From: Janosch Frank <frankja@linux.ibm.com>
To: Claudio Imbrenda <imbrenda@linux.ibm.com>
Cc: kvm@vger.kernel.org, linux-s390@vger.kernel.org,
thuth@redhat.com, david@redhat.com, nsg@linux.ibm.com,
nrb@linux.ibm.com
Subject: Re: [kvm-unit-tests PATCH v2 7/7] lib: s390x: Handle debug prints for SIE exceptions correctly
Date: Fri, 13 Jan 2023 16:50:51 +0100 [thread overview]
Message-ID: <54e21150-714c-4a06-59ba-a599bae76bbf@linux.ibm.com> (raw)
In-Reply-To: <20230112181353.0ed31250@p-imbrenda>
On 1/12/23 18:13, Claudio Imbrenda wrote:
> On Thu, 12 Jan 2023 15:45:48 +0000
> Janosch Frank <frankja@linux.ibm.com> wrote:
>
>> When we leave SIE due to an exception, we'll still have guest values
>> in registers 0 - 13 and that's not clearly portraied in our debug
>> prints. So let's fix that.
>>
>> Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
>> ---
>> lib/s390x/interrupt.c | 46 ++++++++++++++++++++++++++++++++++++++-----
>> lib/s390x/sie.h | 2 ++
>> s390x/cpu.S | 6 ++++--
>> 3 files changed, 47 insertions(+), 7 deletions(-)
>>
>> diff --git a/lib/s390x/interrupt.c b/lib/s390x/interrupt.c
>> index dadb7415..ff47c2c2 100644
>> --- a/lib/s390x/interrupt.c
>> +++ b/lib/s390x/interrupt.c
>> @@ -9,6 +9,7 @@
>> */
>> #include <libcflat.h>
>> #include <asm/barrier.h>
>> +#include <asm/asm-offsets.h>
>> #include <sclp.h>
>> #include <interrupt.h>
>> #include <sie.h>
>> @@ -188,9 +189,12 @@ static void print_storage_exception_information(void)
>> }
>> }
>>
>> -static void print_int_regs(struct stack_frame_int *stack)
>> +static void print_int_regs(struct stack_frame_int *stack, bool sie)
>> {
>> + struct kvm_s390_sie_block *sblk;
>> +
>> printf("\n");
>> + printf("%s\n", sie ? "Guest registers:" : "Host registers:");
>> printf("GPRS:\n");
>> printf("%016lx %016lx %016lx %016lx\n",
>> stack->grs1[0], stack->grs1[1], stack->grs0[0], stack->grs0[1]);
>> @@ -198,24 +202,56 @@ static void print_int_regs(struct stack_frame_int *stack)
>> stack->grs0[2], stack->grs0[3], stack->grs0[4], stack->grs0[5]);
>> printf("%016lx %016lx %016lx %016lx\n",
>> stack->grs0[6], stack->grs0[7], stack->grs0[8], stack->grs0[9]);
>> - printf("%016lx %016lx %016lx %016lx\n",
>> - stack->grs0[10], stack->grs0[11], stack->grs0[12], stack->grs0[13]);
>> +
>> + if (sie) {
>> + sblk = (struct kvm_s390_sie_block *)stack->grs0[12];
>> + printf("%016lx %016lx %016lx %016lx\n",
>> + stack->grs0[10], stack->grs0[11], sblk->gg14, sblk->gg15);
>> + } else {
>> + printf("%016lx %016lx %016lx %016lx\n",
>> + stack->grs0[10], stack->grs0[11], stack->grs0[12], stack->grs0[13]);
>> + }
>> +
>> printf("\n");
>> }
>>
>> static void print_pgm_info(struct stack_frame_int *stack)
>>
>> {
>> - bool in_sie;
>> + bool in_sie, in_sie_gregs;
>> + struct vm_save_area *vregs;
>>
>> in_sie = (lowcore.pgm_old_psw.addr >= (uintptr_t)sie_entry &&
>> lowcore.pgm_old_psw.addr <= (uintptr_t)sie_exit);
>> + in_sie_gregs = (lowcore.pgm_old_psw.addr >= (uintptr_t)sie_entry_gregs &&
>> + lowcore.pgm_old_psw.addr <= (uintptr_t)sie_exit_gregs);
>
> can you explain why <= instead of < ? (I think I know why, but a
> comment would not hurt)
This might be wrong now that the exit label points behind a lmg which
can only cause the operation or the addressing exception.
The operation exception can only happen on a LMY without the long
displacement HW support and the addressing exception should only happen
in circumstances where any processed instruction will likely result in a
new exception being thrown.
>
> with that fixed:
>
> Reviewed-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
>
>>
>> printf("\n");
>> printf("Unexpected program interrupt %s: %#x on cpu %d at %#lx, ilen %d\n",
>> in_sie ? "in SIE" : "",
>> lowcore.pgm_int_code, stap(), lowcore.pgm_old_psw.addr, lowcore.pgm_int_id);
>> - print_int_regs(stack);
>> +
>> + /*
>> + * If we fall out of SIE before loading the host registers,
>> + * then we need to do it here so we print the host registers
>> + * and not the guest registers.
