Hi,
I've the following error :
gcc -D__KERNEL__ -I/usr/src/kernel-sources-2.5.1-pre2/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer
-fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2 -march=k6
-DMODULE -DMODVERSIONS -include
/usr/src/kernel-sources-2.5.1-pre2/include/linux/modversions.h -c -o
aha1542.o aha1542.c
aha1542.c: In function `do_aha1542_intr_handle':
aha1542.c:423: `io_request_lock' undeclared (first use in this function)
aha1542.c:423: (Each undeclared identifier is reported only once
aha1542.c:423: for each function it appears in.)
aha1542.c: In function `aha1542_bus_reset':
aha1542.c:1479: `io_request_lock' undeclared (first use in this function)
aha1542.c: In function `aha1542_host_reset':
aha1542.c:1543: `io_request_lock' undeclared (first use in this function)
aha1542.c: At top level:
aha1542.c:114: warning: `setup_str' defined but not used
make[3]: *** [aha1542.o] Erreur 1
make[3]: Leaving directory `/usr/src/kernel-sources-2.5.1-pre2/drivers/scsi'
make[2]: *** [_modsubdir_scsi] Erreur 2
make[2]: Leaving directory `/usr/src/kernel-sources-2.5.1-pre2/drivers'
make[1]: *** [_mod_drivers] Erreur 2
make[1]: Leaving directory `/usr/src/kernel-sources-2.5.1-pre2'
make: *** [stamp-build] Erreur 2
-------
Regards
Jean-Luc
drivers/scsi/ppa.c (the Iomega low-level parallel port SCSI driver) has the same
problem.
Wayne
f5ibh <[email protected]> on 11/27/2001 02:44:29 PM
To: [email protected]
cc: (bcc: Wayne Brown/Corporate/Altec)
Subject: 2.5.1-pre2 does not compile
Hi,
I've the following error :
gcc -D__KERNEL__ -I/usr/src/kernel-sources-2.5.1-pre2/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer
-fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2 -march=k6
-DMODULE -DMODVERSIONS -include
/usr/src/kernel-sources-2.5.1-pre2/include/linux/modversions.h -c -o
aha1542.o aha1542.c
aha1542.c: In function `do_aha1542_intr_handle':
aha1542.c:423: `io_request_lock' undeclared (first use in this function)
aha1542.c:423: (Each undeclared identifier is reported only once
aha1542.c:423: for each function it appears in.)
aha1542.c: In function `aha1542_bus_reset':
aha1542.c:1479: `io_request_lock' undeclared (first use in this function)
aha1542.c: In function `aha1542_host_reset':
aha1542.c:1543: `io_request_lock' undeclared (first use in this function)
aha1542.c: At top level:
aha1542.c:114: warning: `setup_str' defined but not used
make[3]: *** [aha1542.o] Erreur 1
make[3]: Leaving directory `/usr/src/kernel-sources-2.5.1-pre2/drivers/scsi'
make[2]: *** [_modsubdir_scsi] Erreur 2
make[2]: Leaving directory `/usr/src/kernel-sources-2.5.1-pre2/drivers'
make[1]: *** [_mod_drivers] Erreur 2
make[1]: Leaving directory `/usr/src/kernel-sources-2.5.1-pre2'
make: *** [stamp-build] Erreur 2
-------
Regards
Jean-Luc
In article <[email protected]>,
f5ibh <[email protected]> wrote:
>
>I've the following error :
Yes.
The next-generation block-layer support is starting to be merged into
the 2.5.x tree, and that breaks old drivers that haven't been updated to
the new locking.
In particular, there used to be _one_ lock for the whole IO system
("io_request_lock"), and these days it's a per-block-queue lock.
In many cases the fix is as simple as just replacing the
"io_request_lock" with "host->host_lock", but sometimes this is
complicated by the need to pass the right data structures down far
enough..
Many drivers have been converted (ie IDE, symbios, aic7xxx etc), but
many more have not (especially older SCSI drivers, in your case it's the
classic aha1542).
It will probably take some time until most drivers have been converted.
Tested patches are more than welcome,
Linus
Linus Torvalds writes:
> The next-generation block-layer support is starting to be merged into
> the 2.5.x tree, and that breaks old drivers that haven't been updated to
> the new locking.
>
> In particular, there used to be _one_ lock for the whole IO system
> ("io_request_lock"), and these days it's a per-block-queue lock.
>
> In many cases the fix is as simple as just replacing the
> "io_request_lock" with "host->host_lock", but sometimes this is
> complicated by the need to pass the right data structures down far
> enough..
Is there a description of the new block layer and its interface to
block device drivers somewhere? That would be helpful, since Ben
Herrenschmidt and I are going to have to convert several
powermac-specific drivers.
Thanks,
Paul.
In article <[email protected]> you wrote:
> In many cases the fix is as simple as just replacing the
> "io_request_lock" with "host->host_lock", but sometimes this is
> complicated by the need to pass the right data structures down far
> enough..
>
> Many drivers have been converted (ie IDE, symbios, aic7xxx etc), but
> many more have not (especially older SCSI drivers, in your case it's the
> classic aha1542).
>
> It will probably take some time until most drivers have been converted.
> Tested patches are more than welcome,
While we are at breaking scsi, would you take a patch to remove the
old-style (2.0) scsi error handling completly, forcing drivers still
using it to be fixed? Early 2.5 looks like a good time for that to me..
Christoph
--
Of course it doesn't work. We've performed a software upgrade.
I'm looking a purchasing a dozen computers with the Soyo SY-K7V Dragon
Plus motherboard. It has the KT266A chipset.
My question is Linux support for the onboard ethernet and sound. I've
googled all over and haven't come up with a definitive answer.
They are described as "VIA 10/100 Ethernet" and "CMI 8738 Audio chip".
Does anyone know the scoop?
Dax
On Tue, 2001-11-27 at 17:09, Christoph Hellwig wrote:
> While we are at breaking scsi, would you take a patch to remove the
> old-style (2.0) scsi error handling completly, forcing drivers still
> using it to be fixed? Early 2.5 looks like a good time for that to me..
Linus, please consider this at some early point. There isn't too much
using the scsi_obsolete gunk anyhow, but let's clean it up. Good idea.
Robert Love
Dax Kelson wrote:
> I'm looking a purchasing a dozen computers with the Soyo SY-K7V Dragon
> Plus motherboard. It has the KT266A chipset.
>
> My question is Linux support for the onboard ethernet and sound. I've
> googled all over and haven't come up with a definitive answer.
>
> They are described as "VIA 10/100 Ethernet" and "CMI 8738 Audio chip".
CMI 8738 has a driver in the current kernel.
As for VIA 10/100 Ethernet, I dunno... Maybe the via-rhine would work ?
Not sure, though...
Fran?ois
> Does anyone know the scoop?
>
> Dax
Dax Kelson wrote:
>
> I'm looking a purchasing a dozen computers with the Soyo SY-K7V Dragon
> Plus motherboard. It has the KT266A chipset.
>
> My question is Linux support for the onboard ethernet and sound. I've
> googled all over and haven't come up with a definitive answer.
>
> They are described as "VIA 10/100 Ethernet" and "CMI 8738 Audio chip".
what does 'lspci -vn' show for your system?
--
Jeff Garzik | Only so many songs can be sung
Building 1024 | with two lips, two lungs, and one tongue.
MandrakeSoft | - nomeansno
On Tue, 27 Nov 2001, Christoph Hellwig wrote:
>
> While we are at breaking scsi, would you take a patch to remove the
> old-style (2.0) scsi error handling completly, forcing drivers still
> using it to be fixed? Early 2.5 looks like a good time for that to me..
I agree, that sounds like a good thing, and as I consider the block layer
to be one of the major pushes for 2.5.x it makes perfect sense.
Linus
Given that block is in the transition period, would you consider a
schedule of when you are taking patchs ?
Regard,
Andre Hedrick
Linux Disk Certification Project Linux ATA Development
On Tue, 27 Nov 2001, Linus Torvalds wrote:
> In article <[email protected]>,
> f5ibh <[email protected]> wrote:
> >
> >I've the following error :
>
> Yes.
>
> The next-generation block-layer support is starting to be merged into
> the 2.5.x tree, and that breaks old drivers that haven't been updated to
> the new locking.
>
> In particular, there used to be _one_ lock for the whole IO system
> ("io_request_lock"), and these days it's a per-block-queue lock.
>
> In many cases the fix is as simple as just replacing the
> "io_request_lock" with "host->host_lock", but sometimes this is
> complicated by the need to pass the right data structures down far
> enough..
>
> Many drivers have been converted (ie IDE, symbios, aic7xxx etc), but
> many more have not (especially older SCSI drivers, in your case it's the
> classic aha1542).
>
> It will probably take some time until most drivers have been converted.
> Tested patches are more than welcome,
>
> Linus
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
On Wed, 28 Nov 2001, Paul Mackerras wrote:
>
> Is there a description of the new block layer and its interface to
> block device drivers somewhere? That would be helpful, since Ben
> Herrenschmidt and I are going to have to convert several
> powermac-specific drivers.
Jens has something written up, which he sent to me as an introduction to
the patch. I'll send that out unless he does a cleaned-up version, but I'd
actually prefer for him to do the sending. Jens?
Linus
On Tue, Nov 27, 2001 at 05:04:46PM -0800, Linus Torvalds wrote:
>
> On Wed, 28 Nov 2001, Paul Mackerras wrote:
> >
> > Is there a description of the new block layer and its interface to
> > block device drivers somewhere? That would be helpful, since Ben
> > Herrenschmidt and I are going to have to convert several
> > powermac-specific drivers.
>
> Jens has something written up, which he sent to me as an introduction to
> the patch. I'll send that out unless he does a cleaned-up version, but I'd
> actually prefer for him to do the sending. Jens?
>
> Linus
>
Linus/Jens,
I've just completed my review of submit_bio and the changes to
generic_make_request and I have some questions for whomever
can answer.
1. The changes made to submit_bh indicate I can now send long
chains of variable block size requests to the I/O layer similiar
to the capability of Windows 2000 and NetWare I/O subsystems.
2. The elevator layer is merging these requests, and making a
single sweep request for contiguous sector runs.
3. In theory, I should be able to support page cache capability
for NWFS and possibly NTFS in Linux the way these wierd non-Unix
OS's work.
4. This interface may **NOT** support non-block aligned requests
across all the drivers. I also need to be able to submit a
request chain 512-2048-512-1024-4096 where the first IO requested
may by on a non-block aligned boundry. i.e. Device is configured
for 1024 byte blocks, I start the request as 512 @ LBA 1 -> 1024 @ LBA 2,
etc. The code looks like it will work.
I would love to test this wonderful code and will hopefully this evening,
however, all the SCSI drivers appear to be broken, as well as the
3Ware. :-)
Please advise,
Jeff
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
"Jeff V. Merkey" wrote:
> 1. The changes made to submit_bh indicate I can now send long
> chains of variable block size requests to the I/O layer similiar
> to the capability of Windows 2000 and NetWare I/O subsystems.
I don't want to speak for either Jens or Linus, but from what Jens was
telling me months ago, and from my reading of Jens' earlier patches,
this seems to indeed be the case. I've been looking forward to sending
non-block-sized I/Os to and from a new filesystem I'm working on.
> 3. In theory, I should be able to support page cache capability
> for NWFS and possibly NTFS in Linux the way these wierd non-Unix
> OS's work.
If you are hacking on NTFS please make sure to base changes on
"ntfs-driver-tng" in the "linux-ntfs" sourceforge cvs. It is now, as of
the past week, completely modern to 2.4 vfs standards, and should
support all read-only features except attribute lists.
Regards,
Jeff
--
Jeff Garzik | Only so many songs can be sung
Building 1024 | with two lips, two lungs, and one tongue.
