1 Transactional Memory support
2 ============================
4 POWER kernel support for this feature is currently limited to supporting
5 its use by user programs. It is not currently used by the kernel itself.
7 This file aims to sum up how it is supported by Linux and what behaviour you
8 can expect from your user programs.
14 Hardware Transactional Memory is supported on POWER8 processors, and is a
15 feature that enables a different form of atomic memory access. Several new
16 instructions are presented to delimit transactions; transactions are
17 guaranteed to either complete atomically or roll back and undo any partial
20 A simple transaction looks like this:
26 ld r4, SAVINGS_ACCT(r3)
27 ld r5, CURRENT_ACCT(r3)
30 std r4, SAVINGS_ACCT(r3)
31 std r5, CURRENT_ACCT(r3)
38 ... test for odd failures ...
40 /* Retry the transaction if it failed because it conflicted with
45 The 'tbegin' instruction denotes the start point, and 'tend' the end point.
46 Between these points the processor is in 'Transactional' state; any memory
47 references will complete in one go if there are no conflicts with other
48 transactional or non-transactional accesses within the system. In this
49 example, the transaction completes as though it were normal straight-line code
50 IF no other processor has touched SAVINGS_ACCT(r3) or CURRENT_ACCT(r3); an
51 atomic move of money from the current account to the savings account has been
52 performed. Even though the normal ld/std instructions are used (note no
53 lwarx/stwcx), either *both* SAVINGS_ACCT(r3) and CURRENT_ACCT(r3) will be
54 updated, or neither will be updated.
56 If, in the meantime, there is a conflict with the locations accessed by the
57 transaction, the transaction will be aborted by the CPU. Register and memory
58 state will roll back to that at the 'tbegin', and control will continue from
59 'tbegin+4'. The branch to abort_handler will be taken this second time; the
60 abort handler can check the cause of the failure, and retry.
62 Checkpointed registers include all GPRs, FPRs, VRs/VSRs, LR, CCR/CR, CTR, FPCSR
63 and a few other status/flag regs; see the ISA for details.
65 Causes of transaction aborts
66 ============================
68 - Conflicts with cache lines used by other processors
71 - See the ISA for full documentation of everything that will abort transactions.
77 Performing syscalls from within transaction is not recommended, and can lead
78 to unpredictable results.
80 Syscalls do not by design abort transactions, but beware: The kernel code will
81 not be running in transactional state. The effect of syscalls will always
82 remain visible, but depending on the call they may abort your transaction as a
83 side-effect, read soon-to-be-aborted transactional data that should not remain
84 invisible, etc. If you constantly retry a transaction that constantly aborts
85 itself by calling a syscall, you'll have a livelock & make no progress.
87 Simple syscalls (e.g. sigprocmask()) "could" be OK. Even things like write()
88 from, say, printf() should be OK as long as the kernel does not access any
89 memory that was accessed transactionally.
91 Consider any syscalls that happen to work as debug-only -- not recommended for
92 production use. Best to queue them up till after the transaction is over.
98 Delivery of signals (both sync and async) during transactions provides a second
99 thread state (ucontext/mcontext) to represent the second transactional register
100 state. Signal delivery 'treclaim's to capture both register states, so signals
101 abort transactions. The usual ucontext_t passed to the signal handler
102 represents the checkpointed/original register state; the signal appears to have
103 arisen at 'tbegin+4'.
105 If the sighandler ucontext has uc_link set, a second ucontext has been
106 delivered. For future compatibility the MSR.TS field should be checked to
107 determine the transactional state -- if so, the second ucontext in uc->uc_link
108 represents the active transactional registers at the point of the signal.
110 For 64-bit processes, uc->uc_mcontext.regs->msr is a full 64-bit MSR and its TS
111 field shows the transactional mode.
113 For 32-bit processes, the mcontext's MSR register is only 32 bits; the top 32
114 bits are stored in the MSR of the second ucontext, i.e. in
115 uc->uc_link->uc_mcontext.regs->msr. The top word contains the transactional
118 However, basic signal handlers don't need to be aware of transactions
119 and simply returning from the handler will deal with things correctly:
121 Transaction-aware signal handlers can read the transactional register state
122 from the second ucontext. This will be necessary for crash handlers to
123 determine, for example, the address of the instruction causing the SIGSEGV.
125 Example signal handler:
127 void crash_handler(int sig, siginfo_t *si, void *uc)
129 ucontext_t *ucp = uc;
130 ucontext_t *transactional_ucp = ucp->uc_link;
133 u64 msr = ucp->uc_mcontext.regs->msr;
134 /* May have transactional ucontext! */
135 #ifndef __powerpc64__
136 msr |= ((u64)transactional_ucp->uc_mcontext.regs->msr) << 32;
138 if (MSR_TM_ACTIVE(msr)) {
139 /* Yes, we crashed during a transaction. Oops. */
140 fprintf(stderr, "Transaction to be restarted at 0x%llx, but "
141 "crashy instruction was at 0x%llx\n",
142 ucp->uc_mcontext.regs->nip,
143 transactional_ucp->uc_mcontext.regs->nip);
147 fix_the_problem(ucp->dar);
150 When in an active transaction that takes a signal, we need to be careful with
151 the stack. It's possible that the stack has moved back up after the tbegin.
152 The obvious case here is when the tbegin is called inside a function that
153 returns before a tend. In this case, the stack is part of the checkpointed
154 transactional memory state. If we write over this non transactionally or in
155 suspend, we are in trouble because if we get a tm abort, the program counter and
156 stack pointer will be back at the tbegin but our in memory stack won't be valid
159 To avoid this, when taking a signal in an active transaction, we need to use
160 the stack pointer from the checkpointed state, rather than the speculated
161 state. This ensures that the signal context (written tm suspended) will be
162 written below the stack required for the rollback. The transaction is aborted
163 because of the treclaim, so any memory written between the tbegin and the
164 signal will be rolled back anyway.
166 For signals taken in non-TM or suspended mode, we use the
167 normal/non-checkpointed stack pointer.
170 Failure cause codes used by kernel
171 ==================================
173 These are defined in <asm/reg.h>, and distinguish different reasons why the
174 kernel aborted a transaction:
176 TM_CAUSE_RESCHED Thread was rescheduled.
177 TM_CAUSE_TLBI Software TLB invalide.
178 TM_CAUSE_FAC_UNAV FP/VEC/VSX unavailable trap.
179 TM_CAUSE_SYSCALL Currently unused; future syscalls that must abort
180 transactions for consistency will use this.
181 TM_CAUSE_SIGNAL Signal delivered.
182 TM_CAUSE_MISC Currently unused.
183 TM_CAUSE_ALIGNMENT Alignment fault.
184 TM_CAUSE_EMULATE Emulation that touched memory.
186 These can be checked by the user program's abort handler as TEXASR[0:7]. If
187 bit 7 is set, it indicates that the error is consider persistent. For example
188 a TM_CAUSE_ALIGNMENT will be persistent while a TM_CAUSE_RESCHED will not.q
193 GDB and ptrace are not currently TM-aware. If one stops during a transaction,
194 it looks like the transaction has just started (the checkpointed state is
195 presented). The transaction cannot then be continued and will take the failure
196 handler route. Furthermore, the transactional 2nd register state will be
197 inaccessible. GDB can currently be used on programs using TM, but not sensibly
198 in parts within transactions.