c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2026-05-22 02:27:36 +03:00
Code Issues Packages Projects Releases Wiki Activity

Switch KM_SLEEP to KM_PUSHPAGE

Browse Source 
Differences between how paging is done on Solaris and Linux can cause
deadlocks if KM_SLEEP is used in any the following contexts.

  * The txg_sync thread
  * The zvol write/discard threads
  * The zpl_putpage() VFS callback

This is because KM_SLEEP will allow for direct reclaim which may result
in the VM calling back in to the filesystem or block layer to write out
pages.  If a lock is held over this operation the potential exists to
deadlock the system.  To ensure forward progress all memory allocations
in these contexts must us KM_PUSHPAGE which disables performing any I/O
to accomplish the memory allocation.

Previously, this behavior was acheived by setting PF_MEMALLOC on the
thread.  However, that resulted in unexpected side effects such as the
exhaustion of pages in ZONE_DMA.  This approach touchs more of the zfs
code, but it is more consistent with the right way to handle these cases
under Linux.

This is patch lays the ground work for being able to safely revert the
following commits which used PF_MEMALLOC:

  21ade34 Disable direct reclaim for z_wr_* threads
  cfc9a5c Fix zpl_writepage() deadlock
  eec8164 Fix ASSERTION(!dsl_pool_sync_context(tx->tx_pool))

