Details
-
Bug
-
Resolution: Fixed
-
Major
-
Lustre 2.8.0
-
None
-
3
-
17696
Description
I am unable to build v2_7_50_0-10-g56875fd (LU-3680 osd: reduce osd_thread_info in ldiskfs osd, http://review.whamcloud.com/9726) due to the added CLASSERT() in osd_mod_init() and the that fact that I enable several CONFIG_DEBUG_*.
struct osd_thread_info {
const struct lu_env * oti_env; /* 0 8 */
struct dentry oti_obj_dentry; /* 8 256 */
/* --- cacheline 4 boundary (256 bytes) was 8 bytes ago --- */
struct dentry oti_child_dentry; /* 264 256 */
/* --- cacheline 8 boundary (512 bytes) was 8 bytes ago --- */
struct dentry oti_it_dentry; /* 520 256 */
/* --- cacheline 12 boundary (768 bytes) was 8 bytes ago --- */
union {
struct file oti_file; /* 304 */
struct kstatfs oti_ksfs; /* 120 */
}; /* 776 304 */
/* --- cacheline 16 boundary (1024 bytes) was 56 bytes ago --- */
struct htree_lock * oti_hlock; /* 1080 8 */
/* --- cacheline 17 boundary (1088 bytes) --- */
struct lu_fid oti_fid; /* 1088 16 */
struct lu_fid oti_fid2; /* 1104 16 */
struct lu_fid oti_fid3; /* 1120 16 */
struct osd_inode_id oti_id; /* 1136 8 */
struct osd_inode_id oti_id2; /* 1144 8 */
/* --- cacheline 18 boundary (1152 bytes) --- */
struct osd_inode_id oti_id3; /* 1152 8 */
struct ost_id oti_ostid; /* 1160 16 */
struct timespec oti_time; /* 1176 16 */
struct lustre_capa_key oti_capa_key; /* 1192 72 */
/* --- cacheline 19 boundary (1216 bytes) was 48 bytes ago --- */
struct lustre_capa oti_capa; /* 1264 120 */
/* --- cacheline 21 boundary (1344 bytes) was 40 bytes ago --- */
struct osd_device * oti_dev; /* 1384 8 */
void * oti_it_ea_buf; /* 1392 8 */
unsigned int oti_it_ea_buf_used:1; /* 1400:31 4 */
/* XXX 31 bits hole, try to pack */
/* XXX 4 bytes hole, try to pack */
/* --- cacheline 22 boundary (1408 bytes) --- */
struct iam_iterator oti_idx_it; /* 1408 312 */
/* --- cacheline 26 boundary (1664 bytes) was 56 bytes ago --- */
union {
char oti_name[48]; /* 48 */
char oti_it_ipd[816]; /* 816 */
long long int oti_alignment_lieutenant; /* 8 */
}; /* 1720 816 */
/* --- cacheline 39 boundary (2496 bytes) was 40 bytes ago --- */
union {
char oti_idx_ipd[816]; /* 816 */
long long int oti_alignment_lieutenant_colonel; /* 8 */
}; /* 2536 816 */
/* --- cacheline 52 boundary (3328 bytes) was 24 bytes ago --- */
struct osd_idmap_cache oti_cache; /* 3352 32 */
int oti_r_locks; /* 3384 4 */
int oti_w_locks; /* 3388 4 */
/* --- cacheline 53 boundary (3392 bytes) --- */
int oti_txns; /* 3392 4 */
/* XXX 4 bytes hole, try to pack */
struct lu_buf oti_buf; /* 3400 16 */
struct lu_buf oti_big_buf; /* 3416 16 */
