+ /* Now if all the write area lies past the file end, no point in
+ maping blocks, since there is none, so we just zero out remaining
+ parts of first and last pages in write area (if needed) */
+ if ( (pos & ~(PAGE_CACHE_SIZE - 1)) > inode->i_size ) {
+ if ( from != 0 ) {/* First page needs to be partially zeroed */
+ char *kaddr = kmap_atomic(prepared_pages[0], KM_USER0);
+ memset(kaddr, 0, from);
+ kunmap_atomic( kaddr, KM_USER0);
+ SetPageUptodate(prepared_pages[0]);
+ }
+ if ( to != PAGE_CACHE_SIZE ) { /* Last page needs to be partially zeroed */
+ char *kaddr = kmap_atomic(prepared_pages[num_pages-1], KM_USER0);
+ memset(kaddr+to, 0, PAGE_CACHE_SIZE - to);
+ kunmap_atomic( kaddr, KM_USER0);
+ SetPageUptodate(prepared_pages[num_pages-1]);
+ }
+
+ /* Since all blocks are new - use already calculated value */
+ return blocks;
+ }
+
+ /* Well, since we write somewhere into the middle of a file, there is
+ possibility we are writing over some already allocated blocks, so
+ let's map these blocks and substract number of such blocks out of blocks
+ we need to allocate (calculated above) */
+ /* Mask write position to start on blocksize, we do it out of the
+ loop for performance reasons */
+ pos &= ~(inode->i_sb->s_blocksize - 1);
+ /* Set cpu key to the starting position in a file (on left block boundary)*/
+ make_cpu_key (&key, inode, 1 + ((pos) & ~(inode->i_sb->s_blocksize - 1)), TYPE_ANY, 3/*key length*/);
+
+ reiserfs_write_lock(inode->i_sb); // We need that for at least search_by_key()
+ for ( i = 0; i < num_pages ; i++ ) {
+ int item_pos=-1; /* Position in indirect item */
+
+ head = page_buffers(prepared_pages[i]);
+ /* For each buffer in the page */
+ for(bh = head, block_start = 0; bh != head || !block_start;
+ block_start=block_end, bh = bh->b_this_page) {
+ if (!bh)
+ reiserfs_panic(inode->i_sb, "green-9002: Allocated but absent buffer for a page?");
+ /* Find where this buffer ends */
+ block_end = block_start+inode->i_sb->s_blocksize;
+ if (i == 0 && block_end <= from )
+ /* if this buffer is before requested data to map, skip it*/
+ continue;
+
+ if (i == num_pages - 1 && block_start >= to) {
+ /* If this buffer is after requested data to map, abort
+ processing of current page */
+ break;
+ }
+
+ if ( buffer_mapped(bh) && bh->b_blocknr !=0 ) {
+ /* This is optimisation for a case where buffer is mapped
+ and have blocknumber assigned. In case significant amount
+ of such buffers are present, we may avoid some amount
+ of search_by_key calls.
+ Probably it would be possible to move parts of this code
+ out of BKL, but I afraid that would overcomplicate code
+ without any noticeable benefit.
+ */
+ item_pos++;
+ /* Update the key */
+ set_cpu_key_k_offset( &key, cpu_key_k_offset(&key) + inode->i_sb->s_blocksize);
+ blocks--; // Decrease the amount of blocks that need to be
+ // allocated
+ continue; // Go to the next buffer
+ }
+
+ if ( !itembuf || /* if first iteration */
+ item_pos >= ih_item_len(ih)/UNFM_P_SIZE)
+ { /* or if we progressed past the
+ current unformatted_item */
+ /* Try to find next item */
+ res = search_for_position_by_key(inode->i_sb, &key, &path);
+ /* Abort if no more items */
+ if ( res != POSITION_FOUND )
+ break;
+
+ /* Update information about current indirect item */
+ itembuf = get_last_bh( &path );
+ ih = get_ih( &path );
+ item = get_item( &path );
+ item_pos = path.pos_in_item;
+
+ RFALSE( !is_indirect_le_ih (ih), "green-9003: indirect item expected");
+ }
+
+ /* See if there is some block associated with the file
+ at that position, map the buffer to this block */
+ if ( get_block_num(item,item_pos) ) {
+ map_bh(bh, inode->i_sb, get_block_num(item,item_pos));
+ blocks--; // Decrease the amount of blocks that need to be
+ // allocated
+ }
+ item_pos++;
+ /* Update the key */
+ set_cpu_key_k_offset( &key, cpu_key_k_offset(&key) + inode->i_sb->s_blocksize);
+ }
+ }
+ pathrelse(&path); // Free the path
+ reiserfs_write_unlock(inode->i_sb);
+
+ /* Now zero out unmappend buffers for the first and last pages of
+ write area or issue read requests if page is mapped. */
+ /* First page, see if it is not uptodate */
+ if ( !PageUptodate(prepared_pages[0]) ) {
+ head = page_buffers(prepared_pages[0]);
+
+ /* For each buffer in page */
+ for(bh = head, block_start = 0; bh != head || !block_start;
+ block_start=block_end, bh = bh->b_this_page) {
+
+ if (!bh)
+ reiserfs_panic(inode->i_sb, "green-9002: Allocated but absent buffer for a page?");
+ /* Find where this buffer ends */
+ block_end = block_start+inode->i_sb->s_blocksize;
+ if ( block_end <= from )
+ /* if this buffer is before requested data to map, skip it*/
+ continue;
+ if ( block_start < from ) { /* Aha, our partial buffer */
+ if ( buffer_mapped(bh) ) { /* If it is mapped, we need to
+ issue READ request for it to
+ not loose data */
+ ll_rw_block(READ, 1, &bh);
+ *wait_bh++=bh;
+ } else { /* Not mapped, zero it */
+ char *kaddr = kmap_atomic(prepared_pages[0], KM_USER0);
+ memset(kaddr+block_start, 0, from-block_start);
+ kunmap_atomic( kaddr, KM_USER0);
+ set_bit(BH_Uptodate, &bh->b_state);
+ }
+ }
+ }
+ }
+
+ /* Last page, see if it is not uptodate, or if the last page is past the end of the file. */
+ if ( !PageUptodate(prepared_pages[num_pages-1]) ||
+ ((pos+write_bytes)>>PAGE_CACHE_SHIFT) > (inode->i_size>>PAGE_CACHE_SHIFT) ) {
+ head = page_buffers(prepared_pages[num_pages-1]);
+
+ /* for each buffer in page */
+ for(bh = head, block_start = 0; bh != head || !block_start;
+ block_start=block_end, bh = bh->b_this_page) {
+
+ if (!bh)
+ reiserfs_panic(inode->i_sb, "green-9002: Allocated but absent buffer for a page?");
+ /* Find where this buffer ends */
+ block_end = block_start+inode->i_sb->s_blocksize;
+ if ( block_start >= to )
+ /* if this buffer is after requested data to map, skip it*/
+ break;
+ if ( block_end > to ) { /* Aha, our partial buffer */
+ if ( buffer_mapped(bh) ) { /* If it is mapped, we need to
+ issue READ request for it to
+ not loose data */
+ ll_rw_block(READ, 1, &bh);
+ *wait_bh++=bh;
+ } else { /* Not mapped, zero it */
+ char *kaddr = kmap_atomic(prepared_pages[num_pages-1], KM_USER0);
+ memset(kaddr+to, 0, block_end-to);
+ kunmap_atomic( kaddr, KM_USER0);
+ set_bit(BH_Uptodate, &bh->b_state);
+ }
+ }
+ }
+ }
+
+ /* Wait for read requests we made to happen, if necessary */
+ while(wait_bh > wait) {
+ wait_on_buffer(*--wait_bh);
+ if (!buffer_uptodate(*wait_bh)) {
+ res = -EIO;
+ goto failed_read;
+ }
+ }
+
+ return blocks;
+failed_page_grabbing:
+ num_pages = i;
+failed_read:
+ reiserfs_unprepare_pages(prepared_pages, num_pages);
+ return res;
+}
+
+/* Write @count bytes at position @ppos in a file indicated by @file
+ from the buffer @buf.
+
+ generic_file_write() is only appropriate for filesystems that are not seeking to optimize performance and want
+ something simple that works. It is not for serious use by general purpose filesystems, excepting the one that it was
+ written for (ext2/3). This is for several reasons:
+
+ * It has no understanding of any filesystem specific optimizations.
+
+ * It enters the filesystem repeatedly for each page that is written.
+
+ * It depends on reiserfs_get_block() function which if implemented by reiserfs performs costly search_by_key
+ * operation for each page it is supplied with. By contrast reiserfs_file_write() feeds as much as possible at a time
+ * to reiserfs which allows for fewer tree traversals.
+
+ * Each indirect pointer insertion takes a lot of cpu, because it involves memory moves inside of blocks.
+
+ * Asking the block allocation code for blocks one at a time is slightly less efficient.
+
+ All of these reasons for not using only generic file write were understood back when reiserfs was first miscoded to
+ use it, but we were in a hurry to make code freeze, and so it couldn't be revised then. This new code should make
+ things right finally.
+
+ Future Features: providing search_by_key with hints.
+
+*/
+ssize_t reiserfs_file_write( struct file *file, /* the file we are going to write into */
+ const char *buf, /* pointer to user supplied data
+(in userspace) */
+ size_t count, /* amount of bytes to write */
+ loff_t *ppos /* pointer to position in file that we start writing at. Should be updated to
+ * new current position before returning. */ )
+{
+ size_t already_written = 0; // Number of bytes already written to the file.
+ loff_t pos; // Current position in the file.
+ size_t res; // return value of various functions that we call.
+ unsigned long limit = current->rlim[RLIMIT_FSIZE].rlim_cur; // Current filesize limit
+ struct inode *inode = file->f_dentry->d_inode; // Inode of the file that we are writing to.
