Says to use the RCU_CPU_STALL_TIMEOUT value converted from If the RCU grace period persists, additional CPU stall warningsĪre printed at more widely spaced intervals. sys/module/rcupdate/parameters/rcu_exp_cpu_stall_timeout, howeverĭiff -git a/kernel/rcu/bug b/kernel/rcu/bug This configuration parameter may be changed at runtime via the + used, after conversion to milliseconds, limited to a maximum of +Ě zero value causes the CONFIG_RCU_CPU_STALL_TIMEOUT value to be + grace period until it issues an RCU CPU stall warning. value causes the CONFIG_RCU_CPU_STALL_TIMEOUT value to be used, period is normally 20 milliseconds on Android devices.Ě zero grace period until it issues an RCU CPU stall warning. Of time that RCU will wait from the beginning of an expedited Same as the CONFIG_RCU_CPU_STALL_TIMEOUT parameter but only for sys/module/rcupdate/parameters/rcu_exp_cpu_stall_timeout:21000įixes: 1045a06724f3 ("remove CONFIG_ANDROID")ĭocumentation/RCU/stallwarn.rst | 9 +++++-Ģ files changed, 6 insertions(+), 5 deletions(-)ĭiff -git a/Documentation/RCU/stallwarn.rst b/Documentation/RCU/stallwarn.rst sys/module/rcupdate/parameters/rcu_cpu_stall_timeout:60 Specifying 0 which means to use the CONFIG_RCU_STALL_TIMEOUT the code includes a maximum of 21 seconds, evident when there is no longer a default of 20 ms for Android sinceĬommit 1045a06724f3 ("remove CONFIG_ANDROID"), Make the descriptions of CONFIG_EXT_RCU_CPU_STALL_TIMEOUT ` (6 subsequent siblings) 7 siblings, 0 replies 127+ messages in thread 22:31 ` crypto: x86/sha - limit FPU preemption Robert Elliott 22:31 crypto: x86 - fix RCU stalls Robert 22:31 ` Robert Elliott * rcu: correct CONFIG_EXT_RCU_CPU_STALL_TIMEOUT descriptions sys/module/sha512_ssse3/parameters/rob_call_update_fpu:32 sys/module/sha512_ssse3/parameters/rob_call_update:174 sys/module/sha512_ssse3/parameters/rob_call_finup:1541 During boot, the finup function is actuallyĬalled much more often than update - 1500 calls for 2 GiB viaįinup vs. I added some counters (not posted) to the drivers to observe I haven't doneĬomprehensive tests of different data sizes and alignments, so Hung boot, so the changes seem functionally correct. With various errors inserted, all the drivers failed self-tests or Tcrypt: testing decryption speed of sync skcipher cfb(aes) using cfb(aes-aesni)īut I don't see any problems in the source code. Tcrypt: testing encryption speed of sync skcipher cfb(aes) using cfb(aes-aesni) Tcrypt: testing encryption speed of sync skcipher cts(cbc(aes)) using cts(cbc(aes-aesni)) I originally encountered some RCU stalls with tcrypt in aesni: These are not currently limited, so I've included patches for I've included a patch to make them use the same macro name. Using either the SZ_4K macro or a direct reference to 4096 bytes:īlake-2s, chacha, nhpoly1305, poly1305, polyval This series attempts to fix the RCU stalls triggeredīy the x86 crypto drivers discussed following x86 drivers are enforcing a 4 KiB limit today, To: herbert, davem, tim.c.chen, linux-crypto, linux-kernel +Cc: Robert Elliott ` (7 more replies) 0 siblings, 8 replies 127+ messages in threadįrom: Robert Elliott 22:31 UTC ( / raw) 22:31 ` rcu: correct CONFIG_EXT_RCU_CPU_STALL_TIMEOUT descriptions Robert Elliott Crypto: x86 - fix RCU stalls All of help / color / mirror / Atom feed * crypto: x86 - fix RCU stalls 22:31 Robert Elliott
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |