From: Michael Ellerman <mpe@ellerman.id.au> To: Timothy Pearson <tpearson@raptorengineering.com> Cc: Jens Axboe <axboe@kernel.dk>, regressions <regressions@lists.linux.dev>, npiggin <npiggin@gmail.com>, christophe leroy <christophe.leroy@csgroup.eu>, linuxppc-dev <linuxppc-dev@lists.ozlabs.org> Subject: Re: [PATCH v2] powerpc: Don't clobber fr0/vs0 during fp|altivec register save Date: Wed, 22 Nov 2023 16:01:50 +1100 [thread overview] Message-ID: <87leaqjs8x.fsf@mail.lhotse> (raw) In-Reply-To: <1294229534.48922001.1700539832331.JavaMail.zimbra@raptorengineeringinc.com> Timothy Pearson <tpearson@raptorengineering.com> writes: > ... > > So a little more detail on this, just to put it to rest properly vs. > assuming hand analysis caught every possible pathway. :) > > The debugging that generates this stack trace also verifies the following in __giveup_fpu(): > > 1.) tsk->thread.fp_state.fpr doesn't contain the FPSCR contents prior to calling save_fpu() > 2.) tsk->thread.fp_state.fpr contains the FPSCR contents directly after calling save_fpu() > 3.) MSR_FP is set both in the task struct and in the live MSR. > > Only if all three conditions are met will it generate the trace. This > is a generalization of the hack I used to find the problem in the > first place. > > If the state will subsequently be reloaded from the thread struct, > that means we're reloading the registers from the thread struct that > we just verified was corrupted by the earlier save_fpu() call. There > are only two ways I can see for that to be true -- one is if the > registers were already clobbered when giveup_all() was entered, and > the other is if save_fpu() went ahead and clobbered them right here > inside giveup_all(). > > To see which scenario we were dealing with, I added a bit more > instrumentation to dump the current register state if MSR_FP bit was > already set in registers (i.e. not dumping data from task struct, but > using the live FPU registers instead), and sure enough the registers > are corrupt on entry, so something else has already called save_fpu() > before we even hit giveup_all() in this call chain. Can you share the debug patch you're using? cheers
WARNING: multiple messages have this Message-ID (diff)
From: Michael Ellerman <mpe@ellerman.id.au> To: Timothy Pearson <tpearson@raptorengineering.com> Cc: Jens Axboe <axboe@kernel.dk>, linuxppc-dev <linuxppc-dev@lists.ozlabs.org>, regressions <regressions@lists.linux.dev>, npiggin <npiggin@gmail.com> Subject: Re: [PATCH v2] powerpc: Don't clobber fr0/vs0 during fp|altivec register save Date: Wed, 22 Nov 2023 16:01:50 +1100 [thread overview] Message-ID: <87leaqjs8x.fsf@mail.lhotse> (raw) In-Reply-To: <1294229534.48922001.1700539832331.JavaMail.zimbra@raptorengineeringinc.com> Timothy Pearson <tpearson@raptorengineering.com> writes: > ... > > So a little more detail on this, just to put it to rest properly vs. > assuming hand analysis caught every possible pathway. :) > > The debugging that generates this stack trace also verifies the following in __giveup_fpu(): > > 1.) tsk->thread.fp_state.fpr doesn't contain the FPSCR contents prior to calling save_fpu() > 2.) tsk->thread.fp_state.fpr contains the FPSCR contents directly after calling save_fpu() > 3.) MSR_FP is set both in the task struct and in the live MSR. > > Only if all three conditions are met will it generate the trace. This > is a generalization of the hack I used to find the problem in the > first place. > > If the state will subsequently be reloaded from the thread struct, > that means we're reloading the registers from the thread struct that > we just verified was corrupted by the earlier save_fpu() call. There > are only two ways I can see for that to be true -- one is if the > registers were already clobbered when giveup_all() was entered, and > the other is if save_fpu() went ahead and clobbered them right here > inside giveup_all(). > > To see which scenario we were dealing with, I added a bit more > instrumentation to dump the current register state if MSR_FP bit was > already set in registers (i.e. not dumping data from task struct, but > using the live FPU registers instead), and sure enough the registers > are corrupt on entry, so something else has already called save_fpu() > before we even hit giveup_all() in this call chain. Can you share the debug patch you're using? cheers
next prev parent reply other threads:[~2023-11-22 5:02 UTC|newest] Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top 2023-11-19 15:18 [PATCH v2] powerpc: Don't clobber fr0/vs0 during fp|altivec register save Timothy Pearson 2023-11-20 7:10 ` Michael Ellerman 2023-11-20 14:32 ` Nicholas Piggin 2023-11-20 16:45 ` Timothy Pearson 2023-11-20 16:45 ` Timothy Pearson 2023-11-20 23:39 ` Michael Ellerman 2023-11-20 23:39 ` Michael Ellerman 2023-11-21 0:27 ` Nicholas Piggin 2023-11-21 0:27 ` Nicholas Piggin 2023-11-21 1:23 ` Timothy Pearson 2023-11-21 1:23 ` Timothy Pearson 2023-11-21 7:56 ` Nicholas Piggin 2023-11-21 7:56 ` Nicholas Piggin 2023-11-21 4:10 ` Timothy Pearson 2023-11-21 4:10 ` Timothy Pearson 2023-11-21 4:26 ` Timothy Pearson 2023-11-21 4:26 ` Timothy Pearson 2023-11-21 7:54 ` Nicholas Piggin 2023-11-21 7:54 ` Nicholas Piggin 2023-11-22 5:01 ` Michael Ellerman [this message] 2023-11-22 5:01 ` Michael Ellerman 2023-11-24 0:01 ` Timothy Pearson 2023-11-24 0:01 ` Timothy Pearson 2023-11-27 18:39 ` Timothy Pearson 2023-11-27 18:39 ` Timothy Pearson 2023-11-27 19:58 ` Christophe Leroy 2023-11-27 19:58 ` Christophe Leroy 2023-11-28 0:59 ` Michael Ellerman 2023-11-28 0:59 ` Michael Ellerman 2023-11-28 1:40 ` Nicholas Piggin 2023-11-28 1:40 ` Nicholas Piggin 2023-11-27 22:53 ` Michael Ellerman 2023-11-27 22:53 ` Michael Ellerman 2023-11-28 12:57 ` Michael Ellerman 2023-11-28 12:57 ` Michael Ellerman 2023-11-30 16:29 ` Timothy Pearson 2023-11-30 16:29 ` Timothy Pearson 2023-11-21 0:18 ` Nicholas Piggin 2023-11-21 0:18 ` Nicholas Piggin 2023-12-02 23:00 ` Michael Ellerman
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=87leaqjs8x.fsf@mail.lhotse \ --to=mpe@ellerman.id.au \ --cc=axboe@kernel.dk \ --cc=christophe.leroy@csgroup.eu \ --cc=linuxppc-dev@lists.ozlabs.org \ --cc=npiggin@gmail.com \ --cc=regressions@lists.linux.dev \ --cc=tpearson@raptorengineering.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
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.