+ struct page * prepared_pages[REISERFS_WRITE_PAGES_AT_A_TIME];
+ /* To simplify coding at this time, we store
+ locked pages in array for now */
+ if ( count <= PAGE_CACHE_SIZE || file->f_flags & O_DIRECT)
+ return generic_file_write(file, buf, count, ppos);
+
+ if ( unlikely((ssize_t) count < 0 ))
+ return -EINVAL;
+
+ if (unlikely(!access_ok(VERIFY_READ, buf, count)))
+ return -EFAULT;
+
+ down(&inode->i_sem); // locks the entire file for just us
+
+ // We are going to duplicate a lot of generic_file_write checks here
+ // for now. I do not have any good idea on how to avoid these now.
+
+ res = -EINVAL;
+ if ( *ppos < 0)
+ goto out;
+
+ res = file->f_error;
+ if ( res ) {
+ file->f_error = 0;
+ goto out;
+ }
+
+ if (file->f_flags & O_APPEND)
+ pos=inode->i_size;
+ else
+ pos=*ppos;
+
+ res = -EFBIG;
+ if ( limit != RLIM_INFINITY) {
+ if (pos >= limit) {
+ send_sig(SIGXFSZ, current, 0);
+ goto out;
+ }
+ if ( pos > 0xFFFFFFFFULL || count > limit - (u32)pos) {
+ count = limit - (u32)pos;
+ }
+ }
+
+ // LFS
+ if ( pos + count > MAX_NON_LFS && !(file->f_flags & O_LARGEFILE) ) {
+ if ( pos >= MAX_NON_LFS ) {
+ send_sig(SIGXFSZ, current, 0);
+ goto out;
+ }
+ if ( count > MAX_NON_LFS - (u32)pos ) {
+ count = MAX_NON_LFS - (u32)pos;
+ }
+ }
+
+ // Check we are not going to exceed block limit
+ if ( pos >= inode->i_sb->s_maxbytes) {
+ if ( count || pos > inode->i_sb->s_maxbytes) {
+ send_sig(SIGXFSZ, current, 0);
+ goto out;
+ }
+ // zero length writes are ok.
+ }
+
+ res = 0;
+ if ( count == 0 )
+ goto out;
+
+ remove_suid(file->f_dentry);
+ inode_update_time(inode, 1); /* Both mtime and ctime */
+
+ // Ok, we are done with all the checks.
+
+ // Now we should start real work
+
+ /* If we are going to write past the file's packed tail or if we are going
+ to overwrite part of the tail, we need that tail to be converted into
+ unformatted node */
+ res = reiserfs_check_for_tail_and_convert( inode, pos, count);
+ if (res)
+ goto out;
+
+ while ( count > 0) {
+ /* This is the main loop in which we running until some error occures
+ or until we write all of the data. */
+ int num_pages;/* amount of pages we are going to write this iteration */
+ int write_bytes; /* amount of bytes to write during this iteration */
+ int blocks_to_allocate; /* how much blocks we need to allocate for
+ this iteration */
+
+ /* (pos & (PAGE_CACHE_SIZE-1)) is an idiom for offset into a page of pos*/
+ num_pages = !!((pos+count) & (PAGE_CACHE_SIZE - 1)) + /* round up partial
+ pages */
+ ((count + (pos & (PAGE_CACHE_SIZE-1))) >> PAGE_CACHE_SHIFT);
+ /* convert size to amount of
+ pages */
+ reiserfs_write_lock(inode->i_sb);
+ if ( num_pages > REISERFS_WRITE_PAGES_AT_A_TIME
+ || num_pages > reiserfs_can_fit_pages(inode->i_sb) ) {
+ /* If we were asked to write more data than we want to or if there
+ is not that much space, then we shorten amount of data to write
+ for this iteration. */
+ num_pages = min_t(int, REISERFS_WRITE_PAGES_AT_A_TIME, reiserfs_can_fit_pages(inode->i_sb));
+ /* Also we should not forget to set size in bytes accordingly */
+ write_bytes = num_pages * PAGE_CACHE_SIZE -
+ (pos & (PAGE_CACHE_SIZE-1));
+ /* If position is not on the
+ start of the page, we need
+ to substract the offset
+ within page */
+ } else
+ write_bytes = count;
+
+ /* reserve the blocks to be allocated later, so that later on
+ we still have the space to write the blocks to */
+ reiserfs_claim_blocks_to_be_allocated(inode->i_sb, num_pages * (PAGE_CACHE_SIZE/inode->i_sb->s_blocksize));
+ reiserfs_write_unlock(inode->i_sb);
+
+ if ( !num_pages ) { /* If we do not have enough space even for */
+ res = -ENOSPC; /* single page, return -ENOSPC */
+ if ( pos > (inode->i_size & (inode->i_sb->s_blocksize-1)))
+ break; // In case we are writing past the file end, break.
+ // Otherwise we are possibly overwriting the file, so
+ // let's set write size to be equal or less than blocksize.
+ // This way we get it correctly for file holes.
+ // But overwriting files on absolutelly full volumes would not
+ // be very efficient. Well, people are not supposed to fill
+ // 100% of disk space anyway.
+ write_bytes = min_t(int, count, inode->i_sb->s_blocksize - (pos & (inode->i_sb->s_blocksize - 1)));
+ num_pages = 1;
+ }
+
+ /* Prepare for writing into the region, read in all the
+ partially overwritten pages, if needed. And lock the pages,
+ so that nobody else can access these until we are done.
+ We get number of actual blocks needed as a result.*/
+ blocks_to_allocate = reiserfs_prepare_file_region_for_write(inode, pos, num_pages, write_bytes, prepared_pages);
+ if ( blocks_to_allocate < 0 ) {
+ res = blocks_to_allocate;
+ reiserfs_release_claimed_blocks(inode->i_sb, num_pages * (PAGE_CACHE_SIZE/inode->i_sb->s_blocksize));
+ break;
+ }
+
+ /* First we correct our estimate of how many blocks we need */
+ reiserfs_release_claimed_blocks(inode->i_sb, num_pages * (PAGE_CACHE_SIZE>>inode->i_sb->s_blocksize_bits) - blocks_to_allocate );
+
+ if ( blocks_to_allocate > 0) {/*We only allocate blocks if we need to*/
+ /* Fill in all the possible holes and append the file if needed */
+ res = reiserfs_allocate_blocks_for_region(inode, pos, num_pages, write_bytes, prepared_pages, blocks_to_allocate);
+ } else if ( pos + write_bytes > inode->i_size ) {
+ /* File might have grown even though no new blocks were added */
+ inode->i_size = pos + write_bytes;
+ inode->i_sb->s_op->dirty_inode(inode);
+ }
+
+ /* well, we have allocated the blocks, so it is time to free
+ the reservation we made earlier. */
+ reiserfs_release_claimed_blocks(inode->i_sb, blocks_to_allocate);
+ if ( res ) {
+ reiserfs_unprepare_pages(prepared_pages, num_pages);
+ break;
+ }
+
+/* NOTE that allocating blocks and filling blocks can be done in reverse order
+ and probably we would do that just to get rid of garbage in files after a
+ crash */
+
+ /* Copy data from user-supplied buffer to file's pages */
+ res = reiserfs_copy_from_user_to_file_region(pos, num_pages, write_bytes, prepared_pages, buf);
+ if ( res ) {
+ reiserfs_unprepare_pages(prepared_pages, num_pages);
+ break;
+ }
+
+ /* Send the pages to disk and unlock them. */
+ res = reiserfs_submit_file_region_for_write(pos, num_pages, write_bytes, prepared_pages);
+ if ( res )
+ break;
+
+ already_written += write_bytes;
+ buf += write_bytes;
+ *ppos = pos += write_bytes;
+ count -= write_bytes;
+ }
+
+ if ((file->f_flags & O_SYNC) || IS_SYNC(inode))
+ res = generic_osync_inode(inode, OSYNC_METADATA|OSYNC_DATA);
+
+ up(&inode->i_sem);
+ return (already_written != 0)?already_written:res;
+
+out:
+ up(&inode->i_sem); // unlock the file on exit.
+ return res;
+}
+
struct file_operations reiserfs_file_operations = {
.read = generic_file_read,
- .write = generic_file_write,
+ .write = reiserfs_file_write,
.ioctl = reiserfs_ioctl,
.mmap = generic_file_mmap,
.release = reiserfs_file_release,
diff -Nru a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
--- a/fs/reiserfs/inode.c Sat Dec 7 16:58:48 2002
+++ b/fs/reiserfs/inode.c Sat Dec 7 16:58:48 2002
@@ -759,7 +759,11 @@
pointer to 'block'-th block use block, which is already
allocated */
struct cpu_key tmp_key;
- struct unfm_nodeinfo un = {0, 0};
+ unp_t unf_single=0; // We use this in case we need to allocate only
+ // one block which is a fastpath
+ unp_t *un;
+ __u64 max_to_insert=MAX_ITEM_LEN(inode->i_sb->s_blocksize)/UNFM_P_SIZE;
+ __u64 blocks_needed;
RFALSE( pos_in_item != ih_item_len(ih) / UNFM_P_SIZE,
"vs-804: invalid position for append");
@@ -768,30 +772,58 @@
le_key_k_offset (version, &(ih->ih_key)) + op_bytes_number (ih, inode->i_sb->s_blocksize),
//pos_in_item * inode->i_sb->s_blocksize,
TYPE_INDIRECT, 3);// key type is unimportant
-
- if (cpu_key_k_offset (&tmp_key) == cpu_key_k_offset (&key)) {
+
+ blocks_needed = 1 + ((cpu_key_k_offset (&key) - cpu_key_k_offset (&tmp_key)) >> inode->i_sb->s_blocksize_bits);
+ RFALSE( blocks_needed < 0, "green-805: invalid offset");
+
+ if ( blocks_needed == 1 ) {
+ un = &unf_single;
+ } else {
+ un=kmalloc( min(blocks_needed,max_to_insert)*UNFM_P_SIZE,
+ GFP_ATOMIC); // We need to avoid scheduling.