MandrakeSoft | - nomeansno
On Tue, Nov 27, 2001 at 08:45:24PM -0500, Jeff Garzik wrote:
> "Jeff V. Merkey" wrote:
> > 1. The changes made to submit_bh indicate I can now send long
> > chains of variable block size requests to the I/O layer similiar
> > to the capability of Windows 2000 and NetWare I/O subsystems.
>
> I don't want to speak for either Jens or Linus, but from what Jens was
> telling me months ago, and from my reading of Jens' earlier patches,
> this seems to indeed be the case. I've been looking forward to sending
> non-block-sized I/Os to and from a new filesystem I'm working on.
>
>
> > 3. In theory, I should be able to support page cache capability
> > for NWFS and possibly NTFS in Linux the way these wierd non-Unix
> > OS's work.
>
> If you are hacking on NTFS please make sure to base changes on
> "ntfs-driver-tng" in the "linux-ntfs" sourceforge cvs. It is now, as of
> the past week, completely modern to 2.4 vfs standards, and should
> support all read-only features except attribute lists.
I have a version that supports everything, including the journal,
and does not "barf" on writes. Unfortunately it's written in
"merkey code style" which is ThoseF_ckingLongAssNamesForEveryDamnFunction()
I was trained to write at Novell. Merging it into the sourceforge
version will be nasty. I have a few other projects and code I
am tardy in sending to Brazil :-). I think just letting Anton
review some of it may be a better approach. NWFS is first
up and needs to get finished and in so if I get hit by
a truck or killed by a deranged Novell employee who has
been recently laid off, someone can take maintain it.
:-)
Jeff
>
> Regards,
>
> Jeff
>
>
> --
> Jeff Garzik | Only so many songs can be sung
> Building 1024 | with two lips, two lungs, and one tongue.
> MandrakeSoft | - nomeansno
Oh yeah, I meant to ask: do we get 64-bit inode numbers and 64-bit block
numbers on x86 sometime in 2.5?
--
Jeff Garzik | Only so many songs can be sung
Building 1024 | with two lips, two lungs, and one tongue.
MandrakeSoft | - nomeansno
On Tue, 27 Nov 2001, Jeff Garzik wrote:
>
> Oh yeah, I meant to ask: do we get 64-bit inode numbers and 64-bit block
> numbers on x86 sometime in 2.5?
Well, the 64-bit sector number skeleton is already there in pre2..
We will probably _not_ get 64-bit page index numbers, though. I don't want
to make the page structure bigger/slower for very little gain. So the page
cache is probably going to be limited to about 44 bits (45+ if people
start doing large pages, which is probably worth it). So there would still
be partition/file limits on the order of 16-64 TB in the next few years.
(In a longer timeframe, assuming RAM keeps getting cheaper and cheaper,
and 64-bit computing starts hppening on PC's, a few years down the line we
can re-visit this - that particular transition is not going to be too
painful).
And yes, I realize that you can already build big arrays and use LVM etc
to make them be more than 16TB. I just do not think it's a problem yet,
and I'd rather cater to "normal" people than to peopel who can't bother
to partition their data at all.
Linus
On Tue, Nov 27 2001, Linus Torvalds wrote:
>
> On Wed, 28 Nov 2001, Paul Mackerras wrote:
> >
> > Is there a description of the new block layer and its interface to
> > block device drivers somewhere? That would be helpful, since Ben
> > Herrenschmidt and I are going to have to convert several
> > powermac-specific drivers.
>
> Jens has something written up, which he sent to me as an introduction to
> the patch. I'll send that out unless he does a cleaned-up version, but I'd
> actually prefer for him to do the sending. Jens?
No problem, I'll clean it up and send it out. I also planned on doing a
specific guide to converting drivers to exploit the new features.
--
Jens Axboe
Jeff Garzik wrote:
>
> Oh yeah, I meant to ask: do we get 64-bit inode numbers and 64-bit block
> numbers on x86 sometime in 2.5?
Well at least the patch from Jens makes this much easier...
On Tue, Nov 27 2001, Linus Torvalds wrote:
>
> On Wed, 28 Nov 2001, Paul Mackerras wrote:
> >
> > Is there a description of the new block layer and its interface to
> > block device drivers somewhere? That would be helpful, since Ben
> > Herrenschmidt and I are going to have to convert several
> > powermac-specific drivers.
>
> Jens has something written up, which he sent to me as an introduction to
> the patch. I'll send that out unless he does a cleaned-up version, but I'd
> actually prefer for him to do the sending. Jens?
Ok, here's the stuff I sent Linus + some extra comments on how exactly
it differs for stuff like drivers.
--->
o io_request_lock removal. it's completely gone, although lots of SCSI
drivers still need to be fixed (with it gone they break during compile
though, which is what we want). locking is per-queue (q->queue_lock),
the locking semantics are the same though (locking is still imposed by
the block layer, grabbing the lock before request_fn execution -- I want
to keep it this way, because it means we can leave lots of older drivers
alone and they will automagically be SMP safe still. clever drivers are
free to drop the queue lock whenever they want, of course).
SCSI mid layer is in an ok state, low level drivers that currently work
are sym, sym2, aic7xxx_old, and aic7xxx. IDE should be completely ok,
basically it relies on a new ide_lock to keep the same serialization
that it currently has. Keeping a couple of disks busy in IDE didn't show
this as being a big bottleneck at all, I guess we can later optimize
this if we really want... The "other" block drivers like DAC960,
cpqarray, and cciss are fixed too although at least DAC960 may have
multi-page bio quirks. Not a big deal at this point since only raw I/O
is the user that.
o main unit of I/O is the bio, not the buffer_head. this is the really
major change that requires updates to basically all block drivers. the
exception are drivers that only reference CURRENT and CURRENT->buffer
for transfers (hence forth called "old-style drivers", they will still
work as usual. Any driver that currently traverse segments in a request
must be updated.
A bio has no virtual mapping of the data at all. This basically forces
highmem support on drivers, at least it is no different than doing low
memory I/O which is what I think matters. Of course bounce limits can be
set by the driver. A bio looks like this:
struct bio {
sector_t bi_sector;
struct bio *bi_next; /* request queue link */
atomic_t bi_cnt; /* pin count */
kdev_t bi_dev; /* will be block device */
struct bio_vec_list *bi_io_vec;
unsigned long bi_flags; /* status, command, etc */
unsigned long bi_rw; /* bottom bits READ/WRITE,
* top bits priority
*/
int (*bi_end_io)(struct bio *bio, int nr_sectors);
void *bi_private;
void (*bi_destructor)(struct bio *); /* destructor */
};
Most of the above is self-explanatory, so I'll just brief you on some of
the differences from buffer_head. bi_io_vec is the actual io vector,
it's composed of these:
struct bio_vec {
struct page *bv_page;
unsigned int bv_len;
unsigned int bv_offset;
};
struct bio_vec_list {
unsigned int bvl_cnt; /* how may bio_vec's */
unsigned int bvl_idx; /* current index into bvl_vec */
unsigned int bvl_size; /* total size in bytes */
unsigned int bvl_max; /* max bvl_vecs we can hold, used
as index into pool */
struct bio_vec bvl_vec[0]; /* the iovec array */
};
The main reason for doing an array like this is that it enables easy
pre-clustering of pages for I/O submission. Stuff like read_cluster, raw
I/O, O_DIRECT etc can queue nice big portions of pages at once. And XFS
will really like it too. Also, the vm can pre-alloc a bio + veclist and
do it's own merging if it wants to (again XFS, I know they do this on
Irix (well sort of, pass it down differently, the merging at the vm
layer is what I'm comparing it to)). The I/O scheduler still uses
bi_next to singly link merges at that level.
The end_io callback now has a number of sectors argument, basically so
we don't have to loop around ending requests. No uptodate flag is passed
down, all that can be fitted in the bi_flags now. One bit of completion
info was never enough anyways.
bi_desctructor is only there because a bio can originate from different
sources. Normally it will be gotten with bio_alloc(int gfp_mask, int
nr_iovecs) and the default destructor frees the necessary things on I/O
completion. A bio may come from other sources though, or it may be a
cloned or even a copied bio (ie loop can clone or copy a bio (similar to
network skb stuff) with bio_clone/bio_copy), or maybe someone using just
kmalloc for allocating the bio.
Traversal of data segments in a request becomes even more hairy at this
point though, since a request still holds a list of bios which in turn
can hold a map of segments. I've always thought that drivers had to
initimate knowledge of how that works though, so I've moved this to
something like:
rq_for_each_bio(bio, rq)
bio_for_each_segment(bio_vec, bio, i)
/* bio_vec is now current segment */
which I think we should have had a long time ago. Then we are also free
to change the implementation details later without having to change
drivers again :-). I should mention that ll_rw_blk provides a
blk_rq_map_sg helper to map a struct request to a scatterlist, so we can
rip the scatterlist build out of a lot of drivers. They can simply do
nr_segments = blk_rq_map_sg(q, rq, scatterlist);
I can say a lot more about the bio stuff, but it's probably better if
you just start with commenting/flaming the above and we can go from
there :-). I'll move on to a few more things...
o head active queues are a thing of the past. We never should have
exposted the list implementation of requests to the drivers -- I've
changed this so drivers merely do
rq = elv_next_request(q);
which is the next request that the I/O scheduler thinks should be
handled. That enables us to easily mark requests as active or not, so
the elevator knows not to touch it. Then this head-active crap is
handled transparently instead. Also, it enables us to mark more than the
first request as untouchable, something we've needed for a while.
...
I'll finish off for now with a status... Basically the tree is stable
for IDE and the SCSI stuff listed, same with cpqarray etc. loop probably
needs a bit of fixing, but not much. O_DIRECT is currently a kludge, the
kio blocks/bh stuff is a MESS that needs resolving real soon! Raw I/O is
almost there too, we can pull lots of megs/sec with very little sys time
now.
...
<---
Now, I guess what others are mainly interested in is either a) design of
this stuff or b) what to look out for when converting drivers. Wrt a),
read the source :-). For b, read on.
If you are maintaing and old-style driver that just uses CURRENT and
ignores clustered requests, you should be alright and not need any
changes. Maybe you are dropping the io_request_lock from your request_fn
strategy, if so then just replace that with &q->queue_lock instead. The
generic layer will automatically handle clustered requests, multi-page
bios, etc for you. For a low performance driver or hardware that is PIO
driven or just doesn't support scatter-gather, you can stop reading now.
Lets start by looking at how we transferred data in the 2.4 kernels.
There are two structures you need to know about. The first is the
buffer_head, this is the I/O "atom". For a block driver, the relevant
parts of this struct is:
unsigned long b_rsector; /* where */
kdev_t b_rdev; /* who */
char *b_data; /* what */
struct buffer_head *b_reqnext; /* next buffer */
and that is basically it. b_rsector is offset from b_rdev which includes
minor stuff for partitions. b_data is the virtual mapping of the buffer
that wants to be read or written. actually, b_data is the virtual
mapping og b_page and a possible offset, but now we are already getting
into 2.4 + block-highmem or 2.5 land. b_reqnext is the next pointer to a
chain of buffer_heads.
Most real-hardware (ie not loop, raid, rd, etc) don't receive the
buffer_heads directly though. Instead they receive a list of struct
requests to handle. struct request contains lots of house keeping info,
the interesting part again for block drivers are:
unsigned long sector; /* like b_rsector */
unsigned long nr_sectors; /* total number of sectors */
unsigned long current_nr_sectors; /* "current" ditto of above */
unsigned int nr_segments;
char *buffer; /* like b_data */
struct buffer_head *bh; /* first bh in request */
struct buffer_head *bhtail; /* last bh in request */
So struct request ties a list of buffer_heads together for handing to
the driver. Performance hardware usually builds a scatter-gather list of
these chunks and sends it off to the driver. nr_sectors contains the
total size of all the buffer_heads linked to the request,
current_nr_sectors only the size of the currently first bh in the list
(ie rq->bh). When I/O is ended on a request, the buffer_heads are pealed
off the list as I/O on them completes.