>> + *
>> + * Back tracing is actually not a problem since SIE restores gr15.
>> + */
>> + if (in_sie_gregs) {
>> + print_int_regs(stack, true);
>> + vregs = *((struct vm_save_area **)(stack->grs0[13] + __SF_SIE_SAVEAREA));
>> +
>> + /*
>> + * The grs are not linear on the interrupt stack frame.
>> + * We copy 0 and 1 here and 2 - 15 with the memcopy below.
>> + */
>> + stack->grs1[0] = vregs->host.grs[0];
>> + stack->grs1[1] = vregs->host.grs[1];
>> + /* 2 - 15 */
>> + memcpy(stack->grs0, &vregs->host.grs[2], sizeof(stack->grs0) - 8);
>> + }
>> + print_int_regs(stack, false);
>> dump_stack();
>>
>> /* Dump stack doesn't end with a \n so we add it here instead */
>> diff --git a/lib/s390x/sie.h b/lib/s390x/sie.h
>> index a27a8401..147cb0f2 100644
>> --- a/lib/s390x/sie.h
>> +++ b/lib/s390x/sie.h
>> @@ -273,6 +273,8 @@ struct vm {
>>
>> extern void sie_entry(void);
>> extern void sie_exit(void);
>> +extern void sie_entry_gregs(void);
>> +extern void sie_exit_gregs(void);
>> extern void sie64a(struct kvm_s390_sie_block *sblk, struct vm_save_area *save_area);
>> void sie(struct vm *vm);
>> void sie_expect_validity(struct vm *vm);
>> diff --git a/s390x/cpu.S b/s390x/cpu.S
>> index 45bd551a..9155b044 100644
>> --- a/s390x/cpu.S
>> +++ b/s390x/cpu.S
>> @@ -82,7 +82,8 @@ sie64a:
>> # Store scb and save_area pointer into stack frame
>> stg %r2,__SF_SIE_CONTROL(%r15) # save control block pointer
>> stg %r3,__SF_SIE_SAVEAREA(%r15) # save guest register save area
>> -
>> +.globl sie_entry_gregs
>> +sie_entry_gregs:
>> # Load guest's gprs, fprs and fpc
>> .irp i, 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
>> ld \i, \i * 8 + SIE_SAVEAREA_GUEST_FPRS(%r3)
>> @@ -121,7 +122,8 @@ sie_exit:
>> .endr
>> lfpc SIE_SAVEAREA_HOST_FPC(%r14)
>> lmg %r0,%r14,SIE_SAVEAREA_HOST_GRS(%r14) # restore kernel registers
>> -
>> +.globl sie_exit_gregs
>> +sie_exit_gregs:
>> br %r14
>>
>> .align 8
>
prev parent reply other threads:[~2023-01-13 16:00 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-12 15:45 [kvm-unit-tests PATCH v2 0/7] s390x: Snippet fixes Janosch Frank
2023-01-12 15:45 ` [kvm-unit-tests PATCH v2 1/7] s390x: Cleanup flat.lds Janosch Frank
2023-01-12 17:05 ` Claudio Imbrenda
2023-01-12 15:45 ` [kvm-unit-tests PATCH v2 2/7] s390x: snippets: c: " Janosch Frank
2023-01-12 17:05 ` Claudio Imbrenda
2023-01-12 15:45 ` [kvm-unit-tests PATCH v2 3/7] s390x: Add a linker script to assembly snippets Janosch Frank
2023-01-12 17:05 ` Claudio Imbrenda
2023-01-12 15:45 ` [kvm-unit-tests PATCH v2 4/7] s390x: snippets: Fix SET_PSW_NEW_ADDR macro Janosch Frank
2023-01-12 15:45 ` [kvm-unit-tests PATCH v2 5/7] lib: s390x: sie: Set guest memory pointer Janosch Frank
2023-01-12 15:45 ` [kvm-unit-tests PATCH v2 6/7] s390x: Clear first stack frame and end backtrace early Janosch Frank
2023-01-12 15:45 ` [kvm-unit-tests PATCH v2 7/7] lib: s390x: Handle debug prints for SIE exceptions correctly Janosch Frank
2023-01-12 17:13 ` Claudio Imbrenda
2023-01-13 15:50 ` Janosch Frank [this message]
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=54e21150-714c-4a06-59ba-a599bae76bbf@linux.ibm.com \
--to=frankja@linux.ibm.com \
--cc=david@redhat.com \
--cc=imbrenda@linux.ibm.com \
--cc=kvm@vger.kernel.org \
--cc=linux-s390@vger.kernel.org \
--cc=nrb@linux.ibm.com \
--cc=nsg@linux.ibm.com \
--cc=thuth@redhat.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: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).