+ if ( !un) {
+ un = &unf_single;
+ blocks_needed = 1;
+ max_to_insert = 0;
+ } else
+ memset(un, 0, UNFM_P_SIZE * min(blocks_needed,max_to_insert));
+ }
+ if ( blocks_needed <= max_to_insert) {
/* we are going to add target block to the file. Use allocated
block for that */
- un.unfm_nodenum = cpu_to_le32 (allocated_block_nr);
+ un[blocks_needed-1] = cpu_to_le32 (allocated_block_nr);
set_block_dev_mapped (bh_result, allocated_block_nr, inode);
set_buffer_new(bh_result);
done = 1;
} else {
/* paste hole to the indirect item */
+ /* If kmalloc failed, max_to_insert becomes zero and it means we
+ only have space for one block */
+ blocks_needed=max_to_insert?max_to_insert:1;
}
- retval = reiserfs_paste_into_item (&th, &path, &tmp_key, (char *)&un, UNFM_P_SIZE);
+ retval = reiserfs_paste_into_item (&th, &path, &tmp_key, (char *)un, UNFM_P_SIZE * blocks_needed);
+
+ if (blocks_needed != 1)
+ kfree(un);
+
if (retval) {
reiserfs_free_block (&th, allocated_block_nr);
goto failure;
}
- if (un.unfm_nodenum)
+ if (done) {
inode->i_blocks += inode->i_sb->s_blocksize / 512;
+ } else {
+ /* We need to mark new file size in case this function will be
+ interrupted/aborted later on. And we may do this only for
+ holes. */
+ inode->i_size += inode->i_sb->s_blocksize * blocks_needed;
+ }
//mark_tail_converted (inode);
}
-
+
if (done == 1)
break;
-
+
/* this loop could log more blocks than we had originally asked
** for. So, we have to allow the transaction to end if it is
** too big or too full. Update the inode so things are
@@ -863,7 +895,7 @@
copy_key (INODE_PKEY (inode), &(ih->ih_key));
- inode->i_blksize = PAGE_SIZE;
+ inode->i_blksize = PAGE_SIZE*32;
INIT_LIST_HEAD(&(REISERFS_I(inode)->i_prealloc_list ));
REISERFS_I(inode)->i_flags = 0;
@@ -1554,7 +1586,7 @@
}
// these do not go to on-disk stat data
inode->i_ino = le32_to_cpu (ih.ih_key.k_objectid);
- inode->i_blksize = PAGE_SIZE;
+ inode->i_blksize = PAGE_SIZE*32;
// store in in-core inode the key of stat data and version all
// object items will have (directory items will have old offset
diff -Nru a/fs/reiserfs/super.c b/fs/reiserfs/super.c
--- a/fs/reiserfs/super.c Sat Dec 7 16:58:48 2002
+++ b/fs/reiserfs/super.c Sat Dec 7 16:58:48 2002
@@ -1313,6 +1313,7 @@
reiserfs_proc_register( s, "oidmap", reiserfs_oidmap_in_proc );
reiserfs_proc_register( s, "journal", reiserfs_journal_in_proc );
init_waitqueue_head (&(sbi->s_wait));
+ sbi->bitmap_lock = SPIN_LOCK_UNLOCKED;
return (0);
diff -Nru a/fs/reiserfs/tail_conversion.c b/fs/reiserfs/tail_conversion.c
--- a/fs/reiserfs/tail_conversion.c Sat Dec 7 16:58:48 2002
+++ b/fs/reiserfs/tail_conversion.c Sat Dec 7 16:58:48 2002
@@ -30,7 +30,7 @@
key of unfm pointer to be pasted */
int n_blk_size,
n_retval; /* returned value for reiserfs_insert_item and clones */
- struct unfm_nodeinfo unfm_ptr; /* Handle on an unformatted node
+ unp_t unfm_ptr; /* Handle on an unformatted node
that will be inserted in the
tree. */
@@ -59,8 +59,7 @@
p_le_ih = PATH_PITEM_HEAD (path);
- unfm_ptr.unfm_nodenum = cpu_to_le32 (unbh->b_blocknr);
- unfm_ptr.unfm_freespace = 0; // ???
+ unfm_ptr = cpu_to_le32 (unbh->b_blocknr);
if ( is_statdata_le_ih (p_le_ih) ) {
/* Insert new indirect item. */
diff -Nru a/include/linux/reiserfs_fs.h b/include/linux/reiserfs_fs.h
--- a/include/linux/reiserfs_fs.h Sat Dec 7 16:58:48 2002
+++ b/include/linux/reiserfs_fs.h Sat Dec 7 16:58:48 2002
@@ -2110,6 +2110,7 @@
#endif
void reiserfs_claim_blocks_to_be_allocated( struct super_block *sb, int blocks);
void reiserfs_release_claimed_blocks( struct super_block *sb, int blocks);
+int reiserfs_can_fit_pages(struct super_block *sb);
/* hashes.c */
__u32 keyed_hash (const signed char *msg, int len);
diff -Nru a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h
--- a/include/linux/reiserfs_fs_sb.h Sat Dec 7 16:58:48 2002
+++ b/include/linux/reiserfs_fs_sb.h Sat Dec 7 16:58:48 2002
@@ -397,6 +397,7 @@
reiserfs_proc_info_data_t s_proc_info_data;
struct proc_dir_entry *procdir;
int reserved_blocks; /* amount of blocks reserved for further allocations */
+ spinlock_t bitmap_lock; /* this lock on now only used to protect reserved_blocks variable */
};
/* Definitions of reiserfs on-disk properties: */
diff -Nru a/kernel/ksyms.c b/kernel/ksyms.c
--- a/kernel/ksyms.c Sat Dec 7 16:58:48 2002
+++ b/kernel/ksyms.c Sat Dec 7 16:58:48 2002
@@ -217,6 +217,7 @@
EXPORT_SYMBOL(generic_cont_expand);
EXPORT_SYMBOL(cont_prepare_write);
EXPORT_SYMBOL(generic_commit_write);
+EXPORT_SYMBOL(block_commit_write);
EXPORT_SYMBOL(block_truncate_page);
EXPORT_SYMBOL(generic_block_bmap);
EXPORT_SYMBOL(generic_file_read);
@@ -544,6 +545,8 @@
EXPORT_SYMBOL(make_bad_inode);
EXPORT_SYMBOL(is_bad_inode);
EXPORT_SYMBOL(__inode_dir_notify);
+EXPORT_SYMBOL(generic_osync_inode);
+EXPORT_SYMBOL(remove_suid);
#ifdef CONFIG_UID16
EXPORT_SYMBOL(overflowuid);
^ permalink raw reply
* Re: 2.5.5[01]]: Xircom Cardbus broken (PCI resource collisions)
From: Valdis.Kletnieks @ 2002-12-13 18:46 UTC (permalink / raw)
To: Dave Jones; +Cc: Petr Konecny, linux-kernel
In-Reply-To: <20021213173656.GC1633@suse.de>
[-- Attachment #1: Type: text/plain, Size: 1560 bytes --]
On Fri, 13 Dec 2002 17:36:56 GMT, Dave Jones said:
> It's my understanding that pci_enable_device() *must* be called
> before we fiddle with dev->resource, dev->irq and the like.
OK.. it looks like the problem only hits if it's a PCMCIA card *with an
onboard ROM*.
The immediate problem cause is in pcibios_enable_resources():
if (!r->start && r->end) {
printk(KERN_ERR "PCI: Device %s not available because of resource collisions\n", dev->slot_name);
If there's an onboard ROM, this is set up in pcibios_assign_resources():
if (pci_probe & PCI_ASSIGN_ROMS) {
r = &dev->resource[PCI_ROM_RESOURCE];
r->end -= r->start;
r->start = 0;
if (r->end)
pci_assign_resource(dev, PCI_ROM_RESOURCE);
}
but this hasn't happened yet in cb_alloc():
/* FIXME: Do we need to enable the expansion ROM? */
for (r = 0; r < 7; r++) {
struct resource *res = dev->resource + r;
if (res->flags)
pci_assign_resource(dev, r);
}
So I think right *AFTER* this code is the right place for pci_enable_device()
because otherwise we won't have allocated the ROMs, so we'll get conflicts.
/Valdis (who went nuts looking at <6 and <7 all over the place. Should I
volunteer to clean these up to #defines rather than inline magic numbers? ;)
[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]
^ permalink raw reply
* Re: [PATCH] S4Bios support for 2.4.20 + acpi-20021205
From: Ducrot Bruno @ 2002-12-13 18:36 UTC (permalink / raw)
To: Nils Faerber; +Cc: Ducrot Bruno, acpi-devel-pyega4qmqnRoyOMFzWx49A
In-Reply-To: <20021213181316.0b460077.nils-t93Ne7XHvje5bSeCtf/tX7NAH6kLmebB@public.gmane.org>
On Fri, Dec 13, 2002 at 06:13:16PM +0100, Nils Faerber wrote:
> On Fri, 13 Dec 2002 17:36:15 +0100
> Ducrot Bruno <ducrot-kk6yZipjEM5g9hUCZPvPmw@public.gmane.org> wrote:
> > On Fri, Dec 13, 2002 at 05:14:31PM +0100, Nils Faerber wrote:
> > > On Fri, 13 Dec 2002 16:53:09 +0100
> > > Ducrot Bruno <ducrot-kk6yZipjEM5g9hUCZPvPmw@public.gmane.org> wrote:
> > > > Patch for 2.4.20 and acpi-20021205 for adding s4bios feature.
> > Well, I don't know. Perhaps lphdisk (I don't remember the URI) could
> > help. Some googling permit to get it. Or (who knows?) even a
> > /dev/hda1 as a vfat partition can help?