Now lets take a look at how 2.5 does this differently. Struct
buffer_head now has no relevance for block drivers, it's purely a buffer
cache thing. Instead we deal with struct bio, which I explained the
layout of earlier in this mail. Struct request is basically the same,
except that now we link bio and biotail into it. For a sane block
driver, the request handling loop with bio will look something like:
do {
struct request *rq = elv_next_request(q);
struct scatterlist *sg;
struct my_driver_cmd *cmd;
int nr_segments;
/*
* no more to handle
*/
if (!rq)
break;
blkdev_dequeue_request(rq);
/*
* for queuing, otherwise you will probably just
* have a sg structure allocated at init time (see ide)
*/
sg = my_driver_alloc_sg(rq->nr_segments);
/*
* block layer helper to map a struct request into
* a scatter-gather list
*/
nr_segments = blk_rq_map_sg(q, rq, sg);
cmd = my_driver_alloc_cmd();
/*
* init hardware command with data direction and sg list
*/
my_driver_setup_cmd(cmd, rq->cmd, sg);
my_driver_queue_cmd(cmd);
} while (1);
You are free to loop through the request segments yourself using the
technique described higher up of course, however I recommend using the
blk_rq_map_sg helper to handle the grunt of the mapping work. That will
(probably :-) also help you out later if the internals are changed
again.
blk_rq_map_sg will look at several queue properties to handle stuff
like:
- cluster contigious segments into one, if wanted (QUEUE_FLAG_CLUSTER)
- don't allow a clustered segment to cross a 4GB mem boundary
- don't build bigger segments than q->max_segment_size
and probably more.
Other changes that may affect you:
New queue property settings:
blk_queue_bounce_limit(q, u64 dma_address)
Enable I/O to highmem pages, dma_address being your
limit. No highmem default.
blk_queue_max_sectors(q, max_sectors)
Maximum size request you can handle in units of 512 byte
sectors. 255 default.
blk_queue_max_segments(q, max_segments)
Maximum segments you can handle in a request. 128
default.
blk_queue_max_segment_size(q, max_seg_size)
Maximum size of a clustered segment, 64kB default.
New queue flags:
QUEUE_FLAG_NOSPLIT
can handle a bio with more than one segment. ll_rw_blk
will split bigger bio's for you if needed (not actually
implemented yet :-)
QUEUE_FLAG_CLUSTER
Explained above.
- struct request ->queue is no more. with the introduction of
elv_next_request, you are no longer supposed to handle looping
directly over the request list.
- end_that_request_first takes an additional number_of_sectors argument.
It used to handle always just the first buffer_head in a request, now
it will loop and handle as many sectors (on a bio-segment granularity)
as you want.
- bh->b_end_io is bio->bi_end_io, but you probably want to use
bio_endio(bio, uptodate, nr_sectors) instead.
- you can set max sector size, max segment size etc per queue now.
drivers that used to define their own merge functions to handle things
like this can now just use the blk_queue_* functions at blk_init_queue
time.
- you no longer have to map a {partition, sector offset} into the
correct absolute location anymore, this is done by the block layer. so
when you received a request ala this before:
rq->rq_dev = MKDEV(3, 5); /* /dev/hda5 */
rq->sector = 0; /* first sector on hda5 */
you will now see
rq->rq_dev = MKDEV(3, 0); /* /dev/hda */
rq->sector = 123128; /* offset from start of disk */
- As mentioned, there is no virtual mapping of a bio. For DMA, this is
not a problem as you probably never will need a virtual mapping.
Instead you want a bus mapping so you can ship it to the driver. For
PIO drivers (or drivers that need to revert to PIO transfer once in a
while (IDE for example)), where the CPU is doing the actual data
transfer for you, you do need a virtual mapping though. If you are
supporting highmem I/O, you need to use bio_kmap and bio_kmap_irq to
temporarily map a bio into the virtual address space. See how IDE
handles this with ide_map_buffer.
I've lost track of what else there is to explain, so I'll stop now. If
you have problems convering a driver or questions in general, fire away.
Suparana@IBM has written lots of stuff about bio as it was a WIP, see
http://lse.sourceforge.net/io/bionotes.txt
it may not be completely uptodate right now wrt multi-page bios etc, but
I know that is on its way :-)
--
Jens Axboe
On Tue, Nov 27, 2001 at 04:29:17PM -0800, Linus Torvalds wrote:
>
> On Tue, 27 Nov 2001, Christoph Hellwig wrote:
> >
> > While we are at breaking scsi, would you take a patch to remove the
> > old-style (2.0) scsi error handling completly, forcing drivers still
> > using it to be fixed? Early 2.5 looks like a good time for that to me..
>
> I agree, that sounds like a good thing, and as I consider the block layer
> to be one of the major pushes for 2.5.x it makes perfect sense.
Ok, here is the patch:
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/acorn/scsi/arxescsi.h linux/drivers/acorn/scsi/arxescsi.h
--- ../master/linux-2.5.1-pre2/drivers/acorn/scsi/arxescsi.h Fri May 12 20:21:20 2000
+++ linux/drivers/acorn/scsi/arxescsi.h Wed Nov 28 13:32:29 2001
@@ -61,7 +61,6 @@
eh_bus_reset_handler: fas216_eh_bus_reset, \
eh_device_reset_handler: fas216_eh_device_reset, \
eh_abort_handler: fas216_eh_abort, \
-use_new_eh_code: 1 \
}
#ifndef HOSTS_C
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/acorn/scsi/cumana_2.c linux/drivers/acorn/scsi/cumana_2.c
--- ../master/linux-2.5.1-pre2/drivers/acorn/scsi/cumana_2.c Sun Nov 4 18:31:57 2001
+++ linux/drivers/acorn/scsi/cumana_2.c Wed Nov 28 13:32:36 2001
@@ -575,7 +575,6 @@
eh_bus_reset_handler: fas216_eh_bus_reset,
eh_device_reset_handler: fas216_eh_device_reset,
eh_abort_handler: fas216_eh_abort,
- use_new_eh_code: 1
};
static int __init cumanascsi2_init(void)
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/acorn/scsi/eesox.c linux/drivers/acorn/scsi/eesox.c
--- ../master/linux-2.5.1-pre2/drivers/acorn/scsi/eesox.c Sun Nov 4 18:31:57 2001
+++ linux/drivers/acorn/scsi/eesox.c Wed Nov 28 13:32:41 2001
@@ -577,7 +577,6 @@
eh_bus_reset_handler: fas216_eh_bus_reset,
eh_device_reset_handler: fas216_eh_device_reset,
eh_abort_handler: fas216_eh_abort,
- use_new_eh_code: 1
};
static int __init eesox_init(void)
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/acorn/scsi/powertec.c linux/drivers/acorn/scsi/powertec.c
--- ../master/linux-2.5.1-pre2/drivers/acorn/scsi/powertec.c Sun Nov 4 18:31:57 2001
+++ linux/drivers/acorn/scsi/powertec.c Wed Nov 28 13:32:46 2001
@@ -477,7 +477,6 @@
eh_bus_reset_handler: fas216_eh_bus_reset,
eh_device_reset_handler: fas216_eh_device_reset,
eh_abort_handler: fas216_eh_abort,
- use_new_eh_code: 1
};
static int __init powertecscsi_init(void)
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/message/fusion/mptscsih.h linux/drivers/message/fusion/mptscsih.h
--- ../master/linux-2.5.1-pre2/drivers/message/fusion/mptscsih.h Sat Jul 7 02:03:11 2001
+++ linux/drivers/message/fusion/mptscsih.h Wed Nov 28 13:33:00 2001
@@ -198,7 +198,6 @@
cmd_per_lun: MPT_SCSI_CMD_PER_LUN, \
unchecked_isa_dma: 0, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 1 \
}
#else
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/net/fc/iph5526.c linux/drivers/net/fc/iph5526.c
--- ../master/linux-2.5.1-pre2/drivers/net/fc/iph5526.c Sun Sep 30 21:26:08 2001
+++ linux/drivers/net/fc/iph5526.c Wed Nov 28 13:33:15 2001
@@ -3802,7 +3802,6 @@
fi->host = host;
//host->max_id = MAX_SCSI_TARGETS;
host->max_id = 5;
- host->hostt->use_new_eh_code = 1;
host->this_id = tmpt->this_id;
pci_maddr = pci_resource_start(pdev, 0);
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/3w-xxxx.h linux/drivers/scsi/3w-xxxx.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/3w-xxxx.h Fri Nov 9 23:05:02 2001
+++ linux/drivers/scsi/3w-xxxx.h Wed Nov 28 13:24:40 2001
@@ -461,7 +461,6 @@
present : 0, \
unchecked_isa_dma : 0, \
use_clustering : ENABLE_CLUSTERING, \
- use_new_eh_code : 1, \
emulated : 1 \
}
#endif /* _3W_XXXX_H */
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/53c700.c linux/drivers/scsi/53c700.c
--- ../master/linux-2.5.1-pre2/drivers/scsi/53c700.c Thu Oct 11 18:43:29 2001
+++ linux/drivers/scsi/53c700.c Wed Nov 28 13:35:07 2001
@@ -267,7 +267,6 @@
tpnt->sg_tablesize = NCR_700_SG_SEGMENTS;
tpnt->cmd_per_lun = NCR_700_MAX_TAGS;
tpnt->use_clustering = DISABLE_CLUSTERING;
- tpnt->use_new_eh_code = 1;
tpnt->proc_info = NCR_700_proc_directory_info;
if(tpnt->name == NULL)
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/Makefile linux/drivers/scsi/Makefile
--- ../master/linux-2.5.1-pre2/drivers/scsi/Makefile Wed Nov 21 18:59:11 2001
+++ linux/drivers/scsi/Makefile Wed Nov 28 13:41:00 2001
@@ -133,11 +133,9 @@
obj-$(CONFIG_CHR_DEV_SG) += sg.o
list-multi := scsi_mod.o sd_mod.o sr_mod.o initio.o a100u2w.o cpqfc.o
-scsi_mod-objs := scsi.o hosts.o scsi_ioctl.o constants.o \
- scsicam.o scsi_proc.o scsi_error.o \
- scsi_obsolete.o scsi_queue.o scsi_lib.o \
- scsi_merge.o scsi_dma.o scsi_scan.o \
- scsi_syms.o
+scsi_mod-objs := scsi.o hosts.o scsi_ioctl.o constants.o scsicam.o \
+ scsi_proc.o scsi_error.o scsi_queue.o scsi_lib.o \
+ scsi_merge.o scsi_dma.o scsi_scan.o scsi_syms.o
sd_mod-objs := sd.o
sr_mod-objs := sr.o sr_ioctl.o sr_vendor.o
initio-objs := ini9100u.o i91uscsi.o
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/advansys.h linux/drivers/scsi/advansys.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/advansys.h Fri Jul 20 06:07:47 2001
+++ linux/drivers/scsi/advansys.h Wed Nov 28 13:28:01 2001
@@ -77,7 +77,6 @@
release: advansys_release, \
info: advansys_info, \
queuecommand: advansys_queuecommand, \
- use_new_eh_code: 1, \
eh_bus_reset_handler: advansys_reset, \
bios_param: advansys_biosparam, \
/* \
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/aha152x.h linux/drivers/scsi/aha152x.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/aha152x.h Fri Jul 20 06:08:07 2001
+++ linux/drivers/scsi/aha152x.h Wed Nov 28 13:28:14 2001
@@ -50,7 +50,6 @@
present: 0, \
unchecked_isa_dma: 0, \
use_clustering: DISABLE_CLUSTERING, \
- use_new_eh_code: 1 }
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/aha1542.h linux/drivers/scsi/aha1542.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/aha1542.h Fri Jul 20 06:08:11 2001
+++ linux/drivers/scsi/aha1542.h Wed Nov 28 13:28:36 2001
@@ -166,7 +166,6 @@
sg_tablesize: AHA1542_SCATTER, \
cmd_per_lun: AHA1542_CMDLUN, \
unchecked_isa_dma: 1, \
- use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 1}
-
+ use_clustering: ENABLE_CLUSTERING \
+}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/aic7xxx/aic7xxx_linux_host.h linux/drivers/scsi/aic7xxx/aic7xxx_linux_host.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/aic7xxx/aic7xxx_linux_host.h Wed Nov 28 13:20:39 2001
+++ linux/drivers/scsi/aic7xxx/aic7xxx_linux_host.h Wed Nov 28 13:34:17 2001
@@ -89,7 +89,6 @@
present: 0, /* number of 7xxx's present */\
unchecked_isa_dma: 0, /* no memory DMA restrictions */\
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 1, \
highmem_io: 1 \
}
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/aic7xxx_old/aic7xxx.h linux/drivers/scsi/aic7xxx_old/aic7xxx.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/aic7xxx_old/aic7xxx.h Sun Mar 4 23:30:18 2001
+++ linux/drivers/scsi/aic7xxx_old/aic7xxx.h Wed Nov 28 13:35:21 2001
@@ -55,7 +55,6 @@
present: 0, /* number of 7xxx's present */\
unchecked_isa_dma: 0, /* no memory DMA restrictions */\
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 0 \
}
extern int aic7xxx_queue(Scsi_Cmnd *, void (*)(Scsi_Cmnd *));
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/atp870u.h linux/drivers/scsi/atp870u.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/atp870u.h Fri Jul 20 06:08:56 2001
+++ linux/drivers/scsi/atp870u.h Wed Nov 28 13:28:45 2001
@@ -65,7 +65,6 @@
present: 0, /* number of 7xxx's present */\
unchecked_isa_dma: 0, /* no memory DMA restrictions */\
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 0 \
}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/cpqfcTS.h linux/drivers/scsi/cpqfcTS.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/cpqfcTS.h Sun Aug 5 22:12:41 2001
+++ linux/drivers/scsi/cpqfcTS.h Wed Nov 28 13:28:56 2001
@@ -38,7 +38,6 @@
present: 0, \
unchecked_isa_dma: 0, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 1 \
}
#endif /* CPQFCTS_H */
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/dc390.h linux/drivers/scsi/dc390.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/dc390.h Fri Jul 20 06:08:49 2001
+++ linux/drivers/scsi/dc390.h Wed Nov 28 13:29:27 2001
@@ -64,7 +64,6 @@
this_id: 7, \
sg_tablesize: SG_ALL, \
cmd_per_lun: 16, \
- NEW_EH \
unchecked_isa_dma: 0, \
use_clustering: DISABLE_CLUSTERING \
}
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/dpt_i2o.c linux/drivers/scsi/dpt_i2o.c
--- ../master/linux-2.5.1-pre2/drivers/scsi/dpt_i2o.c Fri Nov 9 23:05:06 2001
+++ linux/drivers/scsi/dpt_i2o.c Wed Nov 28 13:35:35 2001
@@ -178,7 +178,6 @@
adpt_hba* pHba;
adpt_init();
- sht->use_new_eh_code = 1;
PINFO("Detecting Adaptec I2O RAID controllers...\n");
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/dpti.h linux/drivers/scsi/dpti.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/dpti.h Fri Sep 7 18:28:38 2001
+++ linux/drivers/scsi/dpti.h Wed Nov 28 13:30:20 2001
@@ -95,7 +95,6 @@
sg_tablesize: 0, /* max scatter-gather cmds */\
cmd_per_lun: 256, /* cmds per lun (linked cmds) */\
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 1, \
proc_name: "dpt_i2o" /* this is the name of our proc node*/ \
}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/eata.h linux/drivers/scsi/eata.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/eata.h Sun May 20 02:43:06 2001
+++ linux/drivers/scsi/eata.h Wed Nov 28 13:27:54 2001
@@ -30,7 +30,6 @@
this_id: 7, \
unchecked_isa_dma: 1, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 1 /* Enable new error code */ \
}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/esp.h linux/drivers/scsi/esp.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/esp.h Mon Sep 18 22:40:17 2000
+++ linux/drivers/scsi/esp.h Wed Nov 28 13:26:42 2001
@@ -416,7 +416,6 @@
sg_tablesize: SG_ALL, \
cmd_per_lun: 1, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 0 \
}
/* For our interrupt engine. */
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/fcal.h linux/drivers/scsi/fcal.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/fcal.h Tue Mar 16 01:11:31 1999
+++ linux/drivers/scsi/fcal.h Wed Nov 28 13:30:05 2001
@@ -35,7 +35,6 @@
sg_tablesize: 1, \
cmd_per_lun: 1, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: FCP_SCSI_USE_NEW_EH_CODE, \
abort: fcp_old_abort, \
eh_abort_handler: fcp_scsi_abort, \
eh_device_reset_handler:fcp_scsi_dev_reset, \
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/gdth.h linux/drivers/scsi/gdth.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/gdth.h Fri Sep 7 18:28:37 2001
+++ linux/drivers/scsi/gdth.h Wed Nov 28 13:29:55 2001
@@ -1060,7 +1060,6 @@
present: 0, \
unchecked_isa_dma: 1, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 1 /* use new error code */ }
#elif LINUX_VERSION_CODE >= 0x02015F
int gdth_bios_param(Disk *,kdev_t,int *);
@@ -1092,7 +1091,6 @@
present: 0, \
unchecked_isa_dma: 1, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 1 /* use new error code */ }
#elif LINUX_VERSION_CODE >= 0x010300
int gdth_bios_param(Disk *,kdev_t,int *);
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/hosts.h linux/drivers/scsi/hosts.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/hosts.h Wed Nov 28 13:20:39 2001
+++ linux/drivers/scsi/hosts.h Wed Nov 28 13:43:27 2001
@@ -280,13 +280,6 @@
unsigned use_clustering:1;
/*
- * True if this driver uses the new error handling code. This flag is
- * really only temporary until all of the other drivers get converted
- * to use the new error handling code.
- */
- unsigned use_new_eh_code:1;
-
- /*
* True for emulated SCSI host adapters (e.g. ATAPI)
*/
unsigned emulated:1;
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/imm.h linux/drivers/scsi/imm.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/imm.h Sat Mar 3 03:38:38 2001
+++ linux/drivers/scsi/imm.h Wed Nov 28 13:30:14 2001
@@ -175,7 +175,6 @@
eh_device_reset_handler: NULL, \
eh_bus_reset_handler: imm_reset, \
eh_host_reset_handler: imm_reset, \
- use_new_eh_code: 1, \
bios_param: imm_biosparam, \
this_id: 7, \
sg_tablesize: SG_ALL, \
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/in2000.h linux/drivers/scsi/in2000.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/in2000.h Fri Jul 20 06:08:18 2001
+++ linux/drivers/scsi/in2000.h Wed Nov 28 13:27:50 2001
@@ -423,7 +423,6 @@
sg_tablesize: IN2000_SG, /* scatter-gather table size */ \
cmd_per_lun: IN2000_CPL, /* commands per lun */ \
use_clustering: DISABLE_CLUSTERING, /* ENABLE_CLUSTERING may speed things up */ \
- use_new_eh_code: 0 /* new error code - not using it yet */ \
}
#endif /* IN2000_H */
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/ini9100u.h linux/drivers/scsi/ini9100u.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/ini9100u.h Thu Oct 11 18:43:30 2001
+++ linux/drivers/scsi/ini9100u.h Wed Nov 28 13:30:29 2001
@@ -115,7 +115,6 @@
present: 0, \
unchecked_isa_dma: 0, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 0 \
}
#define VIRT_TO_BUS(i) (unsigned int) virt_to_bus((void *)(i))
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/inia100.h linux/drivers/scsi/inia100.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/inia100.h Thu Oct 11 18:43:30 2001
+++ linux/drivers/scsi/inia100.h Wed Nov 28 13:30:35 2001
@@ -110,7 +110,6 @@
present: 0, \
unchecked_isa_dma: 0, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 0 \
}
#define VIRT_TO_BUS(i) (unsigned int) virt_to_bus((void *)(i))
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/ips.h linux/drivers/scsi/ips.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/ips.h Sun Sep 30 21:26:08 2001
+++ linux/drivers/scsi/ips.h Wed Nov 28 13:24:56 2001
@@ -472,7 +472,6 @@
present : 0, \
unchecked_isa_dma : 0, \
use_clustering : ENABLE_CLUSTERING, \
- use_new_eh_code : 1 \
}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/mac53c94.h linux/drivers/scsi/mac53c94.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/mac53c94.h Tue Sep 19 17:31:53 2000
+++ linux/drivers/scsi/mac53c94.h Wed Nov 28 13:30:51 2001
@@ -28,7 +28,6 @@
sg_tablesize: SG_ALL, \
cmd_per_lun: 1, \
use_clustering: DISABLE_CLUSTERING, \
- use_new_eh_code: 1, \
}
/*
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/mac_esp.h linux/drivers/scsi/mac_esp.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/mac_esp.h Mon Jul 10 07:51:43 2000
+++ linux/drivers/scsi/mac_esp.h Wed Nov 28 13:27:37 2001
@@ -35,7 +35,6 @@
sg_tablesize: SG_ALL, \
cmd_per_lun: 1, \
use_clustering: DISABLE_CLUSTERING, \
- use_new_eh_code: 0 }
#endif /* MAC_ESP_H */
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/mesh.h linux/drivers/scsi/mesh.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/mesh.h Tue Sep 19 17:31:53 2000
+++ linux/drivers/scsi/mesh.h Wed Nov 28 13:27:45 2001
@@ -28,7 +28,6 @@
sg_tablesize: SG_ALL, \
cmd_per_lun: 2, \
use_clustering: DISABLE_CLUSTERING, \
- use_new_eh_code: 1, \
}
/*
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/pci2000.h linux/drivers/scsi/pci2000.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/pci2000.h Fri Jul 20 06:07:53 2001
+++ linux/drivers/scsi/pci2000.h Wed Nov 28 13:27:22 2001
@@ -218,7 +218,6 @@
present: 0, \
unchecked_isa_dma:0, \
use_clustering: DISABLE_CLUSTERING, \
- use_new_eh_code:0 \
}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/pci2220i.h linux/drivers/scsi/pci2220i.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/pci2220i.h Fri Jul 20 06:07:54 2001
+++ linux/drivers/scsi/pci2220i.h Wed Nov 28 13:27:28 2001
@@ -56,6 +56,5 @@
present: 0, \
unchecked_isa_dma: 0, \
use_clustering: DISABLE_CLUSTERING, \
- use_new_eh_code: 0 \
}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/pcmcia/nsp_cs.c linux/drivers/scsi/pcmcia/nsp_cs.c
--- ../master/linux-2.5.1-pre2/drivers/scsi/pcmcia/nsp_cs.c Thu Oct 11 18:04:57 2001
+++ linux/drivers/scsi/pcmcia/nsp_cs.c Wed Nov 28 13:35:46 2001
@@ -149,7 +149,6 @@
/* present: 0,*/
/* unchecked_isa_dma: 0,*/
use_clustering: DISABLE_CLUSTERING,
- use_new_eh_code: 0,
/* emulated: 0,*/
};
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/pluto.h linux/drivers/scsi/pluto.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/pluto.h Thu Jan 27 21:53:32 2000
+++ linux/drivers/scsi/pluto.h Wed Nov 28 13:27:08 2001
@@ -53,7 +53,6 @@
sg_tablesize: 1, \
cmd_per_lun: 1, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: FCP_SCSI_USE_NEW_EH_CODE, \
abort: fcp_old_abort, \
eh_abort_handler: fcp_scsi_abort, \
eh_device_reset_handler:fcp_scsi_dev_reset, \
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/ppa.h linux/drivers/scsi/ppa.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/ppa.h Sat Mar 3 03:38:39 2001
+++ linux/drivers/scsi/ppa.h Wed Nov 28 13:26:46 2001
@@ -183,7 +183,6 @@
eh_device_reset_handler: NULL, \
eh_bus_reset_handler: ppa_reset, \
eh_host_reset_handler: ppa_reset, \
- use_new_eh_code: 1, \
bios_param: ppa_biosparam, \
this_id: -1, \
sg_tablesize: SG_ALL, \
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/qla1280.h linux/drivers/scsi/qla1280.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/qla1280.h Mon Sep 17 22:16:31 2001
+++ linux/drivers/scsi/qla1280.h Wed Nov 28 13:27:15 2001
@@ -1726,7 +1726,6 @@
present: 0, /* number of 7xxx's present */\
unchecked_isa_dma: 0, /* no memory DMA restrictions */\
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 0, \
emulated: 0 \
}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/qlogicfc.c linux/drivers/scsi/qlogicfc.c
--- ../master/linux-2.5.1-pre2/drivers/scsi/qlogicfc.c Wed Nov 28 13:20:39 2001
+++ linux/drivers/scsi/qlogicfc.c Wed Nov 28 13:36:21 2001
@@ -734,7 +734,6 @@
scsi_set_pci_device(host, pdev);
host->max_id = QLOGICFC_MAX_ID + 1;
host->max_lun = QLOGICFC_MAX_LUN;
- host->hostt->use_new_eh_code = 1;
hostdata = (struct isp2x00_hostdata *) host->hostdata;
memset(hostdata, 0, sizeof(struct isp2x00_hostdata));
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/qlogicpti.h linux/drivers/scsi/qlogicpti.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/qlogicpti.h Tue Dec 21 07:06:42 1999
+++ linux/drivers/scsi/qlogicpti.h Wed Nov 28 13:27:03 2001
@@ -524,7 +524,6 @@
sg_tablesize: QLOGICPTI_MAX_SG(QLOGICPTI_REQ_QUEUE_LEN), \
cmd_per_lun: 1, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 0 \
}
/* For our interrupt engine. */
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/scsi.c linux/drivers/scsi/scsi.c
--- ../master/linux-2.5.1-pre2/drivers/scsi/scsi.c Wed Nov 28 13:20:39 2001
+++ linux/drivers/scsi/scsi.c Wed Nov 28 13:39:40 2001
@@ -151,14 +151,6 @@
void scsi_build_commandblocks(Scsi_Device * SDpnt);
/*
- * These are the interface to the old error handling code. It should go away
- * someday soon.
- */
-extern void scsi_old_done(Scsi_Cmnd * SCpnt);
-extern void scsi_old_times_out(Scsi_Cmnd * SCpnt);
-
-
-/*
* Function: scsi_initialize_queue()
*
* Purpose: Selects queue handler function for a device.
@@ -670,12 +662,8 @@
mdelay(1 + 999 / HZ);
host->resetting = 0;
}
- if (host->hostt->use_new_eh_code) {
- scsi_add_timer(SCpnt, SCpnt->timeout_per_command, scsi_times_out);
- } else {
- scsi_add_timer(SCpnt, SCpnt->timeout_per_command,
- scsi_old_times_out);
- }
+
+ scsi_add_timer(SCpnt, SCpnt->timeout_per_command, scsi_times_out);
/*
* We will use a queued command if possible, otherwise we will emulate the
@@ -692,50 +680,27 @@
SCSI_LOG_MLQUEUE(3, printk("queuecommand : routine at %p\n",
host->hostt->queuecommand));
/*
- * Use the old error handling code if we haven't converted the driver
- * to use the new one yet. Note - only the new queuecommand variant
- * passes a meaningful return value.
+ * Before we queue this command, check if the command
+ * length exceeds what the host adapter can handle.
*/
- if (host->hostt->use_new_eh_code) {
- /*
- * Before we queue this command, check if the command
- * length exceeds what the host adapter can handle.
- */
- if (CDB_SIZE(SCpnt) <= SCpnt->host->max_cmd_len) {
- spin_lock_irqsave(&host->host_lock, flags);
- rtn = host->hostt->queuecommand(SCpnt, scsi_done);
- spin_unlock_irqrestore(&host->host_lock, flags);
- if (rtn != 0) {
- scsi_delete_timer(SCpnt);
- scsi_mlqueue_insert(SCpnt, SCSI_MLQUEUE_HOST_BUSY);
- SCSI_LOG_MLQUEUE(3, printk("queuecommand : request rejected\n"));
- }
- } else {
- SCSI_LOG_MLQUEUE(3, printk("queuecommand : command too long.\n"));
- SCpnt->result = (DID_ABORT << 16);
- spin_lock_irqsave(&host->host_lock, flags);
- scsi_done(SCpnt);
- spin_unlock_irqrestore(&host->host_lock, flags);
- rtn = 1;
-
+ if (CDB_SIZE(SCpnt) <= SCpnt->host->max_cmd_len) {
+ spin_lock_irqsave(&host->host_lock, flags);
+ rtn = host->hostt->queuecommand(SCpnt, scsi_done);
+ spin_unlock_irqrestore(&host->host_lock, flags);
+ if (rtn != 0) {
+ scsi_delete_timer(SCpnt);
+ scsi_mlqueue_insert(SCpnt, SCSI_MLQUEUE_HOST_BUSY);
+ SCSI_LOG_MLQUEUE(3,
+ printk("queuecommand : request rejected\n"));
}
} else {
- /*
- * Before we queue this command, check if the command
- * length exceeds what the host adapter can handle.
- */
- if (CDB_SIZE(SCpnt) <= SCpnt->host->max_cmd_len) {
- spin_lock_irqsave(&host->host_lock, flags);
- host->hostt->queuecommand(SCpnt, scsi_old_done);
- spin_unlock_irqrestore(&host->host_lock, flags);
- } else {
- SCSI_LOG_MLQUEUE(3, printk("queuecommand : command too long.\n"));
- SCpnt->result = (DID_ABORT << 16);
- spin_lock_irqsave(&host->host_lock, flags);
- scsi_old_done(SCpnt);
- spin_unlock_irqrestore(&host->host_lock, flags);
- rtn = 1;
- }
+ SCSI_LOG_MLQUEUE(3,
+ printk("queuecommand : command too long.\n"));
+ SCpnt->result = (DID_ABORT << 16);
+ spin_lock_irqsave(&host->host_lock, flags);
+ scsi_done(SCpnt);
+ spin_unlock_irqrestore(&host->host_lock, flags);
+ rtn = 1;
}
} else {
int temp;
@@ -755,11 +720,7 @@
host->host_no, temp, host->hostt->command);
spin_lock_irqsave(&host->host_lock, flags);
#endif
- if (host->hostt->use_new_eh_code) {
- scsi_done(SCpnt);
- } else {
- scsi_old_done(SCpnt);
- }
+ scsi_done(SCpnt);
spin_unlock_irqrestore(&host->host_lock, flags);
}
SCSI_LOG_MLQUEUE(3, printk("leaving scsi_dispatch_cmnd()\n"));
@@ -1886,19 +1847,12 @@
/* The detect routine must carefully spinunlock/spinlock if
it enables interrupts, since all interrupt handlers do
- spinlock as well.
- All lame drivers are going to fail due to the following
- spinlock. For the time beeing let's use it only for drivers
- using the new scsi code. NOTE: the detect routine could
- redefine the value tpnt->use_new_eh_code. (DB, 13 May 1998) */
+ spinlock as well. */
/*
* detect should do its own locking
*/
- if (tpnt->use_new_eh_code) {
- tpnt->present = tpnt->detect(tpnt);
- } else
- tpnt->present = tpnt->detect(tpnt);
+ tpnt->present = tpnt->detect(tpnt);
if (tpnt->present) {
if (pcount == next_scsi_host) {
@@ -1932,7 +1886,7 @@
* handle error correction.
*/
for (shpnt = scsi_hostlist; shpnt; shpnt = shpnt->next) {
- if (shpnt->hostt == tpnt && shpnt->hostt->use_new_eh_code) {
+ if (shpnt->hostt == tpnt) {
DECLARE_MUTEX_LOCKED(sem);
shpnt->eh_notify = &sem;
@@ -2140,7 +2094,6 @@
*/
for (shpnt = scsi_hostlist; shpnt; shpnt = shpnt->next) {
if (shpnt->hostt == tpnt
- && shpnt->hostt->use_new_eh_code
&& shpnt->ehandler != NULL) {
DECLARE_MUTEX_LOCKED(sem);
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/scsi.h linux/drivers/scsi/scsi.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/scsi.h Wed Nov 28 13:20:39 2001
+++ linux/drivers/scsi/scsi.h Wed Nov 28 13:43:27 2001
@@ -387,12 +387,6 @@
#define SYNC_RESET 0x40
/*
- * This is the crap from the old error handling code. We have it in a special
- * place so that we can more easily delete it later on.
- */
-#include "scsi_obsolete.h"
-
-/*
* Add some typedefs so that we can prototyope a bunch of the functions.
*/
typedef struct scsi_device Scsi_Device;
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/scsi_debug.h linux/drivers/scsi/scsi_debug.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/scsi_debug.h Mon Dec 11 22:19:52 2000
+++ linux/drivers/scsi/scsi_debug.h Wed Nov 28 13:26:53 2001
@@ -37,6 +37,5 @@
cmd_per_lun: 3, \
unchecked_isa_dma: 0, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 1, \
}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/scsi_obsolete.c linux/drivers/scsi/scsi_obsolete.c
--- ../master/linux-2.5.1-pre2/drivers/scsi/scsi_obsolete.c Wed Nov 28 13:20:39 2001
+++ linux/drivers/scsi/scsi_obsolete.c Thu Jan 1 01:00:00 1970
@@ -1,1119 +0,0 @@
-/*
- * scsi_obsolete.c Copyright (C) 1992 Drew Eckhardt
- * Copyright (C) 1993, 1994, 1995 Eric Youngdale
- *
- * generic mid-level SCSI driver
- * Initial versions: Drew Eckhardt
- * Subsequent revisions: Eric Youngdale
- *
- * <[email protected]>
- *
- * Bug correction thanks go to :
- * Rik Faith <[email protected]>
- * Tommy Thorn <tthorn>
- * Thomas Wuensche <[email protected]>
- *
- * Modified by Eric Youngdale [email protected] to
- * add scatter-gather, multiple outstanding request, and other
- * enhancements.
- *
- * Native multichannel, wide scsi, /proc/scsi and hot plugging
- * support added by Michael Neuffer <[email protected]>
- *
- * Major improvements to the timeout, abort, and reset processing,
- * as well as performance modifications for large queue depths by
- * Leonard N. Zubkoff <[email protected]>
- *
- * Improved compatibility with 2.0 behaviour by Manfred Spraul
- * <[email protected]>
- */
-
-/*
- *#########################################################################
- *#########################################################################
- *#########################################################################
- *#########################################################################
- * NOTE - NOTE - NOTE - NOTE - NOTE - NOTE - NOTE
- *
- *#########################################################################
- *#########################################################################
- *#########################################################################
- *#########################################################################
- *
- * This file contains the 'old' scsi error handling. It is only present
- * while the new error handling code is being debugged, and while the low
- * level drivers are being converted to use the new code. Once the last
- * driver uses the new code this *ENTIRE* file will be nuked.