union {
struct lustre_mdt_attrs oti_mdt_attrs; /* 24 */
char oti_mdt_attrs_old[64]; /* 64 */
struct filter_fid_old oti_ff; /* 32 */
struct filter_fid oti_ff_new; /* 16 */
}; /* 3432 64 */
/* --- cacheline 54 boundary (3456 bytes) was 40 bytes ago --- */
struct osd_iobuf oti_iobuf; /* 3496 184 */
/* XXX last struct has 4 bytes of padding */
/* --- cacheline 57 boundary (3648 bytes) was 32 bytes ago --- */
struct inode * oti_inode; /* 3680 8 */
char oti_ldp[32]; /* 3688 32 */
/* --- cacheline 58 boundary (3712 bytes) was 8 bytes ago --- */
char oti_ldp2[32]; /* 3720 32 */
union {
struct if_dqblk oti_dqblk; /* 72 */
struct if_dqinfo oti_dqinfo; /* 24 */
}; /* 3752 72 */
/* --- cacheline 59 boundary (3776 bytes) was 48 bytes ago --- */
struct lquota_id_info oti_qi; /* 3824 48 */
/* XXX last struct has 7 bytes of padding */
/* --- cacheline 60 boundary (3840 bytes) was 32 bytes ago --- */
struct lquota_trans oti_quota_trans; /* 3872 200 */
/* --- cacheline 63 boundary (4032 bytes) was 40 bytes ago --- */
union lquota_rec oti_quota_rec; /* 4072 32 */
/* --- cacheline 64 boundary (4096 bytes) was 8 bytes ago --- */
__u64 oti_quota_id; /* 4104 8 */
struct lu_seq_range oti_seq_range; /* 4112 24 */
short unsigned int oti_declare_ops[11]; /* 4136 22 */
short unsigned int oti_declare_ops_rb[11]; /* 4158 22 */
/* --- cacheline 65 boundary (4160 bytes) was 20 bytes ago --- */
short unsigned int oti_declare_ops_cred[11]; /* 4180 22 */
bool oti_rollback; /* 4202 1 */
/* size: 4208, cachelines: 66, members: 43 */
/* sum members: 4195, holes: 2, sum holes: 8 */
/* bit holes: 1, sum bit holes: 31 bits */
/* padding: 5 */
/* paddings: 2, sum paddings: 11 */
/* last cacheline: 48 bytes */
};
My config is attached.
Attachments
Issue Links
- is related to
-
-
- Resolved
-
Activity
Oleg, can you post the output of 'pahole lustre/osd-ldiskfs/osd_handler.o'?
that's because you happen to have those two set too.
I guess they are a good proxy for "debug enabled kernel" in the end but it might be interesting to see what is the real reason.
here's a relevant diff between my (orig) and your configs:
# # Kernel hacking @@ -4111,7 +4227,7 @@ # CONFIG_UNUSED_SYMBOLS is not set CONFIG_DEBUG_FS=y CONFIG_HEADERS_CHECK=y -# CONFIG_DEBUG_SECTION_MISMATCH is not set +CONFIG_DEBUG_SECTION_MISMATCH=y CONFIG_DEBUG_KERNEL=y CONFIG_DEBUG_SHIRQ=y CONFIG_DETECT_SOFTLOCKUP=y @@ -4119,8 +4235,8 @@ CONFIG_HARDLOCKUP_DETECTOR=y CONFIG_BOOTPARAM_HARDLOCKUP_ENABLED=y CONFIG_BOOTPARAM_HARDLOCKUP_ENABLED_VALUE=1 -# CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is