diff for duplicates of <1502936450.28554.98.camel@gmail.com> diff --git a/a/1.txt b/N1/1.txt index a542b3a..e5de30c 100644 --- a/a/1.txt +++ b/N1/1.txt @@ -7,7 +7,7 @@ On Fri, 2017-08-11 at 18:02 +0100, James Morse wrote: > > On 07/08/17 06:09, Hoeun Ryu wrote: > > -> > ?Commit 0ee5941 : (x86/panic: replace smp_send_stop() with kdump friendly +> > Commit 0ee5941 : (x86/panic: replace smp_send_stop() with kdump friendly > > version in panic path) introduced crash_smp_send_stop() which is a weak > > function and can be overriden by architecture codes to fix the side effect > (overridden) @@ -20,18 +20,18 @@ It'll be fixed in the next version. > > caused by commit f06e515 : (kernel/panic.c: add "crash_kexec_post_ > > notifiers" option). > > -> > ?ARM64 architecture uses the weak version function and the problem is that +> > ARM64 architecture uses the weak version function and the problem is that > > the weak function simply calls smp_send_stop() which makes other CPUs > > offline and takes away the chance to save crash information for nonpanic > > CPUs in machine_crash_shutdown() when crash_kexec_post_notifiers kernel > > option is enabled. > > -> > ?Calling smp_send_crash_stop() in machine_crash_shutdown() is useless +> > Calling smp_send_crash_stop() in machine_crash_shutdown() is useless > > because all nonpanic CPUs are already offline by smp_send_stop() in this > > case and smp_send_crash_stop() only works against online CPUs. > > > -> > ?The result is that /proc/vmcore is not available with the error messages; +> > The result is that /proc/vmcore is not available with the error messages; > > "Warning: Zero PT_NOTE entries found", "Kdump: vmcore not initialized". > When I tried this I got one of these warnings for each secondary CPU, but the > vmcore file was still available. When I ran 'crash' on the vmcore it reported: @@ -48,7 +48,7 @@ I'll change the commit log with yours. > > > -> > ?crash_smp_send_stop() is implemented to fix this problem by replacing the +> > crash_smp_send_stop() is implemented to fix this problem by replacing the > > exising smp_send_crash_stop() and adding a check for multiple calling to > (existing) @@ -64,18 +64,18 @@ It'll be fixed in the next version. > > > > * crash_kexec_post_notifiers : false > > -> > ? panic() -> > ????__crash_kexec() -> > ??????machine_crash_shutdown() -> > ????????crash_smp_send_stop()????<= save crash dump for nonpanic cores +> > panic() +> > __crash_kexec() +> > machine_crash_shutdown() +> > crash_smp_send_stop() <= save crash dump for nonpanic cores > > > > * crash_kexec_post_notifiers : true > > -> > ? panic() -> > ????crash_smp_send_stop()????????<= save crash dump for nonpanic cores -> > ????__crash_kexec() -> > ??????machine_crash_shutdown() -> > ????????crash_smp_send_stop()????<= just return. +> > panic() +> > crash_smp_send_stop() <= save crash dump for nonpanic cores +> > __crash_kexec() +> > machine_crash_shutdown() +> > crash_smp_send_stop() <= just return. > > > > > diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c @@ -83,20 +83,20 @@ It'll be fixed in the next version. > > --- a/arch/arm64/kernel/smp.c > > +++ b/arch/arm64/kernel/smp.c > > @@ -977,11 +977,21 @@ void smp_send_stop(void) -> > ?} -> > ? -> > ?#ifdef CONFIG_KEXEC_CORE +> > } +> > +> > #ifdef CONFIG_KEXEC_CORE > > -void smp_send_crash_stop(void) > > +void crash_smp_send_stop(void) -> > ?{ +> > { > > + static int cpus_stopped; -> > ? cpumask_t mask; -> > ? unsigned long timeout; -> > ? +> > cpumask_t mask; +> > unsigned long timeout; +> > > > + /* -> > + ?* This function can be called twice in panic path, but obviously -> > + ?* we execute this only once. -> > + ?*/ +> > + * This function can be called twice in panic path, but obviously +> > + * we execute this only once. +> > + */ > > + if (cpus_stopped) > > + return; > > + @@ -110,20 +110,20 @@ It'll be fixed in the next version. This variable is simply to prevent from calling 'smp_cross_call(&mask, IPI_CPU_CRASH_STOP)' twice in the machine_crash_shutdown(). Please look at following call path. -? + * crash_kexec_post_notifiers : true -?panic() -? ? ?crash_smp_send_stop() { -? ? ? ? ?... -? ? ? ? ?cpu_stopped = 1 ? ? ? ? ? ? ?