- */
-
-#define __NO_VERSION__
-#include <linux/module.h>
-
-#include <linux/sched.h>
-#include <linux/timer.h>
-#include <linux/string.h>
-#include <linux/slab.h>
-#include <linux/ioport.h>
-#include <linux/kernel.h>
-#include <linux/stat.h>
-#include <linux/blk.h>
-#include <linux/interrupt.h>
-#include <linux/delay.h>
-
-#include <asm/system.h>
-#include <asm/irq.h>
-#include <asm/dma.h>
-
-#include "scsi.h"
-#include "hosts.h"
-#include "constants.h"
-
-#undef USE_STATIC_SCSI_MEMORY
-
-/*
- static const char RCSid[] = "$Header: /mnt/ide/home/eric/CVSROOT/linux/drivers/scsi/scsi_obsolete.c,v 1.1 1997/05/18 23:27:21 eric Exp $";
- */
-
-
-#define INTERNAL_ERROR (panic ("Internal error in file %s, line %d.\n", __FILE__, __LINE__))
-
-
-static int scsi_abort(Scsi_Cmnd *, int code);
-static int scsi_reset(Scsi_Cmnd *, unsigned int);
-
-extern void scsi_old_done(Scsi_Cmnd * SCpnt);
-int update_timeout(Scsi_Cmnd *, int);
-extern void scsi_old_times_out(Scsi_Cmnd * SCpnt);
-
-extern int scsi_dispatch_cmd(Scsi_Cmnd * SCpnt);
-
-#define SCSI_BLOCK(HOST) (HOST->can_queue && HOST->host_busy >= HOST->can_queue)
-
-static unsigned char generic_sense[6] =
-{REQUEST_SENSE, 0, 0, 0, 255, 0};
-
-/*
- * This is the number of clock ticks we should wait before we time out
- * and abort the command. This is for where the scsi.c module generates
- * the command, not where it originates from a higher level, in which
- * case the timeout is specified there.
- *
- * ABORT_TIMEOUT and RESET_TIMEOUT are the timeouts for RESET and ABORT
- * respectively.
- */
-
-#ifdef DEBUG_TIMEOUT
-static void scsi_dump_status(void);
-#endif
-
-
-#ifdef DEBUG
-#define SCSI_TIMEOUT (5*HZ)
-#else
-#define SCSI_TIMEOUT (2*HZ)
-#endif
-
-#ifdef DEBUG
-#define SENSE_TIMEOUT SCSI_TIMEOUT
-#define ABORT_TIMEOUT SCSI_TIMEOUT
-#define RESET_TIMEOUT SCSI_TIMEOUT
-#else
-#define SENSE_TIMEOUT (5*HZ/10)
-#define RESET_TIMEOUT (5*HZ/10)
-#define ABORT_TIMEOUT (5*HZ/10)
-#endif
-
-
-/* Do not call reset on error if we just did a reset within 15 sec. */
-#define MIN_RESET_PERIOD (15*HZ)
-
-
-
-/*
- * Flag bits for the internal_timeout array
- */
-#define IN_ABORT 1
-#define IN_RESET 2
-#define IN_RESET2 4
-#define IN_RESET3 8
-
-/*
- * This is our time out function, called when the timer expires for a
- * given host adapter. It will attempt to abort the currently executing
- * command, that failing perform a kernel panic.
- */
-
-void scsi_old_times_out(Scsi_Cmnd * SCpnt)
-{
- struct Scsi_Host *host = SCpnt->host;
- unsigned long flags;
-
- spin_lock_irqsave(&host->host_lock, flags);
-
- /* Set the serial_number_at_timeout to the current serial_number */
- SCpnt->serial_number_at_timeout = SCpnt->serial_number;
-
- switch (SCpnt->internal_timeout & (IN_ABORT | IN_RESET | IN_RESET2 | IN_RESET3)) {
- case NORMAL_TIMEOUT:
- {
-#ifdef DEBUG_TIMEOUT
- scsi_dump_status();
-#endif
- }
-
- if (!scsi_abort(SCpnt, DID_TIME_OUT))
- break;
- case IN_ABORT:
- printk("SCSI host %d abort (pid %ld) timed out - resetting\n",
- host->host_no, SCpnt->pid);
- if (!scsi_reset(SCpnt, SCSI_RESET_ASYNCHRONOUS))
- break;
- case IN_RESET:
- case (IN_ABORT | IN_RESET):
- /* This might be controversial, but if there is a bus hang,
- * you might conceivably want the machine up and running
- * esp if you have an ide disk.
- */
- printk("SCSI host %d channel %d reset (pid %ld) timed out - "
- "trying harder\n",
- host->host_no, SCpnt->channel, SCpnt->pid);
- SCpnt->internal_timeout &= ~IN_RESET;
- SCpnt->internal_timeout |= IN_RESET2;
- scsi_reset(SCpnt,
- SCSI_RESET_ASYNCHRONOUS | SCSI_RESET_SUGGEST_BUS_RESET);
- break;
- case IN_RESET2:
- case (IN_ABORT | IN_RESET2):
- /* Obviously the bus reset didn't work.
- * Let's try even harder and call for an HBA reset.
- * Maybe the HBA itself crashed and this will shake it loose.
- */
- printk("SCSI host %d reset (pid %ld) timed out - trying to shake it loose\n",
- host->host_no, SCpnt->pid);
- SCpnt->internal_timeout &= ~(IN_RESET | IN_RESET2);
- SCpnt->internal_timeout |= IN_RESET3;
- scsi_reset(SCpnt,
- SCSI_RESET_ASYNCHRONOUS | SCSI_RESET_SUGGEST_HOST_RESET);
- break;
-
- default:
- printk("SCSI host %d reset (pid %ld) timed out again -\n",
- host->host_no, SCpnt->pid);
- printk("probably an unrecoverable SCSI bus or device hang.\n");
- break;
-
- }
- spin_unlock_irqrestore(&host->host_lock, flags);
-
-}
-
-/*
- * From what I can find in scsi_obsolete.c, this function is only called
- * by scsi_old_done and scsi_reset. Both of these functions run with the
- * host_lock already held, so we need do nothing here about grabbing
- * any locks.
- */
-static void scsi_request_sense(Scsi_Cmnd * SCpnt)
-{
- SCpnt->flags |= WAS_SENSE | ASKED_FOR_SENSE;
- update_timeout(SCpnt, SENSE_TIMEOUT);
-
- memcpy((void *) SCpnt->cmnd, (void *) generic_sense,
- sizeof(generic_sense));
- memset((void *) SCpnt->sense_buffer, 0,
- sizeof(SCpnt->sense_buffer));
-
- if (SCpnt->device->scsi_level <= SCSI_2)
- SCpnt->cmnd[1] = SCpnt->lun << 5;
- SCpnt->cmnd[4] = sizeof(SCpnt->sense_buffer);
-
- SCpnt->request_buffer = &SCpnt->sense_buffer;
- SCpnt->request_bufflen = sizeof(SCpnt->sense_buffer);
- SCpnt->use_sg = 0;
- SCpnt->cmd_len = COMMAND_SIZE(SCpnt->cmnd[0]);
- SCpnt->result = 0;
- SCpnt->sc_data_direction = SCSI_DATA_READ;
-
- /*
- * Ugly, ugly. The newer interfaces all assume that the lock
- * isn't held. Mustn't disappoint, or we deadlock the system.
- */
- spin_unlock_irq(&SCpnt->host->host_lock);
- scsi_dispatch_cmd(SCpnt);
- spin_lock_irq(&SCpnt->host->host_lock);
-}
-
-
-
-
-static int check_sense(Scsi_Cmnd * SCpnt)
-{
- /* If there is no sense information, request it. If we have already
- * requested it, there is no point in asking again - the firmware must
- * be confused.
- */
- if (((SCpnt->sense_buffer[0] & 0x70) >> 4) != 7) {
- if (!(SCpnt->flags & ASKED_FOR_SENSE))
- return SUGGEST_SENSE;
- else
- return SUGGEST_RETRY;
- }
- SCpnt->flags &= ~ASKED_FOR_SENSE;
-
-#ifdef DEBUG_INIT
- printk("scsi%d, channel%d : ", SCpnt->host->host_no, SCpnt->channel);
- print_sense("", SCpnt);
- printk("\n");
-#endif
- if (SCpnt->sense_buffer[2] & 0xe0)
- return SUGGEST_ABORT;
-
- switch (SCpnt->sense_buffer[2] & 0xf) {
- case NO_SENSE:
- return 0;
- case RECOVERED_ERROR:
- return SUGGEST_IS_OK;
-
- case ABORTED_COMMAND:
- return SUGGEST_RETRY;
- case NOT_READY:
- case UNIT_ATTENTION:
- /*
- * If we are expecting a CC/UA because of a bus reset that we
- * performed, treat this just as a retry. Otherwise this is
- * information that we should pass up to the upper-level driver
- * so that we can deal with it there.
- */
- if (SCpnt->device->expecting_cc_ua) {
- SCpnt->device->expecting_cc_ua = 0;
- return SUGGEST_RETRY;
- }
- return SUGGEST_ABORT;
-
- /* these three are not supported */
- case COPY_ABORTED:
- case VOLUME_OVERFLOW:
- case MISCOMPARE:
-
- case MEDIUM_ERROR:
- return SUGGEST_REMAP;
- case BLANK_CHECK:
- case DATA_PROTECT:
- case HARDWARE_ERROR:
- case ILLEGAL_REQUEST:
- default:
- return SUGGEST_ABORT;
- }
-}
-
-/* This function is the mid-level interrupt routine, which decides how
- * to handle error conditions. Each invocation of this function must
- * do one and *only* one of the following:
- *
- * (1) Call last_cmnd[host].done. This is done for fatal errors and
- * normal completion, and indicates that the handling for this
- * request is complete.
- * (2) Call internal_cmnd to requeue the command. This will result in
- * scsi_done being called again when the retry is complete.
- * (3) Call scsi_request_sense. This asks the host adapter/drive for
- * more information about the error condition. When the information
- * is available, scsi_done will be called again.
- * (4) Call reset(). This is sort of a last resort, and the idea is that
- * this may kick things loose and get the drive working again. reset()
- * automatically calls scsi_request_sense, and thus scsi_done will be
- * called again once the reset is complete.
- *
- * If none of the above actions are taken, the drive in question
- * will hang. If more than one of the above actions are taken by
- * scsi_done, then unpredictable behavior will result.
- */
-void scsi_old_done(Scsi_Cmnd * SCpnt)
-{
- int status = 0;
- int exit = 0;
- int checked;
- int oldto;
- struct Scsi_Host *host = SCpnt->host;
- Scsi_Device * device = SCpnt->device;
- int result = SCpnt->result;
- SCpnt->serial_number = 0;
- SCpnt->serial_number_at_timeout = 0;
- oldto = update_timeout(SCpnt, 0);
-
-#ifdef DEBUG_TIMEOUT
- if (result)
- printk("Non-zero result in scsi_done %x %d:%d\n",
- result, SCpnt->target, SCpnt->lun);
-#endif
-
- /* If we requested an abort, (and we got it) then fix up the return
- * status to say why
- */
- if (host_byte(result) == DID_ABORT && SCpnt->abort_reason)
- SCpnt->result = result = (result & 0xff00ffff) |
- (SCpnt->abort_reason << 16);
-
-
-#define CMD_FINISHED 0
-#define MAYREDO 1
-#define REDO 3
-#define PENDING 4
-
-#ifdef DEBUG
- printk("In scsi_done(host = %d, result = %06x)\n", host->host_no, result);
-#endif
-
- if (SCpnt->flags & SYNC_RESET) {
- /*
- * The behaviou of scsi_reset(SYNC) was changed in 2.1.? .
- * The scsi mid-layer does a REDO after every sync reset, the driver
- * must not do that any more. In order to prevent old drivers from
- * crashing, all scsi_done() calls during sync resets are ignored.
