From: Janosch Frank <frankja@linux.ibm.com>
To: Christian Borntraeger <borntraeger@de.ibm.com>, qemu-devel@nongnu.org
Cc: thuth@redhat.com, cohuck@redhat.com, david@redhat.com
Subject: Re: [PATCH] pc-bios: s390x: Add a comment to the io and external new PSW setup
Date: Wed, 22 Jul 2020 10:05:53 +0200 [thread overview]
Message-ID: <cd2c8bb2-413f-b6b9-ae16-e79e838b503d@linux.ibm.com> (raw)
In-Reply-To: <2e6134fb-3e34-4d3a-1bf8-1665ceffef38@de.ibm.com>
[-- Attachment #1.1: Type: text/plain, Size: 2901 bytes --]
On 7/22/20 9:39 AM, Christian Borntraeger wrote:
>
>
> On 22.07.20 09:24, Janosch Frank wrote:
>> On 7/22/20 8:43 AM, Christian Borntraeger wrote:
>>>
>>>
>>> On 15.07.20 16:08, Janosch Frank wrote:
>>>> Normally they don't need to be set up before waiting for an interrupt
>>>> but are set up on boot. The BIOS however might overwrite the lowcore
>>>> (and hence the PSWs) when loading a blob into memory and therefore
>>>> needs to set up those PSWs more often.
>>>
>>> Now when I read the new comment this actually inidicates a bug.
>>> When do we restore the original content? If the loaded program
>>> does have interrupt handlers in the original image and relies on that
>>> then we are broken, no?
>>
>> I haven't seen references to a save/restore functionality for those
>> PSWs. And I also think it's not that easy to do because we have multiple
>> ways of loading data and if we want to print when loading we might end
>> up overwriting and then saving the written value for a later restore.
>>
>> I need to have a closer look at how virtio works, but wouldn't we have a
>> chicken - egg problem with IO interrupts for IO that writes the prefix?
>>
>> The BIOS often has "interesting" solutions to problems.
>> If you have a quick fix, be my guest and send it. If not I'd put it on
>> my todo list or let Stefan make it a proper dev item.
>
> Maybe a global fixup table in BIOS memory that restores all the memory that
> we messed with when we hand over control? Can you at least change the comment
> here to add a fixme?
Sure
>
>>
>>>
>>>>
>>>> Signed-off-by: Janosch Frank <frankja@linux.ibm.com>
>>>> ---
>>>> pc-bios/s390-ccw/start.S | 10 ++++++++--
>>>> 1 file changed, 8 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/pc-bios/s390-ccw/start.S b/pc-bios/s390-ccw/start.S
>>>> index 01c4c21b26..b0fcb918cc 100644
>>>> --- a/pc-bios/s390-ccw/start.S
>>>> +++ b/pc-bios/s390-ccw/start.S
>>>> @@ -64,7 +64,10 @@ consume_sclp_int:
>>>> stctg %c0,%c0,0(%r15)
>>>> oi 6(%r15),0x2
>>>> lctlg %c0,%c0,0(%r15)
>>>> - /* prepare external call handler */
>>>> + /*
>>>> + * Prepare external new PSW as it might have been overwritten
>>>> + * by a loaded blob
>>>> + */
>>>> larl %r1, external_new_code
>>>> stg %r1, 0x1b8
>>>> larl %r1, external_new_mask
>>>> @@ -84,7 +87,10 @@ consume_io_int:
>>>> stctg %c6,%c6,0(%r15)
>>>> oi 4(%r15), 0xff
>>>> lctlg %c6,%c6,0(%r15)
>>>> - /* prepare i/o call handler */
>>>> + /*
>>>> + * Prepare i/o new PSW as it might have been overwritten
>>>> + * by a loaded blob
>>>> + */
>>>> larl %r1, io_new_code
>>>> stg %r1, 0x1f8
>>>> larl %r1, io_new_mask
>>>>
>>
>>
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
next prev parent reply other threads:[~2020-07-22 8:10 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-15 9:40 [PATCH 0/7] pc-bios: s390x: Cleanup part 2 Janosch Frank
2020-07-15 9:40 ` [PATCH 1/7] pc-bios: s390x: Fix bootmap.c zipl component entry data handling Janosch Frank
2020-07-17 15:05 ` Thomas Huth
2020-07-22 6:50 ` Christian Borntraeger
2020-07-22 7:30 ` Janosch Frank
2020-07-22 7:33 ` Christian Borntraeger
2020-07-22 8:06 ` Janosch Frank
2020-07-15 9:40 ` [PATCH 2/7] pc-bios: s390x: Cleanup jump to ipl code Janosch Frank
2020-07-17 15:13 ` Thomas Huth
2020-07-21 13:07 ` Janosch Frank
2020-07-21 13:54 ` Christian Borntraeger
2020-07-15 9:40 ` [PATCH 3/7] pc-bios: s390x: Remove unneeded dasd-ipl.c reset psw mask changes Janosch Frank
2020-07-20 11:45 ` Thomas Huth
2020-07-20 12:16 ` Janosch Frank
2020-07-21 7:10 ` Thomas Huth
2020-07-15 9:40 ` [PATCH 4/7] pc-bios: s390x: Rework data initialization Janosch Frank
2020-07-20 11:56 ` Thomas Huth
2020-07-20 12:10 ` Janosch Frank
2020-07-15 9:40 ` [PATCH 5/7] pc-bios: s390x: Replace lowcore offsets with pointers in dasd-ipl.c Janosch Frank
2020-07-21 7:00 ` Thomas Huth
2020-07-15 9:40 ` [PATCH 6/7] pc-bios: s390x: Use PSW constants in start.S Janosch Frank
2020-07-21 7:05 ` Thomas Huth
2020-07-22 6:47 ` Christian Borntraeger
2020-07-15 9:40 ` [PATCH 7/7] pc-bios: s390x: Setup io and ext new psws only once Janosch Frank
2020-07-15 13:13 ` Janosch Frank
2020-07-15 13:16 ` Christian Borntraeger
2020-07-15 14:08 ` [PATCH] pc-bios: s390x: Add a comment to the io and external new PSW setup Janosch Frank
2020-07-21 7:03 ` Thomas Huth
2020-07-22 6:43 ` Christian Borntraeger
2020-07-22 7:24 ` Janosch Frank
2020-07-22 7:39 ` Christian Borntraeger
2020-07-22 8:05 ` Janosch Frank [this message]
2020-08-27 9:20 ` Thomas Huth
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=cd2c8bb2-413f-b6b9-ae16-e79e838b503d@linux.ibm.com \
--to=frankja@linux.ibm.com \
--cc=borntraeger@de.ibm.com \
--cc=cohuck@redhat.com \
--cc=david@redhat.com \
--cc=qemu-devel@nongnu.org \
--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).