<= make it '1' -? ? ? ? ?smp_cross_call() ? ? ? ? ? ? <= save crash dump for nonpanic cores -? ? ?} -?? ? __crash_kexec() -? ? ? ? ?machine_crash_shutdown() -? ? ? ? ? ? ?crash_smp_send_stop() {? -? ? ? ? ? ? ? ? ?if (cpu_stopped) -? ? ? ? ? ? ? ? ? ? ?return ? ? ? ? ? <= just return. -? ? ? ? ? ? ?} + panic() + crash_smp_send_stop() { + ... + cpu_stopped = 1 <= make it '1' + smp_cross_call() <= save crash dump for nonpanic cores + } + __crash_kexec() + machine_crash_shutdown() + crash_smp_send_stop() { + if (cpu_stopped) + return <= just return. + } > > > Other than my '/proc/vmcore is not available' question above, this looks fine to me: diff --git a/a/content_digest b/N1/content_digest index 0d223fc..47dc6a8 100644 --- a/a/content_digest +++ b/N1/content_digest @@ -1,9 +1,24 @@ "ref\01502082623-23952-1-git-send-email-hoeun.ryu@gmail.com\0" "ref\0598DE33E.6050606@arm.com\0" - "From\0hoeun.ryu@gmail.com (Hoeun Ryu)\0" - "Subject\0[PATCHv2] arm64:kexec: have own crash_smp_send_stop() for crash dump for nonpanic cores\0" + "From\0Hoeun Ryu <hoeun.ryu@gmail.com>\0" + "Subject\0Re: [PATCHv2] arm64:kexec: have own crash_smp_send_stop() for crash dump for nonpanic cores\0" "Date\0Thu, 17 Aug 2017 11:20:50 +0900\0" - "To\0linux-arm-kernel@lists.infradead.org\0" + "To\0James Morse <james.morse@arm.com>\0" + "Cc\0Catalin Marinas <catalin.marinas@arm.com>" + Will Deacon <will.deacon@arm.com> + Mark Rutland <mark.rutland@arm.com> + AKASHI Takahiro <takahiro.akashi@linaro.org> + Robin Murphy <robin.murphy@arm.com> + Ard Biesheuvel <ard.biesheuvel@linaro.org> + Ingo Molnar <mingo@kernel.org> + Peter Zijlstra (Intel) <peterz@infradead.org> + Suzuki K Poulose <suzuki.poulose@arm.com> + David Daney <david.daney@cavium.com> + Rob Herring <robh@kernel.org> + Kefeng Wang <wangkefeng.wang@huawei.com> + Thomas Gleixner <tglx@linutronix.de> + linux-arm-kernel@lists.infradead.org + " linux-kernel@vger.kernel.org\0" "\00:1\0" "b\0" "Hello, James.\n" @@ -15,7 +30,7 @@ "> \n" "> On 07/08/17 06:09, Hoeun Ryu wrote:\n" "> > \n" - "> > ?Commit 0ee5941 : (x86/panic: replace smp_send_stop() with kdump friendly\n" + "> > \302\240Commit 0ee5941 : (x86/panic: replace smp_send_stop() with kdump friendly\n" "> > version in panic path) introduced crash_smp_send_stop() which is a weak\n" "> > function and can be overriden by architecture codes to fix the side effect\n" "> (overridden)\n" @@ -28,18 +43,18 @@ "> > caused by commit f06e515 : (kernel/panic.c: add \"crash_kexec_post_\n" "> > notifiers\" option).\n" "> > \n" - "> > ?ARM64 architecture uses the weak version function and the problem is that\n" + "> > \302\240ARM64 architecture uses the weak version function and the problem is that\n" "> > the weak function simply calls smp_send_stop() which makes other CPUs\n" "> > offline and takes away the chance to save crash information for nonpanic\n" "> > CPUs in machine_crash_shutdown() when crash_kexec_post_notifiers kernel\n" "> > option is enabled.\n" "> > \n" - "> > ?Calling smp_send_crash_stop() in machine_crash_shutdown() is useless\n" + "> > \302\240Calling smp_send_crash_stop() in machine_crash_shutdown() is useless\n" "> > because all nonpanic CPUs are already offline by smp_send_stop() in this\n" "> > case and smp_send_crash_stop() only works against online CPUs.\n" "> \n" "> > \n" - "> > ?The result is that /proc/vmcore is not available with the error messages;\n" + "> > \302\240The result is that /proc/vmcore is not available with the error messages;\n" "> > \"Warning: Zero PT_NOTE entries found\", \"Kdump: vmcore not initialized\".\n" "> When I tried this I got one of these warnings for each secondary CPU, but the\n" "> vmcore file was still available. When I ran 'crash' on the vmcore it reported:\n" @@ -56,7 +71,7 @@ "\n" "> \n" "> > \n" - "> > ?crash_smp_send_stop() is implemented to fix this problem by replacing the\n" + "> > \302\240crash_smp_send_stop() is implemented to fix this problem by replacing the\n" "> > exising smp_send_crash_stop() and adding a check for multiple calling to\n" "> (existing)\n" "\n" @@ -72,18 +87,18 @@ "> > \n" "> > * crash_kexec_post_notifiers : false\n" "> > \n" - "> > ? panic()\n" - "> > ????__crash_kexec()\n" - "> > ??????machine_crash_shutdown()\n" - "> > ????????crash_smp_send_stop()????