- */
- printk("scsi%d: device driver called scsi_done() "
- "for a synchronous reset.\n", SCpnt->host->host_no);
- return;
- }
- if (SCpnt->flags & WAS_SENSE) {
- SCpnt->use_sg = SCpnt->old_use_sg;
- SCpnt->cmd_len = SCpnt->old_cmd_len;
- SCpnt->sc_data_direction = SCpnt->sc_old_data_direction;
- SCpnt->underflow = SCpnt->old_underflow;
- }
- switch (host_byte(result)) {
- case DID_OK:
- if (status_byte(result) && (SCpnt->flags & WAS_SENSE))
- /* Failed to obtain sense information */
- {
- SCpnt->flags &= ~WAS_SENSE;
-#if 0 /* This cannot possibly be correct. */
- SCpnt->internal_timeout &= ~SENSE_TIMEOUT;
-#endif
-
- if (!(SCpnt->flags & WAS_RESET)) {
- printk("scsi%d : channel %d target %d lun %d request sense"
- " failed, performing reset.\n",
- SCpnt->host->host_no, SCpnt->channel, SCpnt->target,
- SCpnt->lun);
- scsi_reset(SCpnt, SCSI_RESET_SYNCHRONOUS);
- status = REDO;
- break;
- } else {
- exit = (DRIVER_HARD | SUGGEST_ABORT);
- status = CMD_FINISHED;
- }
- } else
- switch (msg_byte(result)) {
- case COMMAND_COMPLETE:
- switch (status_byte(result)) {
- case GOOD:
- if (SCpnt->flags & WAS_SENSE) {
-#ifdef DEBUG
- printk("In scsi_done, GOOD status, COMMAND COMPLETE, "
- "parsing sense information.\n");
-#endif
- SCpnt->flags &= ~WAS_SENSE;
-#if 0 /* This cannot possibly be correct. */
- SCpnt->internal_timeout &= ~SENSE_TIMEOUT;
-#endif
-
- switch (checked = check_sense(SCpnt)) {
- case SUGGEST_SENSE:
- case 0:
-#ifdef DEBUG
- printk("NO SENSE. status = REDO\n");
-#endif
- update_timeout(SCpnt, oldto);
- status = REDO;
- break;
- case SUGGEST_IS_OK:
- break;
- case SUGGEST_REMAP:
-#ifdef DEBUG
- printk("SENSE SUGGEST REMAP - status = CMD_FINISHED\n");
-#endif
- status = CMD_FINISHED;
- exit = DRIVER_SENSE | SUGGEST_ABORT;
- break;
- case SUGGEST_RETRY:
-#ifdef DEBUG
- printk("SENSE SUGGEST RETRY - status = MAYREDO\n");
-#endif
- status = MAYREDO;
- exit = DRIVER_SENSE | SUGGEST_RETRY;
- break;
- case SUGGEST_ABORT:
-#ifdef DEBUG
- printk("SENSE SUGGEST ABORT - status = CMD_FINISHED");
-#endif
- status = CMD_FINISHED;
- exit = DRIVER_SENSE | SUGGEST_ABORT;
- break;
- default:
- printk("Internal error %s %d \n", __FILE__,
- __LINE__);
- }
- }
- /* end WAS_SENSE */
- else {
-#ifdef DEBUG
- printk("COMMAND COMPLETE message returned, "
- "status = CMD_FINISHED. \n");
-#endif
- exit = DRIVER_OK;
- status = CMD_FINISHED;
- }
- break;
-
- case CHECK_CONDITION:
- case COMMAND_TERMINATED:
- switch (check_sense(SCpnt)) {
- case 0:
- update_timeout(SCpnt, oldto);
- status = REDO;
- break;
- case SUGGEST_REMAP:
- status = CMD_FINISHED;
- exit = DRIVER_SENSE | SUGGEST_ABORT;
- break;
- case SUGGEST_RETRY:
- status = MAYREDO;
- exit = DRIVER_SENSE | SUGGEST_RETRY;
- break;
- case SUGGEST_ABORT:
- status = CMD_FINISHED;
- exit = DRIVER_SENSE | SUGGEST_ABORT;
- break;
- case SUGGEST_SENSE:
- scsi_request_sense(SCpnt);
- status = PENDING;
- break;
- }
- break;
-
- case CONDITION_GOOD:
- case INTERMEDIATE_GOOD:
- case INTERMEDIATE_C_GOOD:
- break;
-
- case BUSY:
- case QUEUE_FULL:
- update_timeout(SCpnt, oldto);
- status = REDO;
- break;
-
- case RESERVATION_CONFLICT:
- printk("scsi%d, channel %d : RESERVATION CONFLICT performing"
- " reset.\n", SCpnt->host->host_no, SCpnt->channel);
- scsi_reset(SCpnt, SCSI_RESET_SYNCHRONOUS);
- status = REDO;
- break;
- default:
- printk("Internal error %s %d \n"
- "status byte = %d \n", __FILE__,
- __LINE__, status_byte(result));
-
- }
- break;
- default:
- panic("scsi: unsupported message byte %d received\n",
- msg_byte(result));
- }
- break;
- case DID_TIME_OUT:
-#ifdef DEBUG
- printk("Host returned DID_TIME_OUT - ");
-#endif
-
- if (SCpnt->flags & WAS_TIMEDOUT) {
-#ifdef DEBUG
- printk("Aborting\n");
-#endif
- /*
- Allow TEST_UNIT_READY and INQUIRY commands to timeout early
- without causing resets. All other commands should be retried.
- */
- if (SCpnt->cmnd[0] != TEST_UNIT_READY &&
- SCpnt->cmnd[0] != INQUIRY)
- status = MAYREDO;
- exit = (DRIVER_TIMEOUT | SUGGEST_ABORT);
- } else {
-#ifdef DEBUG
- printk("Retrying.\n");
-#endif
- SCpnt->flags |= WAS_TIMEDOUT;
- SCpnt->internal_timeout &= ~IN_ABORT;
- status = REDO;
- }
- break;
- case DID_BUS_BUSY:
- case DID_PARITY:
- status = REDO;
- break;
- case DID_NO_CONNECT:
-#ifdef DEBUG
- printk("Couldn't connect.\n");
-#endif
- exit = (DRIVER_HARD | SUGGEST_ABORT);
- break;
- case DID_ERROR:
- status = MAYREDO;
- exit = (DRIVER_HARD | SUGGEST_ABORT);
- break;
- case DID_BAD_TARGET:
- case DID_ABORT:
- exit = (DRIVER_INVALID | SUGGEST_ABORT);
- break;
- case DID_RESET:
- if (SCpnt->flags & IS_RESETTING) {
- SCpnt->flags &= ~IS_RESETTING;
- status = REDO;
- break;
- }
- if (msg_byte(result) == GOOD &&
- status_byte(result) == CHECK_CONDITION) {
- switch (check_sense(SCpnt)) {
- case 0:
- update_timeout(SCpnt, oldto);
- status = REDO;
- break;
- case SUGGEST_REMAP:
- case SUGGEST_RETRY:
- status = MAYREDO;
- exit = DRIVER_SENSE | SUGGEST_RETRY;
- break;
- case SUGGEST_ABORT:
- status = CMD_FINISHED;
- exit = DRIVER_SENSE | SUGGEST_ABORT;
- break;
- case SUGGEST_SENSE:
- scsi_request_sense(SCpnt);
- status = PENDING;
- break;
- }
- } else {
- status = REDO;
- exit = SUGGEST_RETRY;
- }
- break;
- default:
- exit = (DRIVER_ERROR | SUGGEST_DIE);
- }
-
- switch (status) {
- case CMD_FINISHED:
- case PENDING:
- break;
- case MAYREDO:
-#ifdef DEBUG
- printk("In MAYREDO, allowing %d retries, have %d\n",
- SCpnt->allowed, SCpnt->retries);
-#endif
- if ((++SCpnt->retries) < SCpnt->allowed) {
- if ((SCpnt->retries >= (SCpnt->allowed >> 1))
- && !(SCpnt->host->resetting && time_before(jiffies, SCpnt->host->last_reset + MIN_RESET_PERIOD))
- && !(SCpnt->flags & WAS_RESET)) {
- printk("scsi%d channel %d : resetting for second half of retries.\n",
- SCpnt->host->host_no, SCpnt->channel);
- scsi_reset(SCpnt, SCSI_RESET_SYNCHRONOUS);
- /* fall through to REDO */
- }
- } else {
- status = CMD_FINISHED;
- break;
- }
- /* fall through to REDO */
-
- case REDO:
-
- if (SCpnt->flags & WAS_SENSE)
- scsi_request_sense(SCpnt);
- else {
- memcpy((void *) SCpnt->cmnd,
- (void *) SCpnt->data_cmnd,
- sizeof(SCpnt->data_cmnd));
- memset((void *) SCpnt->sense_buffer, 0,
- sizeof(SCpnt->sense_buffer));
- SCpnt->request_buffer = SCpnt->buffer;
- SCpnt->request_bufflen = SCpnt->bufflen;
- SCpnt->use_sg = SCpnt->old_use_sg;
- SCpnt->cmd_len = SCpnt->old_cmd_len;
- SCpnt->sc_data_direction = SCpnt->sc_old_data_direction;
- SCpnt->underflow = SCpnt->old_underflow;
- SCpnt->result = 0;
- /*
- * Ugly, ugly. The newer interfaces all
- * assume that the lock isn't held. Mustn't
- * disappoint, or we deadlock the system.
- */
- spin_unlock_irq(&host->host_lock);
- scsi_dispatch_cmd(SCpnt);
- spin_lock_irq(&host->host_lock);
- }
- break;
- default:
- INTERNAL_ERROR;
- }
-
- if (status == CMD_FINISHED) {
- Scsi_Request *SRpnt;
-#ifdef DEBUG
- printk("Calling done function - at address %p\n", SCpnt->done);
-#endif
- host->host_busy--; /* Indicate that we are free */
- device->device_busy--; /* Decrement device usage counter. */
-
- SCpnt->result = result | ((exit & 0xff) << 24);
- SCpnt->use_sg = SCpnt->old_use_sg;
- SCpnt->cmd_len = SCpnt->old_cmd_len;
- SCpnt->sc_data_direction = SCpnt->sc_old_data_direction;
- SCpnt->underflow = SCpnt->old_underflow;
- /*
- * The upper layers assume the lock isn't held. We mustn't
- * disappoint them. When the new error handling code is in
- * use, the upper code is run from a bottom half handler, so
- * it isn't an issue.
- */
- spin_unlock_irq(&host->host_lock);
- SRpnt = SCpnt->sc_request;
- if( SRpnt != NULL ) {
- SRpnt->sr_result = SRpnt->sr_command->result;
- if( SRpnt->sr_result != 0 ) {
- memcpy(SRpnt->sr_sense_buffer,
- SRpnt->sr_command->sense_buffer,
- sizeof(SRpnt->sr_sense_buffer));
- }
- }
-
- SCpnt->done(SCpnt);
- spin_lock_irq(&host->host_lock);
- }
-#undef CMD_FINISHED
-#undef REDO
-#undef MAYREDO
-#undef PENDING
-}
-
-/*
- * The scsi_abort function interfaces with the abort() function of the host
- * we are aborting, and causes the current command to not complete. The
- * caller should deal with any error messages or status returned on the
- * next call.
- *
- * This will not be called reentrantly for a given host.
- */
-
-/*
- * Since we're nice guys and specified that abort() and reset()
- * can be non-reentrant. The internal_timeout flags are used for
- * this.
- */
-
-
-static int scsi_abort(Scsi_Cmnd * SCpnt, int why)
-{
- int oldto;
- struct Scsi_Host *host = SCpnt->host;
-
- while (1) {
-
- /*
- * Protect against races here. If the command is done, or we are
- * on a different command forget it.