>
> AFAIK this is only good for Phoenix BIOS whereas mine is AWARD
> Medallion. Couldn't this also be specified in ACPI specs, i.e. how a
> suspend partition has to look like?
You will find nothing. This is not ACPI problem, but a BIOS one.
>
> > > does it suspend? I know for Phoenix BIOSes there is a tool on the
> > > net to create that partition. I have a Asus L3800C notebook which
> > > has AFAIK an AWARD BIOS. And AFAIK there is no tool available from
> > > ASUS or AWARD to create such a partition.
> > In a really perfect world:
> > echo 1 > /proc/acpi/sleep # for standby
> > echo 2 > /proc/acpi/sleep # for suspend to ram
> > echo 3 > /proc/acpi/sleep # for suspend to ram, but with more power conservative
> > echo 4 > /proc/acpi/sleep # for suspend to disk
> > echo 5 > /proc/acpi/sleep # for shutdown unfriendly the system
I wanted only to speak for a 'perfect world', not for what is the patch.
My apologies for confusion.
>
> That would really be perfect :)
> S1 on mine does just switch of the fan and stops the CPU. Nothing more.
> S2 does not exist.
> S3 does nothing except for funny ACPI messages:
> acpi: sleep 3
> acpi: GO!
> acpi: acpi_suspend call 3
>
> And your S4BIOS patch ends up freezing the machine after "saving CPU
> context". Oh, but S5 behaves as expected :)
My patch is NOT for S3.
S4bios requrire that the bios have necessary support. I check that
it has before going to s4bios. Could you please :
get pmtools from Intel site:
http://developer.intel.com/technology/iapc/acpi/downloads.htm
Then under pmtools-20010730:
make
cd acpidmp
Could you then send me output of:
./acpidmp FACP | ./acpitbl
and
./acpidmp FACS | ./acpitbl
I check for those values:
For ./acpidmp FACP | ./acpitbl
S4BIOS_REQ is not 0;
For ./acpidmp FACS | ./acpitbl:
Flags have bit 0 set.
I supposed that the latest check would be sufficient?
--
Ducrot Bruno
-- Which is worse: ignorance or apathy?
-- Don't know. Don't care.
-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
^ permalink raw reply
* Re: Re: [LARTC] VRRPD (rfc2338)
From: Dmitry Golubev @ 2002-12-13 18:35 UTC (permalink / raw)
To: lartc
In-Reply-To: <marc-lartc-103970435030477@msgid-missing>
Hello,
And then bridge these interfaces together? I do not want to use any king of
switch simply to make the card respond for multiple MAC addrs. Moreover
I have seen VLAN probems with some eth cards that cannot work with
1504 byte packets
Anyway, Alexandre claims VLAN solution hacky. Do You use it in production?
BR, Dmitry
==== At 2002-12-12, 15:05:00 you wrote: ===
>Dmitry Golubev wrote:
>
>>But as far as I know, there is no VRRP implementations that fully comply
>>with rfc2338 as it requires multiple MAC addresses for the one poor linux
>>box's interface. Maybe, someone can suggest a working solution of this
>>problem?
>>
>Yes, there is a way -- the VLAN code in the linux kernel supports
>setting the MAC address of virtual interfaces (eth0.5, for instance).
>AFAIC, this is much superior (in concept) to multicast MACs, given the
>Cisco problem.
>
>>
>>I have seen one idea, but haven't tested it yet (hope someone can try it out):
>>
>>To bridge the physical iface with TAP on which the vrrpd (or keepalived) is
>>running. In that case we could make the VRRP-router that fully comply with RFC.
>>
>>For more info see: http://www.math.leidenuniv.nl/pipermail/bridge/2002-June/002021.html
>>
>>BR, Dmitry
>>
>>==== At 2002-12-11, 03:56:00 you wrote: ===>>
>>
>>
>>>The daemon at http://www.keepalived.org/ is the VRRPd implementation
>>>that's supposed to be the best. It's actually part of the Linux Virtual
>>>Server project (layer 4 load balancer), but the author claims you should
>>>be able to use it as a pure VRRP daemon -- although when I've read the
>>>doc, I couldn't figure out how. (But don't be discouraged by my
>>>impatience. :) It's supposed to be the most mature and ready-for-production.
>>>
>>>There's also Jerome Etienne's reference implementation (don't have a
>>>URL, but it's easy to Google). However, I've heard from more than place
>>>that this is too proof-of-concept and perhaps not production-worthy.
>>>Here's a link to a paper about running VRRPd as the hotspare protocol
>>>for linux firewalls (uses Jerome Etienne's implementation):
>>>http://www.gnusec.com/resource/security/docs/HAFirewallLinux-VRRP.pdf.
>>>
>>>BTW, keep in mind that if you intend to use VRRP in an environment with
>>>Cisco routers, you'll need to do some work on them too. Cisco routers do
>>>not accept multicast MAC addresses as legit ARP replies by default.
>>>Unfortunately, the VRRP RFC and all implementations use multicast MACs.
>>>What that means is that you'll need to either 1) turn the switch on the
>>>Cisco routers that makes them accept multicast MAC ARP replies (good),
>>>or 2) put a static ARP entry in the Cisco routers for the VRRP multicast
>>>MACs (better).
>>>
>>>Hope that helps.
>>>
>>>-S
>>>
>>>
>>>Anton Tinchev wrote:
>>>
>>>
>>>
>>>>Can someone point me for good VRRPD (rfc2338) implementation on linux.
>>>>Some stable and live project
>>>>Thanks
>>>>
>>>>_______________________________________________
>>>>LARTC mailing list / LARTC@mailman.ds9a.nl
>>>>http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
>>>>
>>>>
>>>>
>>>>
>>>_______________________________________________
>>>LARTC mailing list / LARTC@mailman.ds9a.nl
>>>http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
>>>
>>>
>>
>>
>>
>>
>>
>>_______________________________________________
>>LARTC mailing list / LARTC@mailman.ds9a.nl
>>http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
>>
>>
_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
^ permalink raw reply
* Re: pci-skeleton duplex check
From: David S. Miller @ 2002-12-13 18:29 UTC (permalink / raw)
To: becker; +Cc: jgarzik, rl, netdev, linux-kernel
In-Reply-To: <Pine.LNX.4.44.0212131109180.1399-100000@beohost.scyld.com>
From: Donald Becker <becker@scyld.com>
Date: Fri, 13 Dec 2002 11:56:17 -0500 (EST)
The development criteria used to be technically based, and that is still
the public statement. Now, as your statement makes clear, working code
is an irrelevant criteria.
No, working code is only part of the equation. If you're a total and
complete asshole, your work is likely to get lost to the sands of
time. In such a case nobody wants to deal with you.
Welcome to the real world where you have to interact with other human
beings (not just be technically capable) in order to accomplish
things.
It's always been like this Donald. If you piss off, or are a jerk to,
the primary maintainers you're going to get the short end of the
stick.
^ permalink raw reply
* Re: Re: [Alsa-user] fm801 driver status?
From: Friedrich Ewaldt @ 2002-12-13 18:24 UTC (permalink / raw)
To: Takashi Iwai; +Cc: Thierry Vignaud, alsa-user, alsa-devel
In-Reply-To: <s5hel8lua27.wl@alsa2.suse.de>
Takashi Iwai wrote:
>>lspci -n output:
>>
>>00:0b.0 Class 0401: 1319:0801 (rev a0)
>>00:0b.1 Class 0980: 1319:0801 (rev a0)
>>
>>
>
>thanks, could you try the attached patch?
>at least, the weird messages for allocation of invalid i/o ports
>should disappear. not sure whether this cures the lock-up problem,
>though.
>
>
>Takashi
>
>
done. The error in /var/log/messages doesn't appear any longer (I only
get these 'sharing IRQ ...' messages). Thanks! But the system still
locks up when stopping playback :-(
fe
-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
^ permalink raw reply
* Re: [PATCH 2.4.21-BK] Fix typo in arch/arm/config.in
From: Russell King @ 2002-12-13 18:29 UTC (permalink / raw)
To: Marc-Christian Petersen; +Cc: linux-kernel, Marcelo Tosatti
In-Reply-To: <200212131844.45280.m.c.p@wolk-project.de>
On Fri, Dec 13, 2002 at 06:47:08PM +0100, Marc-Christian Petersen wrote:
> this fixes a typo in arch/arm/config.in.
>
> old: source drivers/ssi/Config.in
> new: source drivers/scsi/Config.in
>
> Without it, make menuconfig crashes.
Only that? I'm surprised - there's a hell of a lot of outstanding
stuff for 2.4 for ARM.
Bluntly, I'm not interested in reports and fixes against Marcelo's
tree for ARM stuff because its fairly out of date, and a fair amount
of required generic changes didn't make it into what was Linus' tree
before Marcelo took it over.
--
Russell King (rmk@arm.linux.org.uk) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html
^ permalink raw reply
* Re: where could I download the 2.4.20-ac1-ish kernel? thx
From: Marc-Christian Petersen @ 2002-12-13 18:27 UTC (permalink / raw)
To: linux-kernel; +Cc: Hu, Boris
In-Reply-To: <957BD1C2BF3CD411B6C500A0C944CA260216C27D@pdsmsx32.pd.intel.com>
On Friday 13 December 2002 19:20, Hu, Boris wrote:
Hi Boris,
> where could I download the 2.4.20-ac1-ish kernel?
> I have tried
> http://rawhide.redhat.com/pub/redhat/linux/rawhide/SRPMS/SRPMS/ and
> http://www.kernel.org/pub/linux/kernel/people/alan/linux-2.4/2.4.20/. but
> couldn't find it. :(
Consider buying new glasses ;)
ftp://ftp.kernel.org/pub/linux/kernel/people/alan/linux-2.4/2.4.20/patch-2.4.20-ac2.bz2
ciao, Marc
^ permalink raw reply
* Re: usbaudio won't do 24-bit or 32-bit i/o, and won't do 96000 frames per second
From: Takashi Iwai @ 2002-12-13 18:19 UTC (permalink / raw)
To: Patrick Shirkey; +Cc: Clemens Ladisch, John S. Denker, alsa-devel
In-Reply-To: <3DF0482E.8010109@boosthardware.com>
Hi Patrick,
At Fri, 06 Dec 2002 15:48:14 +0900,
Patrick Shirkey wrote:
>
> Clemens Ladisch wrote:
> > John S. Denker wrote:
> >
> >>Request #1: USB driver supporting 24-bit i/o.
> >>Request #2: USB driver supporting 96000 frames per second.
> >
> >
> > What you want is already supported by the snd-usb-audio driver. IIRC
> > Patrick Shirkey reported that 24bit@96kHz works with the M-Audio Quattro.
> >
>
> Actually I have just yesterday noticed a specific problem with the 24
> bit support.
>
> Also you should know that the quattro only supports 24_3le which is 24
> bits, three bytes. This is contrary to possibly all other pro devices
> which support 24bits, four bytes or something like that. Meaning that
> you have to use a special bit depth just for the quattro. This could be
> the problem you are seeing.
>
> Currently I am using the cvs from the 18 November. I will update and
> check again. I have initialised both pcms and I cannot record a signal
> through the first pcm hw:1,0 although arecord doesn't complain.
could you tell me the rcs version numbers of the files on
alsa-kernel/usb you are using (18 Nov.) ? i've checked the files via
cvs but i couldn't see any differences around the date.
Takashi
-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
^ permalink raw reply
* Re: [PATCH 2.4.21-BK] Fix typo in arch/arm/config.in
From: Russell King @ 2002-12-13 18:26 UTC (permalink / raw)
To: Marc-Christian Petersen; +Cc: linux-kernel, Marcelo Tosatti
In-Reply-To: <200212131859.16039.m.c.p@wolk-project.de>
On Fri, Dec 13, 2002 at 06:59:16PM +0100, Marc-Christian Petersen wrote:
> if [ "$CONFIG_ARCH_CLPS711X" = "y" ]; then
> source drivers/ssi/Config.in
> fi
>
> drivers/ssi/Config.in does not exist, make menuconfig crashes.
> I thought it is a typo, but source'ing it twice also crashes, for sure.
>
> So what is drivers/ssi/* ?
Its something that isn't merged, and something that I lost access to the
hardware to complete, so it isn't likely to be merged.
--
Russell King (rmk@arm.linux.org.uk) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html
^ permalink raw reply
* Re: Integrating Linux Desktop With MS Exchange Server
From: Jorge R . Csapo @ 2002-12-13 18:18 UTC (permalink / raw)
To: Kumar, Pradeep (MED, TCS); +Cc: 'linux-admin@vger.kernel.org'
In-Reply-To: <8608421EC5CBD511B5090002A55C00480478E219@uswaumsx04medge.med.ge.com>
According to Microsoft support in my area, the only way to do this is buying
SFU (Services For Unix, IIRC) and install that on your PDC, then use NIS on the
Linux side. The PDC shouls appear as a NIS Master Server.
Probably there's some way of doing this with Samba, though. I think you'd better
start with the Samba documentation.
Finally, my own opinion: this should probably be done the other way around, with
Linux as the authentication server.
Jorge
assim falou Kumar, Pradeep (MED, TCS) (em 13/12/2002):
> Hi,
> I am sorry if I am posting this to a wrong mailing list. In that case,
> please suggest me the correct one.
>
> We are currently running Microsoft Exchange 5.5 server. Now we would
> like to run Red Hat Linux on some desktops.
> But we would like to have a unified logon procedure i.e all the windows
> users should be able to login on the linux desktop
> with their windows username and passwords. Whenever a user wants to
> login to the linux desktop, the authentication
> should be redirected to the Windows Primary Domain Controller. As a
> whole , it is just putting a linux desktop into
> a Windows domain. But all the usernames and passwords are created and
> maintained on the Exchange Server database and
> and authentication of all users should be done on the Exchange server.
>
> Please help me how to proceed on this. I would appreciate any useful
> information provided. What should I do to achieve this?
> What and all should be ruuning at Linux side? How to achieve this
> integration?
>
> Thanks in anticipation,
>
> Regards,
> Pradeep
>
> S.Pradeep Kumar
> Pradeep.Kumar@med.ge.com
> -
> To unsubscribe from this list: send the line "unsubscribe linux-admin" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
Jorge R. Csapo
--------------------------------------------------
/"\
\ / CAMPANHA DA FITA ASCII - CONTRA MAIL HTML
X ASCII RIBBON CAMPAIGN - AGAINST HTML MAIL
/ \
--------------------------------------------------
http://www.completo.com.br/~jorge
===========================================
With a PC, I always felt limited
by the software available.
On Unix, I am limited only by my knowledge.
--Peter J. Schoenster
^ permalink raw reply
* Re: [PATCH] Add CONFIG_ACPI_RELAXED_AML option
From: Ducrot Bruno @ 2002-12-13 18:15 UTC (permalink / raw)
To: Carlos Morgado; +Cc: acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f
In-Reply-To: <20021213171815.GS5382-V9yvzcrGID8XxY44YfPCZCanxOoIfzq+@public.gmane.org>
On Fri, Dec 13, 2002 at 05:18:15PM +0000, Carlos Morgado wrote:
>
> On 2002.12.13 16:22:43 +0000 "Moore, Robert" wrote:
>
> >
> >What is really being proposed here is for Linux ACPI to be bug-for-bug
> >compatible with Microsoft. This is impossible to do deterministically
> >because the MS interpreter is closed source. The only standard that we
> >have
> >that we can code to is the ACPI specification, and this has to be the last
> >word on the matter.
> >
>
> not to mention windows tends to ignore builtin dsdts afaik and use the ones
> distributed in windows drivers.
I suppose you think perticuliary to their ACPI EC smbus driver?
Well, I don't think that they use a perticular custom dsdt for that.
--
Ducrot Bruno
-- Which is worse: ignorance or apathy?
-- Don't know. Don't care.
-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
^ permalink raw reply
* where could I download the 2.4.20-ac1-ish kernel? thx
From: Hu, Boris @ 2002-12-13 18:20 UTC (permalink / raw)
To: Linux-Kernel (E-mail)
where could I download the 2.4.20-ac1-ish kernel?
I have tried http://rawhide.redhat.com/pub/redhat/linux/rawhide/SRPMS/SRPMS/
and http://www.kernel.org/pub/linux/kernel/people/alan/linux-2.4/2.4.20/.
but couldn't find it. :(
thanks.
Boris
=========================
To know what I don't know
To learn what I don't know
To contribute what I know
=========================
^ permalink raw reply
* Re: [PATCH] Add CONFIG_ACPI_RELAXED_AML option
From: Matthew Tippett @ 2002-12-13 18:14 UTC (permalink / raw)
To: chbm-tNiY1ywYjSU; +Cc: acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f
In-Reply-To: <20021213171815.GS5382-V9yvzcrGID8XxY44YfPCZCanxOoIfzq+@public.gmane.org>
Carlos Morgado wrote:
>
> On 2002.12.13 16:22:43 +0000 "Moore, Robert" wrote:
>
>>
>> What is really being proposed here is for Linux ACPI to be
bug-for-bug
>> compatible with Microsoft. This is impossible to do
deterministically
>> because the MS interpreter is closed source. The only standard that
>> we have
>> that we can code to is the ACPI specification, and this has to be the
>> last
>> word on the matter.
>>
>
> not to mention windows tends to ignore builtin dsdts afaik and use the
ones
> distributed in windows drivers.
Continuing with this thread, would it make sense to be ACPI compliant
but allow 'custom' dsdts to be passed to the acpi subsystem to allow
users to work around less than perfect implementations from
manufacturers.
Although I am not as clued into the inner workings of ACPI, it may be
extensible to allow developers to highlight to manufacturers where there
dstd is broken and how to have it fixed.
Regards,
Matthew
-----
The information contained in this message is proprietary of Casero Inc.,
protected from disclosure, and may be privileged. The information is
intended to be conveyed only to the designated recipient(s) of the
message. If the reader of this message is not the intended recipient,
you are hereby notified that any dissemination, use, distribution or
copying of this communication is strictly prohibited and may be
unlawful. If you have received this communication in error, please
notify us immediately by replying to the message and deleting it from
your computer. Thank you.
-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
^ permalink raw reply
* fan behaviour
From: David Douard @ 2002-12-13 18:12 UTC (permalink / raw)
To: acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f
Hi,
there have been quite a long time since I've played wit ACPI on my Compaq Evo
N150...
Remember that I nedd a pached DSDT to get many things to work (bat, processor
C states etc.).
I used to have pb with fans : they where not autmatically turned on/off
according to CPU temperature. Thus I made a little daemon to hold that.
But then I tried a nex ACPI version (one month or two ago), and I realized
that fan where handled fine. But this was with the original DSDT.
I recompiled the kernel with my DSDT. And then fans where not handled.
So I compiled the lates acpi patch on a 2.4.20. And I realized that I can make
automatic fan handling if I play with the
/proc/acpi/thermal_zone/THRM/cooling_mode.
The things I have to do so are quite esoteric : set passive cooling mode, make
cpu hot (reach the trip point); then set active cooling mode. (well I guess
the manipulation is that, I'm not very sure).
Then, the fans are nicely handled.
This sound very strange to me, and my question is this : does anyone have any
idea on if that may come from ACPI CA (the AML interpreter maybe), or from
the DSDT code which (I know) is bugged.