not set -CONFIG_BOOTPARAM_HARDLOCKUP_PANIC_VALUE=0 +CONFIG_BOOTPARAM_HARDLOCKUP_PANIC=y +CONFIG_BOOTPARAM_HARDLOCKUP_PANIC_VALUE=1 # CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC is not set CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC_VALUE=0 CONFIG_DETECT_HUNG_TASK=y @@ -4132,33 +4248,39 @@ CONFIG_TIMER_STATS=y CONFIG_DEBUG_OBJECTS=y # CONFIG_DEBUG_OBJECTS_SELFTEST is not set -CONFIG_DEBUG_OBJECTS_FREE=y +# CONFIG_DEBUG_OBJECTS_FREE is not set # CONFIG_DEBUG_OBJECTS_TIMERS is not set CONFIG_DEBUG_OBJECTS_ENABLE_DEFAULT=1 CONFIG_DEBUG_SLAB=y # CONFIG_DEBUG_SLAB_LEAK is not set -# CONFIG_DEBUG_KMEMLEAK is not set -# CONFIG_DEBUG_RT_MUTEXES is not set +CONFIG_DEBUG_KMEMLEAK=y +CONFIG_DEBUG_KMEMLEAK_EARLY_LOG_SIZE=400 +# CONFIG_DEBUG_KMEMLEAK_TEST is not set +# CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF is not set +CONFIG_DEBUG_RT_MUTEXES=y +CONFIG_DEBUG_PI_LIST=y # CONFIG_RT_MUTEX_TESTER is not set CONFIG_DEBUG_SPINLOCK=y CONFIG_DEBUG_MUTEXES=y -# CONFIG_DEBUG_WW_MUTEX_SLOWPATH is not set -# CONFIG_DEBUG_LOCK_ALLOC is not set -# CONFIG_PROVE_LOCKING is not set -# CONFIG_LOCK_STAT is not set +CONFIG_DEBUG_LOCK_ALLOC=y +CONFIG_PROVE_LOCKING=y +CONFIG_LOCKDEP=y +CONFIG_LOCK_STAT=y +# CONFIG_DEBUG_LOCKDEP is not set +CONFIG_TRACE_IRQFLAGS=y CONFIG_DEBUG_SPINLOCK_SLEEP=y # CONFIG_DEBUG_LOCKING_API_SELFTESTS is not set CONFIG_STACKTRACE=y # CONFIG_DEBUG_KOBJECT is not set CONFIG_DEBUG_BUGVERBOSE=y CONFIG_DEBUG_INFO=y -# CONFIG_DEBUG_VM is not set -# CONFIG_DEBUG_VIRTUAL is not set -CONFIG_DEBUG_WRITECOUNT=y +CONFIG_DEBUG_VM=y +CONFIG_DEBUG_VIRTUAL=y +# CONFIG_DEBUG_WRITECOUNT is not set CONFIG_DEBUG_MEMORY_INIT=y CONFIG_DEBUG_LIST=y # CONFIG_DEBUG_SG is not set -CONFIG_DEBUG_NOTIFIERS=y +# CONFIG_DEBUG_NOTIFIERS is not set # CONFIG_DEBUG_CREDENTIALS is not set CONFIG_ARCH_WANT_FRAME_POINTERS=y CONFIG_FRAME_POINTER=y @@ -4170,15 +4292,10 @@ # CONFIG_DEBUG_BLOCK_EXT_DEVT is not set # CONFIG_DEBUG_FORCE_WEAK_PER_CPU is not set # CONFIG_LKDTM is not set -CONFIG_FAULT_INJECTION=y -CONFIG_FAILSLAB=y -CONFIG_FAIL_PAGE_ALLOC=y -CONFIG_FAIL_MAKE_REQUEST=y -# CONFIG_FAIL_IO_TIMEOUT is not set -# CONFIG_FAULT_INJECTION_DEBUG_FS is not set -CONFIG_LATENCYTOP=y -CONFIG_SYSCTL_SYSCALL_CHECK=y -CONFIG_DEBUG_PAGEALLOC=y +# CONFIG_FAULT_INJECTION is not set +# CONFIG_LATENCYTOP is not set +# CONFIG_SYSCTL_SYSCALL_CHECK is not set +# CONFIG_DEBUG_PAGEALLOC is not set CONFIG_USER_STACKTRACE_SUPPORT=y CONFIG_NOP_TRACER=y CONFIG_HAVE_FTRACE_NMI_ENTER=y @@ -4221,12 +4338,12 @@ # CONFIG_FTRACE_STARTUP_TEST is not set # CONFIG_MMIOTRACE is not set CONFIG_RING_BUFFER_BENCHMARK=m -# CONFIG_PROVIDE_OHCI1394_DMA_INIT is not set +CONFIG_PROVIDE_OHCI1394_DMA_INIT=y # CONFIG_FIREWIRE_OHCI_REMOTE_DMA is not set -# CONFIG_BUILD_DOCSRC is not set +CONFIG_BUILD_DOCSRC=y CONFIG_DYNAMIC_DEBUG=y # CONFIG_DMA_API_DEBUG is not set -# CONFIG_ASYNC_RAID6_TEST is not set +CONFIG_ASYNC_RAID6_TEST=m # CONFIG_SAMPLES is not set CONFIG_HAVE_ARCH_KGDB=y CONFIG_KGDB=y @@ -4235,18 +4352,17 @@ # CONFIG_KGDB_TESTS_ON_BOOT is not set CONFIG_HAVE_ARCH_KMEMCHECK=y # CONFIG_TEST_KSTRTOX is not set -# CONFIG_STRICT_DEVMEM is not set +CONFIG_STRICT_DEVMEM=y # CONFIG_X86_VERBOSE_BOOTUP is not set CONFIG_EARLY_PRINTK=y -# CONFIG_EARLY_PRINTK_DBGP is not set -# CONFIG_EARLY_PRINTK_EFI is not set +CONFIG_EARLY_PRINTK_DBGP=y CONFIG_DEBUG_STACKOVERFLOW=y -CONFIG_DEBUG_STACK_USAGE=y +# CONFIG_DEBUG_STACK_USAGE is not set # CONFIG_DEBUG_PER_CPU_MAPS is not set # CONFIG_X86_PTDUMP is not set CONFIG_DEBUG_RODATA=y -# CONFIG_DEBUG_RODATA_TEST is not set -# CONFIG_DEBUG_NX_TEST is not set +CONFIG_DEBUG_RODATA_TEST=y +CONFIG_DEBUG_NX_TEST=m # CONFIG_IOMMU_DEBUG is not set # CONFIG_IOMMU_STRESS is not set CONFIG_HAVE_MMIOTRACE_SUPPORT=y
This is not the real reason for blown up structure sizes.
I have all of those set and my struct size is still about 3500
[green@intelbox lustre-release]$ grep CONFIG_DEBUG_MUTEXES ~/bk/linux-2.6.32-504.3.3.el6-debug/.config CONFIG_DEBUG_MUTEXES=y [green@intelbox lustre-release]$ grep CONFIG_DEBUG_SPINLOCK ~/bk/linux-2.6.32-504.3.3.el6-debug/.config CONFIG_DEBUG_SPINLOCK=y CONFIG_DEBUG_SPINLOCK_SLEEP=y Reading symbols from /home/green/git/lustre-release/lustre/osd-ldiskfs/osd_ldiskfs.ko...done. (gdb) p sizeof(struct osd_thread_info) $1 = 3576
The culprit must be something else
John L. Hammond (john.hammond@intel.com) uploaded a new patch: http://review.whamcloud.com/13955
Subject: LU-6324 osd: allow larger osd_thread_info for debug
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 9d2a721fa5289f788f9c3c718752abe5fb1d8517
John, would you mind to put CLASSERT() under #ifndef CONFIG_DEBUG_MUTEXES and #ifndef CONFIG_DEBUG_SPINLOCK please ?
Andreas, that's possible with http://review.whamcloud.com/#/c/5572/ (capa support removal):
(gdb) p sizeof(struct lustre_capa_key)
$1 = 72
(gdb) p sizeof(struct lustre_capa)
$2 = 120
Oleg Drokin (oleg.drokin@intel.com) merged in patch http://review.whamcloud.com/13955/
Subject:
LU-6324osd: allow larger osd_thread_info for debugProject: fs/lustre-release
Branch: master
Current Patch Set:
Commit: c65d83dd93fc61c44a1f0503e196a123d57dfb88