<= save crash dump for nonpanic cores\n" + "> > \302\240 panic()\n" + "> > \302\240\302\240\302\240\302\240__crash_kexec()\n" + "> > \302\240\302\240\302\240\302\240\302\240\302\240machine_crash_shutdown()\n" + "> > \302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240crash_smp_send_stop()\302\240\302\240\302\240\302\240<= save crash dump for nonpanic cores\n" "> > \n" "> > * crash_kexec_post_notifiers : true\n" "> > \n" - "> > ? panic()\n" - "> > ????crash_smp_send_stop()????????<= save crash dump for nonpanic cores\n" - "> > ????__crash_kexec()\n" - "> > ??????machine_crash_shutdown()\n" - "> > ????????crash_smp_send_stop()????<= just return.\n" + "> > \302\240 panic()\n" + "> > \302\240\302\240\302\240\302\240crash_smp_send_stop()\302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240<= save crash dump for nonpanic cores\n" + "> > \302\240\302\240\302\240\302\240__crash_kexec()\n" + "> > \302\240\302\240\302\240\302\240\302\240\302\240machine_crash_shutdown()\n" + "> > \302\240\302\240\302\240\302\240\302\240\302\240\302\240\302\240crash_smp_send_stop()\302\240\302\240\302\240\302\240<= just return.\n" "> \n" "> > \n" "> > diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c\n" @@ -91,20 +106,20 @@ "> > --- a/arch/arm64/kernel/smp.c\n" "> > +++ b/arch/arm64/kernel/smp.c\n" "> > @@ -977,11 +977,21 @@ void smp_send_stop(void)\n" - "> > ?}\n" - "> > ?\n" - "> > ?#ifdef CONFIG_KEXEC_CORE\n" + "> > \302\240}\n" + "> > \302\240\n" + "> > \302\240#ifdef CONFIG_KEXEC_CORE\n" "> > -void smp_send_crash_stop(void)\n" "> > +void crash_smp_send_stop(void)\n" - "> > ?{\n" + "> > \302\240{\n" "> > +\tstatic int cpus_stopped;\n" - "> > ?\tcpumask_t mask;\n" - "> > ?\tunsigned long timeout;\n" - "> > ?\n" + "> > \302\240\tcpumask_t mask;\n" + "> > \302\240\tunsigned long timeout;\n" + "> > \302\240\n" "> > +\t/*\n" - "> > +\t?* This function can be called twice in panic path, but obviously\n" - "> > +\t?* we execute this only once.\n" - "> > +\t?*/\n" + "> > +\t\302\240* This function can be called twice in panic path, but obviously\n" + "> > +\t\302\240* we execute this only once.\n" + "> > +\t\302\240*/\n" "> > +\tif (cpus_stopped)\n" "> > +\t\treturn;\n" "> > +\n" @@ -118,20 +133,20 @@ "This variable is simply to prevent from calling\n" "'smp_cross_call(&mask, IPI_CPU_CRASH_STOP)' twice in the machine_crash_shutdown().\n" "Please look at following call path.\n" - "?\n" + "\302\240\n" "* crash_kexec_post_notifiers : true\n" - "?panic()\n" - "? ? ?crash_smp_send_stop() {\n" - "? ? ? ? ?...\n" - "? ? ? ? ?cpu_stopped = 1 ? ? ? ? ? ? ?<= make it '1'\n" - "? ? ? ? ?smp_cross_call() ? ? ? ? ? ? <= save crash dump for nonpanic cores\n" - "? ? ?}\n" - "?? ? __crash_kexec()\n" - "? ? ? ? ?machine_crash_shutdown()\n" - "? ? ? ? ? ? ?crash_smp_send_stop() {?\n" - "? ? ? ? ? ? ? ? ?if (cpu_stopped)\n" - "? ? ? ? ? ? ? ? ? ? ?return ? ? ? ? ? <= just return.\n" - "? ? ? ? ? ? ?}\n" + "\302\240panic()\n" + "\302\240 \302\240 \302\240crash_smp_send_stop() {\n" + "\302\240 \302\240 \302\240 \302\240 \302\240...\n" + "\302\240 \302\240 \302\240 \302\240 \302\240cpu_stopped = 1 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240<= make it '1'\n" + "\302\240 \302\240 \302\240 \302\240 \302\240smp_cross_call() \302\240 \302\240 \302\240 \302\240 \302\240 \302\240 <= save crash dump for nonpanic cores\n" + "\302\240 \302\240 \302\240}\n" + "\302\240\302\240 \302\240 __crash_kexec()\n" + "\302\240 \302\240 \302\240 \302\240 \302\240machine_crash_shutdown()\n" + "\302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240crash_smp_send_stop() {\302\240\n" + "\302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240if (cpu_stopped)\n" + "\302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240return \302\240 \302\240 \302\240 \302\240 \302\240 <= just return.\n" + "\302\240 \302\240 \302\240 \302\240 \302\240 \302\240 \302\240}\n" "> \n" "> \n" "> Other than my '/proc/vmcore is not available' question above, this looks fine to me:\n" @@ -146,4 +161,4 @@ "> \n" > -eca7a9ef830b2f36f25f3b015ed70a2939eb2da2b69e34543115a83f35166012 +78fe39ff801742e13c902f49d70c67eeb0684f9dfb0439aed62635ca50a5f4d9
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.