Note too another strange thing (I made it twice) :
I set (in single user mode) the cooling mode to passive and then lauch a
little cpuburner program. Throttling started to increase while temperature
reach the passive trip point (this is the first time I see throttling
working. It used to hang the machine if I try to change its value).
But when I killed the cpuburner process, it hanged the machine !
Thanks for any tip !
David Douard
-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
^ permalink raw reply
* [OOPS] 2.5.51-mm2
From: Paul P Komkoff Jr @ 2002-12-13 18:11 UTC (permalink / raw)
To: Linux Kernel Mailing List
This is very funny.
mke2fs -j -O dir_index -J size=192 -T news -N 1000100
atest3 1000000
(creat & write 1 byte to 1000000 files)
free space on device became 0 and voila
Unable to handle kernel paging request at virtual address 5a5a5b9e
printing eip:
c01a5ed2
*pde = 00000000
Oops: 0000
CPU: 1
EIP: 0060:[<c01a5ed2>] Not tainted
EFLAGS: 00010202
EIP is at ext3_get_inode_loc+0x32/0x1a0
eax: c5e9a2d0 ebx: 00000000 ecx: 00000000 edx: 5a5a5a5a
esi: 5a5a5a5a edi: c8999ec8 ebp: c5e9a2d0 esp: c8999e5c
ds: 0068 es: 0068 ss: 0068
Process atest3 (pid: 1482, threadinfo=c8998000 task=c7f17960)
Stack: c5e9a060 c0146111 cbd389b0 c5e9a240 ca50f1f4 cbd32c00 00000000 ca50f1f4
c8999ec8 c5e9a2d0 c01a6c91 c5e9a2d0 c8999ec8 ffffffe4 c57fe3ec cb5b30d0
c01aca2d cbd389b0 c5e9a244 c8999ec8 c5e9a2d0 ca50f1f4 cb5b30d0 c01a6d7b
Call Trace:
[<c0146111>] cache_free_debugcheck+0x131/0x1c0
[<c01a6c91>] ext3_reserve_inode_write+0x31/0xf0
[<c01aca2d>] ext3_destroy_inode+0x1d/0x30
[<c01a6d7b>] ext3_mark_inode_dirty+0x2b/0x60
[<c01aa03b>] ext3_add_nondir+0x5b/0x60
[<c01aa1ae>] ext3_create+0x16e/0x230
[<c016eae7>] permission+0x57/0x70
[<c016ffd6>] vfs_create+0xb6/0x130
[<c0170659>] open_namei+0x3b9/0x420
[<c015da53>] filp_open+0x43/0x70
[<c015df5b>] sys_open+0x5b/0x90
[<c015dfaf>] sys_creat+0x1f/0x30
[<c010998f>] syscall_call+0x7/0xb
Code: 8b 86 44 01 00 00 3b 50 50 72 0d 8b 9e 44 01 00 00 8b 43 2c
--
Paul P 'Stingray' Komkoff 'Greatest' Jr /// (icq)23200764 /// (http)stingr.net
When you're invisible, the only one really watching you is you (my keychain)
^ permalink raw reply
* How do I add a 1:1 port-forwarding capability for a DNAT port-range to NetFilter ?
From: Ranjeet Shetye @ 2002-12-13 18:01 UTC (permalink / raw)
To: netfilter-devel
This is what I posted on the user-mailing list yesterday.
---------------------------------------------------------------
DNAT:
Is it possible to have a SINGLE (DNAT?) rule that will let me do 1:1
port-forwarding over a range of ports while doing Destination NAT.
e.g. Any incoming connections to 64.1.0.20:100-101 need to be mapped to
172.16.0.100:200-201 for the TCP protocol.
such that a connection to port 101 will ALWAYS map to port 201 and a
connection to port 100 will ALWAYS map to port 200.
Under current DNAT port range scenario, the connection goes to the
lowest port that is free e.g. a port 101 connection will be DNATt'ed to
port 200 if port 200 is free.
The reason for wanting a 1:1 rule is for X windows and other fat port
ranges. Dont want hundreds of rules in there if one can do the job. Can
IPTables do it ? If so how ? If not, I guess I'll have to get in touch
with the developers for tips on a good starting point.
---------------------------------------------------------------
The answer I got is that basically I have to modify the nat code. I have
been reading the code for the last 2-3 days but I am competely lost as
to where to start. Should I make the change in ip_nat_setup_info or
get_unique_tuple or manip_pkt ? Any pointers/any help here is
appreciated.
Also, if anyone has an architectural diagram of NetFilter :D (mapping
out the interactions between the various layers/components), that would
be heaven!!!
Thanks in advance,
Ranjeet Shetye
Senior Software Engineer
Zultys Technologies
771 Vaqueros Avenue
Sunnyvale CA 94085
USA
Ranjeet.Shetye@Zultys.com
http://www.zultys.com/
^ permalink raw reply
* Re: R: Kernel bug handling TCP_RTO_MAX?
From: Nivedita Singhvi @ 2002-12-13 18:07 UTC (permalink / raw)
To: Andrew McGregor
Cc: Bogdan Costescu, David S. Miller, dlstevens, matti.aarnio, alan,
stefano.andreani.ap, linux-kernel, linux-net
In-Reply-To: <19000000.1039780134@localhost.localdomain>
Andrew McGregor wrote:
> In a closed network, why not have SOCK_STREAM map to something faster than
> TCP anyway? That is, if I connect(address matching localnet), SOCK_STREAM
> maps to (eg) SCTP. That would be a far more dramatic performance hack!
>
> Andrew
Not that simple. SCTP (if that is what Matti was referring to) is
a SOCK_STREAM socket, with a protocol of IPPROTO_SCTP. I'm just
getting done implementing a testsuite against the SCTP API.
i.e. You have to know you want an SCTP socket at the time you
open the socket. You certainly have no idea whether youre on
a closed network or not, for that matter, the app may want to talk
on multiple interfaces etc. (Most hosts will have one interface
on a public net)..
Currently, Linux SCTP doesn't yet support TCP style i.e SOCK_STREAM
sockets, we only do udp-style sockets (SOCK_SEQPACKET). We will be
putting in SOCK_STREAM support next, but understand that performance
is not something that has been addressed yet, and a performant SCTP
is still some ways away (though I'm sure Jon and Sridhar will be
working their tails off to do so ;)).
But dont expect SCTP to be the surreptitious underlying layer
carrying TCP traffic, if thats an expectation that anyone has :)
Solving this problem without application involvement is a
more limited scenario..
thanks,
Nivedita
^ permalink raw reply
* Re: IDE feature request & problem
From: Milan Roubal @ 2002-12-13 18:03 UTC (permalink / raw)
To: Vojtech Pavlik, Pavel Machek
Cc: Vojtech Pavlik, Alan Cox, Petr Sebor, Linux Kernel Mailing List
In-Reply-To: <20021213154156.A6001@ucw.cz>
Power Supply is not problem for sure,
there are 1x 420 W and 3x 300W HotSwap,
so I think its enough for it.
Milan
----- Original Message -----
From: "Vojtech Pavlik" <vojtech@suse.cz>
To: "Pavel Machek" <pavel@suse.cz>
Cc: "Vojtech Pavlik" <vojtech@suse.cz>; "Alan Cox"
<alan@lxorguk.ukuu.org.uk>; "Milan Roubal"
<roubm9am@barbora.ms.mff.cuni.cz>; "Petr Sebor" <petr@scssoft.com>; "Linux
Kernel Mailing List" <linux-kernel@vger.kernel.org>
Sent: Friday, December 13, 2002 3:41 PM
Subject: Re: IDE feature request & problem
> On Thu, Dec 12, 2002 at 07:12:50PM +0100, Pavel Machek wrote:
> > Hi!
> >
> > > > > I have got xfs partition and man fsck.xfs say
> > > > > that it will run automatically on reboot.
> > > >
> > > > You need to force one. Something (I assume XFS) asked the disk for a
> > > > stupid sector number. Thats mostly likely due to some kind of
internal
> > > > corruption on the XFS
> > >
> > > Or the power supply doesn't give enough power to the drives anymore
(my
> > > 350W PSU is having heavy problems with five or more drives), and the
IDE
> > > transfers get garbled. Note that there is no CRC protection for
non-data
> > > xfers even when UDMA is in use, which includes LBA sector addressing.
> >
> > But kernel would not log bogus LBA in such case.
>
> It could, if the drive has read a different sector than it was supposed
> to and the filesystem got confused by the data ...
>
> --
> Vojtech Pavlik
> SuSE Labs
^ permalink raw reply
* RE: natting specific ports
From: Ranjeet Shetye @ 2002-12-13 17:57 UTC (permalink / raw)
To: netfilter
In-Reply-To: <CAFAAEC91CC8D511952000062938C6F12ECDC0@ozlan.fcdomain.net>
Hi Doug,
Do you want to NAT for traffic coming in or for traffic going out ?
If you want your internal network to be able to reach external telnet
and smtp servers, then your destination port will be 23 or 25, not your
source port.
If you want to host telnet and smtp servers behind a firewall and allow
only NATted access to these servers, then you should be using DNAT, not
SNAT.
Hope this helps,
Ranjeet Shetye
Senior Software Engineer
Zultys Technologies
771 Vaqueros Avenue
Sunnyvale CA 94085
USA
Ranjeet.Shetye@Zultys.com
http://www.zultys.com/
> -----Original Message-----
> From: netfilter-admin@lists.netfilter.org
> [mailto:netfilter-admin@lists.netfilter.org] On Behalf Of
> Simpson, Doug
> Sent: Friday, December 13, 2002 9:49 AM
> To: 'netfilter@lists.netfilter.org'
> Subject: natting specific ports
>
>
> I want to "NAT" just specific ports to my Public IP. Do the
> commands below make sense? I want my internal network to be
> able to telnet and send email. (eth0 is my External NIC - it
> is exposed to the internet)
> iptables -t nat -A POSTROUTING -p tcp --sport 25 -o eth0 -s
> $INTERNAL_IP -j SNAT --to $EXTERNAL_IP iptables -t nat -A
> POSTROUTING -p tcp --sport 23 -o eth0 -s $INTERNAL_IP -j SNAT
> --to $EXTERNAL_IP
>
> Thank you,
> Doug
>
^ permalink raw reply
* Re: 2.5.51: new warning from lilo
From: Andries Brouwer @ 2002-12-13 18:02 UTC (permalink / raw)
To: Pavel Machek; +Cc: kernel list
In-Reply-To: <20021212193451.GA458@elf.ucw.cz>
On Thu, Dec 12, 2002 at 08:34:51PM +0100, Pavel Machek wrote:
> Lilo now presents me with new warning:
>
> Warning: Kernel & BIOS return differing head/sector geometries for
> device 0x80
> Kernel: 8944 cylidners, 15 heads, 63 sectors
> BIOS: 525 cylinders, 255 heads, 63 sectors
>
> lilo did not warn under 2.5.50. Now... Will it boot?
That depends on the options you gave it.
With linear or lba32 the geometry does not play any role.
If you don't give these options then the geometry does play
a role, at least for the versions of LILO I have looked at.
You can give LILO explicit geometry options if for some
reason you do not want to use "linear".
Andries
^ permalink raw reply
* Re: [PATCH 2.4.21-BK] Fix typo in arch/arm/config.in
From: Marc-Christian Petersen @ 2002-12-13 17:59 UTC (permalink / raw)
To: linux-kernel; +Cc: Marcelo Tosatti
In-Reply-To: <200212131844.45280.m.c.p@wolk-project.de>
On Friday 13 December 2002 18:47, Marc-Christian Petersen wrote:
Hi again,
> this fixes a typo in arch/arm/config.in.
> old: source drivers/ssi/Config.in
> new: source drivers/scsi/Config.in
> Without it, make menuconfig crashes.
ignore this patch. It is wrong.
So we have this:
if [ "$CONFIG_SCSI" != "n" ]; then
source drivers/scsi/Config.in
fi
endmenu
if [ "$CONFIG_ARCH_CLPS711X" = "y" ]; then
source drivers/ssi/Config.in
fi
drivers/ssi/Config.in does not exist, make menuconfig crashes.
I thought it is a typo, but source'ing it twice also crashes, for sure.
So what is drivers/ssi/* ?
Or should this be drivers/sgi/Config.in ?
ciao, Marc
^ permalink raw reply
* /proc/acpi/battery empty on a samsung V20
From: Parrenin Frédéric @ 2002-12-13 17:51 UTC (permalink / raw)
To: acpi-devel
Hi,
I have posted this problem already to the acpi-support mailing list, but
even with the help of Gunter and Derek (thanks again), I cannot solve
it.
ACPI seems to work fine on my laptop, except the battery and ac status.
The related directories in /proc/acpi exist, but are empty.
There are no messages related to battery or ac in my dmesg.
Note that there is the same problem with apm : suspend mode works (apm
-s), but the battery status is unavailable.
Perhaps a related problem : I have a interupts conflict.
My dmesg follows.
I would be happy to assist with any information you need.
Frederic
lspci:
00:00.0 Host bridge: Intel Corp.: Unknown device 2560 (rev 01)
00:02.0 VGA compatible controller: Intel Corp.: Unknown device 2562 (rev
01)
00:1d.0 USB Controller: Intel Corp.: Unknown device 24c2 (rev 01)
00:1d.1 USB Controller: Intel Corp.: Unknown device 24c4 (rev 01)
00:1d.2 USB Controller: Intel Corp.: Unknown device 24c7 (rev 01)
00:1d.7 USB Controller: Intel Corp.: Unknown device 24cd (rev 01)
00:1e.0 PCI bridge: Intel Corp. 82801BA/CA PCI Bridge (rev 81)
00:1f.0 ISA bridge: Intel Corp.: Unknown device 24c0 (rev 01)
00:1f.1 IDE interface: Intel Corp. 82801DB ICH4 IDE (rev 01)
00:1f.3 SMBus: Intel Corp.: Unknown device 24c3 (rev 01)
00:1f.5 Multimedia audio controller: Intel Corp.: Unknown device 24c5
(rev 01)
00:1f.6 Modem: Intel Corp.: Unknown device 24c6 (rev 01)
02:03.0 CardBus bridge: Ricoh Co Ltd RL5c475 (rev 80)
02:08.0 Ethernet controller: Intel Corp.: Unknown device 1039 (rev 81)
02:0d.0 FireWire (IEEE 1394): Texas Instruments TSB43AB22 1394a-2000
Controller
DMESG :
Linux version 2.4.20-1mdk (quintela-JAOzawC1ADxn+L6o72McsUEOCMrvLtNR@public.gmane.org) (gcc version
3.2 (Mandrake Linux 9.1 3.2-4mdk)) #1 Tue Dec 3 21:24:49 CET 2002
BIOS-provided physical RAM map:
BIOS-e820: 0000000000000000 - 000000000009f800 (usable)
BIOS-e820: 000000000009f800 - 00000000000a0000 (reserved)
BIOS-e820: 00000000000ce000 - 00000000000d0000 (reserved)
BIOS-e820: 00000000000dc000 - 0000000000100000 (reserved)
BIOS-e820: 0000000000100000 - 000000000f6e0000 (usable)
BIOS-e820: 000000000f6e0000 - 000000000f6f0000 (ACPI data)
BIOS-e820: 000000000f6f0000 - 000000000f700000 (ACPI NVS)
BIOS-e820: 000000000f700000 - 000000000f780000 (usable)
BIOS-e820: 000000000f780000 - 0000000010000000 (reserved)
BIOS-e820: 00000000ff800000 - 00000000ffc00000 (reserved)
BIOS-e820: 00000000fffffc00 - 0000000100000000 (reserved)
247MB LOWMEM available.
ACPI: have wakeup address 0xc0001000
On node 0 totalpages: 63360
zone(0): 4096 pages.
zone(1): 59264 pages.
zone(2): 0 pages.
ACPI: RSDP (v000 PTLTD ) @ 0x000f68a0
ACPI: RSDT (v001 PTLTD RSDT 01540.00000) @ 0x0f6eb074
ACPI: FADT (v001 SEC DRACO 01540.00000) @ 0x0f6eff64
ACPI: BOOT (v001 PTLTD $SBFTBL$ 01540.00000) @ 0x0f6effd8
ACPI: DSDT (v001 SEC DRACO 01540.00000) @ 0x00000000
ACPI: BIOS passes blacklist
ACPI: MADT not present
Kernel command line: BOOT_IMAGE=2.420.-1mdk ro root=305 quiet devfs=mount hdc=ide-scsi
ide_setup: hdc=ide-scsi
No local APIC present or hardware disabled
Initializing CPU#0
Detected 1695.027 MHz processor.
Console: colour VGA+ 80x25
Calibrating delay loop... 3381.65 BogoMIPS
Memory: 247808k/253440k available (1302k kernel code, 5116k reserved, 491k data, 140k init, 0k highmem)
Dentry cache hash table entries: 32768 (order: 6, 262144 bytes)
Inode cache hash table entries: 16384 (order: 5, 131072 bytes)
Mount-cache hash table entries: 4096 (order: 3, 32768 bytes)
Buffer-cache hash table entries: 16384 (order: 4, 65536 bytes)
Page-cache hash table entries: 65536 (order: 6, 262144 bytes)
CPU: L1 I cache: 0K, L1 D cache: 8K
Intel machine check architecture supported.
Intel machine check reporting enabled on CPU#0.
CPU: After generic, caps: 3febf9ff 00000000 00000000 00000000
CPU: Common caps: 3febf9ff 00000000 00000000 00000000
CPU: Intel(R) Celeron(R) CPU 1.70GHz stepping 03
Enabling fast FPU save and restore... done.
Enabling unmasked SIMD FPU exception support... done.
Checking 'hlt' instruction... OK.
POSIX conformance testing by UNIFIX
mtrr: v1.40 (20010327) Richard Gooch (rgooch-r1x6VkxMR+00zabcByZE4g@public.gmane.org)
mtrr: detected mtrr type: Intel
ACPI: Subsystem revision 20021115
PCI: PCI BIOS revision 2.10 entry at 0xfd994, last bus=3
PCI: Using configuration type 1
ACPI-0508: *** Info: GPE Block0 defined as GPE0 to GPE31
ACPI: Interpreter enabled
ACPI: Using PIC for interrupt routing
ACPI: System [ACPI] (supports S0 S3 S4 S5)
ACPI: PCI Root Bridge [PCI0] (00:00)
PCI: Probing PCI hardware (bus 00)
Transparent bridge - Intel Corp. 82801BA/CA/DB PCI Bridge
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.PCIB._PRT]
ACPI: PCI Interrupt Link [LNKA] (IRQs *10)
ACPI: PCI Interrupt Link [LNKB] (IRQs *5)
ACPI: PCI Interrupt Link [LNKC] (IRQs *5)
ACPI: PCI Interrupt Link [LNKD] (IRQs *10)
ACPI: PCI Interrupt Link [LNKE] (IRQs *3)
ACPI: PCI Interrupt Link [LNKF] (IRQs 4 5 9 10 11, disabled)
ACPI: PCI Interrupt Link [LNKG] (IRQs 4 5 9 10 11, disabled)
ACPI: PCI Interrupt Link [LNKH] (IRQs *3)
ACPI: Embedded Controller [H_EC] (gpe 24)
ACPI: Power Resource [PFAN] (on)
PCI: Probing PCI hardware
ACPI: PCI Interrupt Link [LNKF] enabled at IRQ 11
ACPI: PCI Interrupt Link [LNKG] enabled at IRQ 9
PCI: Using ACPI for IRQ routing
PCI: if you experience problems, try using option 'pci=noacpi' or even 'acpi=off'
isapnp: Scanning for PnP cards...
isapnp: No Plug & Play device found
Linux NET4.0 for Linux 2.4
Based upon Swansea University Computer Society NET3.039
Initializing RT netlink socket
apm: BIOS version 1.2 Flags 0x03 (Driver version 1.16)
apm: overridden by ACPI.