Signed-off-by: Richard Yao <ryao@cs.stonybrook.edu>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Issue #726
This commit is contained in:
Richard Yao
2012-05-07 13:49:51 -04:00
committed by Brian Behlendorf
parent 991fc1d7ae
commit b8d06fca08
45 changed files with 187 additions and 185 deletions
Download Patch File Download Diff File Expand all files Collapse all files
module/zfs/sa.c
+13 -13
View File
@@ -433,10 +433,10 @@ sa_add_layout_entry(objset_t *os, sa_attr_type_t *attrs, int attr_count,
avl_index_t loc;
ASSERT(MUTEX_HELD(&sa->sa_lock));
tb = kmem_zalloc(sizeof (sa_lot_t), KM_SLEEP);
tb = kmem_zalloc(sizeof (sa_lot_t), KM_PUSHPAGE);
tb->lot_attr_count = attr_count;
tb->lot_attrs = kmem_alloc(sizeof (sa_attr_type_t) * attr_count,
KM_SLEEP);
KM_PUSHPAGE);
bcopy(attrs, tb->lot_attrs, sizeof (sa_attr_type_t) * attr_count);
tb->lot_num = lot_num;
tb->lot_hash = hash;
@@ -721,7 +721,7 @@ sa_build_layouts(sa_handle_t *hdl, sa_bulk_attr_t *attr_desc, int attr_count,
buf_space = hdl->sa_bonus->db_size - hdrsize;
attrs_start = attrs = kmem_alloc(sizeof (sa_attr_type_t) * attr_count,
KM_SLEEP);
KM_PUSHPAGE);
lot_count = 0;
for (i = 0, len_idx = 0, hash = -1ULL; i != attr_count; i++) {
@@ -842,7 +842,7 @@ sa_attr_table_setup(objset_t *os, sa_attr_reg_t *reg_attrs, int count)
dmu_objset_type_t ostype = dmu_objset_type(os);
sa->sa_user_table =
kmem_zalloc(count * sizeof (sa_attr_type_t), KM_SLEEP);
kmem_zalloc(count * sizeof (sa_attr_type_t), KM_PUSHPAGE);
sa->sa_user_table_sz = count * sizeof (sa_attr_type_t);
if (sa->sa_reg_attr_obj != 0) {
@@ -901,7 +901,7 @@ sa_attr_table_setup(objset_t *os, sa_attr_reg_t *reg_attrs, int count)
sa->sa_num_attrs = sa_attr_count;
tb = sa->sa_attr_table =
kmem_zalloc(sizeof (sa_attr_table_t) * sa_attr_count, KM_SLEEP);
kmem_zalloc(sizeof (sa_attr_table_t) * sa_attr_count, KM_PUSHPAGE);
/*
* Attribute table is constructed from requested attribute list,
@@ -926,7 +926,7 @@ sa_attr_table_setup(objset_t *os, sa_attr_reg_t *reg_attrs, int count)
continue;
}
tb[ATTR_NUM(value)].sa_name =
kmem_zalloc(strlen(za.za_name) +1, KM_SLEEP);
kmem_zalloc(strlen(za.za_name) +1, KM_PUSHPAGE);
(void) strlcpy(tb[ATTR_NUM(value)].sa_name, za.za_name,
strlen(za.za_name) +1);
}
@@ -952,7 +952,7 @@ sa_attr_table_setup(objset_t *os, sa_attr_reg_t *reg_attrs, int count)
tb[i].sa_registered = B_FALSE;
tb[i].sa_name =
kmem_zalloc(strlen(sa_legacy_attrs[i].sa_name) +1,
KM_SLEEP);
KM_PUSHPAGE);
(void) strlcpy(tb[i].sa_name,
sa_legacy_attrs[i].sa_name,
strlen(sa_legacy_attrs[i].sa_name) + 1);
@@ -970,7 +970,7 @@ sa_attr_table_setup(objset_t *os, sa_attr_reg_t *reg_attrs, int count)
tb[attr_id].sa_byteswap = reg_attrs[i].sa_byteswap;
tb[attr_id].sa_attr = attr_id;
tb[attr_id].sa_name =
kmem_zalloc(strlen(reg_attrs[i].sa_name) + 1, KM_SLEEP);
kmem_zalloc(strlen(reg_attrs[i].sa_name) + 1, KM_PUSHPAGE);
(void) strlcpy(tb[attr_id].sa_name, reg_attrs[i].sa_name,
strlen(reg_attrs[i].sa_name) + 1);
}
@@ -1007,7 +1007,7 @@ sa_setup(objset_t *os, uint64_t sa_obj, sa_attr_reg_t *reg_attrs, int count,
return (0);
}
sa = kmem_zalloc(sizeof (sa_os_t), KM_SLEEP);
sa = kmem_zalloc(sizeof (sa_os_t), KM_PUSHPAGE);
mutex_init(&sa->sa_lock, NULL, MUTEX_DEFAULT, NULL);
sa->sa_master_obj = sa_obj;
@@ -1055,7 +1055,7 @@ sa_setup(objset_t *os, uint64_t sa_obj, sa_attr_reg_t *reg_attrs, int count,
uint64_t lot_num;
lot_attrs = kmem_zalloc(sizeof (sa_attr_type_t) *
za.za_num_integers, KM_SLEEP);
za.za_num_integers, KM_PUSHPAGE);
if ((error = (zap_lookup(os, sa->sa_layout_attr_obj,
za.za_name, 2, za.za_num_integers,
@@ -1540,14 +1540,14 @@ sa_find_idx_tab(objset_t *os, dmu_object_type_t bonustype, void *data)
}
/* No such luck, create a new entry */
idx_tab = kmem_zalloc(sizeof (sa_idx_tab_t), KM_SLEEP);
idx_tab = kmem_zalloc(sizeof (sa_idx_tab_t), KM_PUSHPAGE);
idx_tab->sa_idx_tab =
kmem_zalloc(sizeof (uint32_t) * sa->sa_num_attrs, KM_SLEEP);
kmem_zalloc(sizeof (uint32_t) * sa->sa_num_attrs, KM_PUSHPAGE);
idx_tab->sa_layout = tb;
refcount_create(&idx_tab->sa_refcount);
if (tb->lot_var_sizes)
idx_tab->sa_variable_lengths = kmem_alloc(sizeof (uint16_t) *
tb->lot_var_sizes, KM_SLEEP);
tb->lot_var_sizes, KM_PUSHPAGE);
sa_attr_iter(os, hdr, bonustype, sa_build_idx_tab,
tb, idx_tab);