Starting kswapd
VFS: Disk quotas vdquot_6.5.1
devfs: v1.12c (20020818) Richard Gooch (rgooch-r1x6VkxMR+00zabcByZE4g@public.gmane.org)
devfs: boot_options: 0x1
pty: 256 Unix98 ptys configured
Serial driver version 5.05c (2001-07-08) with HUB-6 MANY_PORTS MULTIPORT SHARE_IRQ SERIAL_PCI ISAPNP enabled
Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
ICH4: IDE controller on PCI bus 00 dev f9
PCI: Device 00:1f.1 not available because of resource collisions
ICH4: BIOS setup was incomplete.
ICH4: chipset revision 1
ICH4: not 100% native mode: will probe irqs later
ide0: BM-DMA at 0x1860-0x1867, BIOS settings: hda:DMA, hdb:pio
ide1: BM-DMA at 0x1868-0x186f, BIOS settings: hdc:DMA, hdd:pio
hda: IC25N020ATCS04-0, ATA DISK drive
hdc: Samsung CD-RW/DVD-ROM SN-324B, ATAPI CD/DVD-ROM drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
blk: queue c03172ec, I/O limit 4095Mb (mask 0xffffffff)
hda: 39070080 sectors (20004 MB) w/1768KiB Cache, CHS=2432/255/63, UDMA(100)
Partition check:
/dev/ide/host0/bus0/target0/lun0: p1 p2 < p5 p6 p7 >
RAMDISK driver initialized: 16 RAM disks of 32000K size 1024 blocksize
md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27
md: Autodetecting RAID arrays.
md: autorun ...
md: ... autorun DONE.
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP, IGMP
IP: routing cache hash table of 2048 buckets, 16Kbytes
TCP: Hash tables configured (established 16384 bind 32768)
Linux IP multicast router 0.06 plus PIM-SM
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
RAMDISK: Compressed image found at block 0
Freeing initrd memory: 113k freed
VFS: Mounted root (ext2 filesystem).
Mounted devfs on /dev
Journalled Block Device driver loaded
kjournald starting. Commit interval 5 seconds
EXT3-fs: mounted filesystem with ordered data mode.
Mounted devfs on /dev
Freeing unused kernel memory: 140k freed
Real Time Clock Driver v1.10e
usb.c: registered new driver usbdevfs
usb.c: registered new driver hub
usb-uhci.c: $Revision: 1.275 $ time 22:00:44 Dec 3 2002
usb-uhci.c: High bandwidth mode enabled
PCI: Setting latency timer of device 00:1d.0 to 64
usb-uhci.c: USB UHCI at I/O 0x1800, IRQ 10
usb-uhci.c: Detected 2 ports
usb.c: new USB bus registered, assigned bus number 1
hub.c: USB hub found
hub.c: 2 ports detected
PCI: Setting latency timer of device 00:1d.1 to 64
usb-uhci.c: USB UHCI at I/O 0x1820, IRQ 10
usb-uhci.c: Detected 2 ports
usb.c: new USB bus registered, assigned bus number 2
hub.c: USB hub found
hub.c: 2 ports detected
PCI: Setting latency timer of device 00:1d.2 to 64
usb-uhci.c: USB UHCI at I/O 0x1840, IRQ 5
usb-uhci.c: Detected 2 ports
usb.c: new USB bus registered, assigned bus number 3
hub.c: USB hub found
hub.c: 2 ports detected
usb-uhci.c: v1.275:USB Universal Host Controller Interface driver
PCI: Setting latency timer of device 00:1d.7 to 64
hcd.c: ehci-hcd @ 00:1d.7, Intel Corp. 82801DB USB EHCI Controller
hcd.c: irq 3, pci mem d007b000
usb.c: new USB bus registered, assigned bus number 4
ehci-hcd.c: restricting 64bit DMA mappings to segment 0 ...
ehci-hcd.c: USB 2.0 support enabled, EHCI rev 1. 0
hub.c: USB hub found
hub.c: 6 ports detected
usbdevfs: remount parameter error
EXT3 FS 2.4-0.9.19, 19 August 2002 on ide0(3,5), internal journal
Adding Swap: 522040k swap-space (priority -1)
SCSI subsystem driver Revision: 1.00
scsi0 : SCSI host adapter emulation for IDE ATAPI devices
Vendor: SAMSUNG Model: CDRW/DVD SN-324B Rev: U100
Type: CD-ROM ANSI SCSI revision: 02
ACPI-0323: *** Error: Handler for [Embedded_control] returned AE_BAD_PARAMETER
ACPI-1103: *** Error: Method execution failed, AE_AML_NO_RETURN_VALUE
ACPI-0323: *** Error: Handler for [Embedded_control] returned AE_BAD_PARAMETER
ACPI-1103: *** Error: Method execution failed, AE_BAD_PARAMETER
ACPI: Power Button (FF) [PWRF]
ACPI: Sleep Button (CM) [SLPB]
ACPI: Processor [CPU0] (supports C1 C2)
ACPI: Fan [FAN0] (on)
ACPI-0323: *** Error: Handler for [Embedded_control] returned AE_BAD_PARAMETER
ACPI-1103: *** Error: Method execution failed, AE_BAD_PARAMETER
kjournald starting. Commit interval 5 seconds
EXT3 FS 2.4-0.9.19, 19 August 2002 on ide0(3,7), internal journal
EXT3-fs: mounted filesystem with ordered data mode.
MSDOS FS: IO charset iso8859-15
MSDOS FS: Using codepage 850
ohci1394: $Rev: 578 $ Ben Collins <bcollins-8fiUuRrzOP0dnm+yROfE0A@public.gmane.org>
ohci1394_0: OHCI-1394 1.1 (PCI): IRQ=[10] MMIO=[d0105000-d01057ff] Max Packet=[2048]
ieee1394: Host added: Node[00:1023] GUID[0000f04120001f68] [Linux OHCI-1394]
hdc: DMA disabled
hda: task_no_data_intr: status=0x51 { DriveReady SeekComplete Error }
hda: task_no_data_intr: error=0x04 { DriveStatusError }
eepro100.c:v1.09j-t 9/29/99 Donald Becker http://www.scyld.com/network/eepro100.html
eepro100.c: $Revision: 1.36 $ 2000/11/17 Modified by Andrey V. Savochkin <saw-5bpFXmC1L3aJ4eUNlOKu3Q@public.gmane.org> and others
eth0: OEM i82557/i82558 10/100 Ethernet, 00:00:F0:6E:31:1A, IRQ 3.
Board assembly 000000-000, Physical connectors present: RJ45
Primary interface chip i82555 PHY #1.
General self-test: passed.
Serial sub-system self-test: passed.
Internal registers self-test: passed.
ROM checksum self-test: passed (0x04f4518b).
-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/
^ permalink raw reply
* natting specific ports
From: Simpson, Doug @ 2002-12-13 17:48 UTC (permalink / raw)
To: 'netfilter@lists.netfilter.org'
I want to "NAT" just specific ports to my Public IP. Do the commands below
make sense? I want my internal network to be able to telnet and send email.
(eth0 is my External NIC - it is exposed to the internet)
iptables -t nat -A POSTROUTING -p tcp --sport 25 -o eth0 -s $INTERNAL_IP -j
SNAT --to $EXTERNAL_IP
iptables -t nat -A POSTROUTING -p tcp --sport 23 -o eth0 -s $INTERNAL_IP -j
SNAT --to $EXTERNAL_IP
Thank you,
Doug
^ permalink raw reply
* RE: Does IPTables have a 1:1 port-forwarding capability for a DNAT port-range ?
From: Ranjeet Shetye @ 2002-12-13 17:45 UTC (permalink / raw)
To: netfilter
In-Reply-To: <3DF9CC70.3080603@istitutocolli.org>
Hi Andrea,
I was hoping that the answer would be "Yes, IPTables can do it", but I
think you are right. As a matter of fact, I have been looking into the
NetFilter code for the last couple of days and I was hoping that I
didn't need to hack the kernel code (cos its not documented). But it
looks like I will have to.
Anyways, I will take my query to the developer-mailing list and post a
reply here once everything is sorted out. In the meantime if anyone has
an architectural overview of Netfilter, one that maps out the
interactions between the various components and layers, I'd really
really like to get my hands on it!
Thanks for your help, Andrea.
Ranjeet Shetye
Senior Software Engineer
Zultys Technologies
771 Vaqueros Avenue
Sunnyvale CA 94085
USA
Ranjeet.Shetye@Zultys.com
http://www.zultys.com/
> -----Original Message-----
> From: netfilter-admin@lists.netfilter.org
> [mailto:netfilter-admin@lists.netfilter.org] On Behalf Of
> Andrea Rossato
> Sent: Friday, December 13, 2002 4:03 AM
> To: netfilter@lists.netfilter.org
> Subject: Re: Does IPTables have a 1:1 port-forwarding
> capability for a DNAT port-range ?
>
>
> Ranjeet Shetye wrote:
> > The reason for wanting a 1:1 rule is for X windows and
> other fat port
> > ranges. Dont want hundreds of rules in there if one can do the job.
> > Can IPTables do it ? If so how ? If not, I guess I'll have
> to get in
> > touch with the developers for tips on a good starting point.
>
> I believe that the only way is to hack nat code.
> I will start looking in
> net/ipv4/netfilter/ip_nat_core.c
> and the function manip_pkt that, as far as I understand, is
> actually writing the NATed packet andrea
>
>
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.