- */
- if (SCpnt->serial_number != SCpnt->serial_number_at_timeout) {
- return 0;
- }
- if (SCpnt->internal_timeout & IN_ABORT) {
- spin_unlock_irq(&host->host_lock);
- while (SCpnt->internal_timeout & IN_ABORT)
- barrier();
- spin_lock_irq(&host->host_lock);
- } else {
- SCpnt->internal_timeout |= IN_ABORT;
- oldto = update_timeout(SCpnt, ABORT_TIMEOUT);
-
- if ((SCpnt->flags & IS_RESETTING) && SCpnt->device->soft_reset) {
- /* OK, this command must have died when we did the
- * reset. The device itself must have lied.
- */
- printk("Stale command on %d %d:%d appears to have died when"
- " the bus was reset\n",
- SCpnt->channel, SCpnt->target, SCpnt->lun);
- }
- if (!host->host_busy) {
- SCpnt->internal_timeout &= ~IN_ABORT;
- update_timeout(SCpnt, oldto);
- return 0;
- }
- printk("scsi : aborting command due to timeout : pid %lu, scsi%d,"
- " channel %d, id %d, lun %d ",
- SCpnt->pid, SCpnt->host->host_no, (int) SCpnt->channel,
- (int) SCpnt->target, (int) SCpnt->lun);
- print_command(SCpnt->cmnd);
- if (SCpnt->serial_number != SCpnt->serial_number_at_timeout)
- return 0;
- SCpnt->abort_reason = why;
- switch (host->hostt->abort(SCpnt)) {
- /* We do not know how to abort. Try waiting another
- * time increment and see if this helps. Set the
- * WAS_TIMEDOUT flag set so we do not try this twice
- */
- case SCSI_ABORT_BUSY: /* Tough call - returning 1 from
- * this is too severe
- */
- case SCSI_ABORT_SNOOZE:
- if (why == DID_TIME_OUT) {
- SCpnt->internal_timeout &= ~IN_ABORT;
- if (SCpnt->flags & WAS_TIMEDOUT) {
- return 1; /* Indicate we cannot handle this.
- * We drop down into the reset handler
- * and try again
- */
- } else {
- SCpnt->flags |= WAS_TIMEDOUT;
- oldto = SCpnt->timeout_per_command;
- update_timeout(SCpnt, oldto);
- }
- }
- return 0;
- case SCSI_ABORT_PENDING:
- if (why != DID_TIME_OUT) {
- update_timeout(SCpnt, oldto);
- }
- return 0;
- case SCSI_ABORT_SUCCESS:
- /* We should have already aborted this one. No
- * need to adjust timeout
- */
- SCpnt->internal_timeout &= ~IN_ABORT;
- return 0;
- case SCSI_ABORT_NOT_RUNNING:
- SCpnt->internal_timeout &= ~IN_ABORT;
- update_timeout(SCpnt, 0);
- return 0;
- case SCSI_ABORT_ERROR:
- default:
- SCpnt->internal_timeout &= ~IN_ABORT;
- return 1;
- }
- }
- }
-}
-
-
-/* Mark a single SCSI Device as having been reset. */
-
-static inline void scsi_mark_device_reset(Scsi_Device * Device)
-{
- Device->was_reset = 1;
- Device->expecting_cc_ua = 1;
-}
-
-
-/* Mark all SCSI Devices on a specific Host as having been reset. */
-
-void scsi_mark_host_reset(struct Scsi_Host *Host)
-{
- Scsi_Cmnd *SCpnt;
- Scsi_Device *SDpnt;
-
- for (SDpnt = Host->host_queue; SDpnt; SDpnt = SDpnt->next) {
- for (SCpnt = SDpnt->device_queue; SCpnt; SCpnt = SCpnt->next)
- scsi_mark_device_reset(SCpnt->device);
- }
-}
-
-
-/* Mark all SCSI Devices on a specific Host Bus as having been reset. */
-
-static void scsi_mark_bus_reset(struct Scsi_Host *Host, int channel)
-{
- Scsi_Cmnd *SCpnt;
- Scsi_Device *SDpnt;
-
- for (SDpnt = Host->host_queue; SDpnt; SDpnt = SDpnt->next) {
- for (SCpnt = SDpnt->device_queue; SCpnt; SCpnt = SCpnt->next)
- if (SCpnt->channel == channel)
- scsi_mark_device_reset(SCpnt->device);
- }
-}
-
-
-static int scsi_reset(Scsi_Cmnd * SCpnt, unsigned int reset_flags)
-{
- int temp;
- Scsi_Cmnd *SCpnt1;
- Scsi_Device *SDpnt;
- struct Scsi_Host *host = SCpnt->host;
-
- printk("SCSI bus is being reset for host %d channel %d.\n",
- host->host_no, SCpnt->channel);
-
-#if 0
- /*
- * First of all, we need to make a recommendation to the low-level
- * driver as to whether a BUS_DEVICE_RESET should be performed,
- * or whether we should do a full BUS_RESET. There is no simple
- * algorithm here - we basically use a series of heuristics
- * to determine what we should do.
- */
- SCpnt->host->suggest_bus_reset = FALSE;
-
- /*
- * First see if all of the active devices on the bus have
- * been jammed up so that we are attempting resets. If so,
- * then suggest a bus reset. Forcing a bus reset could
- * result in some race conditions, but no more than
- * you would usually get with timeouts. We will cross
- * that bridge when we come to it.
- *
- * This is actually a pretty bad idea, since a sequence of
- * commands will often timeout together and this will cause a
- * Bus Device Reset followed immediately by a SCSI Bus Reset.
- * If all of the active devices really are jammed up, the
- * Bus Device Reset will quickly timeout and scsi_times_out
- * will follow up with a SCSI Bus Reset anyway.
- */
- SCpnt1 = host->host_queue;
- while (SCpnt1) {
- if (SCpnt1->request.rq_status != RQ_INACTIVE
- && (SCpnt1->flags & (WAS_RESET | IS_RESETTING)) == 0)
- break;
- SCpnt1 = SCpnt1->next;
- }
- if (SCpnt1 == NULL) {
- reset_flags |= SCSI_RESET_SUGGEST_BUS_RESET;
- }
- /*
- * If the code that called us is suggesting a hard reset, then
- * definitely request it. This usually occurs because a
- * BUS_DEVICE_RESET times out.
- *
- * Passing reset_flags along takes care of this automatically.
- */
- if (reset_flags & SCSI_RESET_SUGGEST_BUS_RESET) {
- SCpnt->host->suggest_bus_reset = TRUE;
- }
-#endif
-
- while (1) {
-
- /*
- * Protect against races here. If the command is done, or we are
- * on a different command forget it.
- */
- if (reset_flags & SCSI_RESET_ASYNCHRONOUS)
- if (SCpnt->serial_number != SCpnt->serial_number_at_timeout) {
- return 0;
- }
- if (SCpnt->internal_timeout & IN_RESET) {
- spin_unlock_irq(&host->host_lock);
- while (SCpnt->internal_timeout & IN_RESET)
- barrier();
- spin_lock_irq(&host->host_lock);
- } else {
- SCpnt->internal_timeout |= IN_RESET;
- update_timeout(SCpnt, RESET_TIMEOUT);
-
- if (reset_flags & SCSI_RESET_SYNCHRONOUS)
- SCpnt->flags |= SYNC_RESET;
- if (host->host_busy) {
- for (SDpnt = host->host_queue; SDpnt; SDpnt = SDpnt->next) {
- SCpnt1 = SDpnt->device_queue;
- while (SCpnt1) {
- if (SCpnt1->request.rq_status != RQ_INACTIVE) {
-#if 0
- if (!(SCpnt1->flags & IS_RESETTING) &&
- !(SCpnt1->internal_timeout & IN_ABORT))
- scsi_abort(SCpnt1, DID_RESET);
-#endif
- SCpnt1->flags |= (WAS_RESET | IS_RESETTING);
- }
- SCpnt1 = SCpnt1->next;
- }
- }
-
- host->last_reset = jiffies;
- host->resetting = 1;
- /*
- * I suppose that the host reset callback will not play
- * with the resetting field. We have just set the resetting
- * flag here. -arca
- */
- temp = host->hostt->reset(SCpnt, reset_flags);
- /*
- This test allows the driver to introduce an additional bus
- settle time delay by setting last_reset up to 20 seconds in
- the future. In the normal case where the driver does not
- modify last_reset, it must be assumed that the actual bus
- reset occurred immediately prior to the return to this code,
- and so last_reset must be updated to the current time, so
- that the delay in internal_cmnd will guarantee at least a
- MIN_RESET_DELAY bus settle time.
- */
- if (host->last_reset - jiffies > 20UL * HZ)
- host->last_reset = jiffies;
- } else {
- host->host_busy++;
- host->last_reset = jiffies;
- host->resetting = 1;
- SCpnt->flags |= (WAS_RESET | IS_RESETTING);
- /*
- * I suppose that the host reset callback will not play
- * with the resetting field. We have just set the resetting
- * flag here. -arca
- */
- temp = host->hostt->reset(SCpnt, reset_flags);
- if (time_before(host->last_reset, jiffies) ||
- (time_after(host->last_reset, jiffies + 20 * HZ)))
- host->last_reset = jiffies;
- host->host_busy--;
- }
- if (reset_flags & SCSI_RESET_SYNCHRONOUS)
- SCpnt->flags &= ~SYNC_RESET;
-
-#ifdef DEBUG
- printk("scsi reset function returned %d\n", temp);
-#endif
-
- /*
- * Now figure out what we need to do, based upon
- * what the low level driver said that it did.
- * If the result is SCSI_RESET_SUCCESS, SCSI_RESET_PENDING,
- * or SCSI_RESET_WAKEUP, then the low level driver did a
- * bus device reset or bus reset, so we should go through
- * and mark one or all of the devices on that bus
- * as having been reset.
- */
- switch (temp & SCSI_RESET_ACTION) {
- case SCSI_RESET_SUCCESS:
- if (temp & SCSI_RESET_HOST_RESET)
- scsi_mark_host_reset(host);
- else if (temp & SCSI_RESET_BUS_RESET)
- scsi_mark_bus_reset(host, SCpnt->channel);
- else
- scsi_mark_device_reset(SCpnt->device);
- SCpnt->internal_timeout &= ~(IN_RESET | IN_RESET2 | IN_RESET3);
- return 0;
- case SCSI_RESET_PENDING:
- if (temp & SCSI_RESET_HOST_RESET)
- scsi_mark_host_reset(host);
- else if (temp & SCSI_RESET_BUS_RESET)
- scsi_mark_bus_reset(host, SCpnt->channel);
- else
- scsi_mark_device_reset(SCpnt->device);
- case SCSI_RESET_NOT_RUNNING:
- return 0;
- case SCSI_RESET_PUNT:
- SCpnt->internal_timeout &= ~(IN_RESET | IN_RESET2 | IN_RESET3);
- scsi_request_sense(SCpnt);
- return 0;
- case SCSI_RESET_WAKEUP:
- if (temp & SCSI_RESET_HOST_RESET)
- scsi_mark_host_reset(host);
- else if (temp & SCSI_RESET_BUS_RESET)
- scsi_mark_bus_reset(host, SCpnt->channel);
- else
- scsi_mark_device_reset(SCpnt->device);
- SCpnt->internal_timeout &= ~(IN_RESET | IN_RESET2 | IN_RESET3);
- scsi_request_sense(SCpnt);
- /*
- * If a bus reset was performed, we
- * need to wake up each and every command
- * that was active on the bus or if it was a HBA
- * reset all active commands on all channels
- */
- if (temp & SCSI_RESET_HOST_RESET) {
- for (SDpnt = host->host_queue; SDpnt; SDpnt = SDpnt->next) {
- SCpnt1 = SDpnt->device_queue;
- while (SCpnt1) {
- if (SCpnt1->request.rq_status != RQ_INACTIVE
- && SCpnt1 != SCpnt)
- scsi_request_sense(SCpnt1);
- SCpnt1 = SCpnt1->next;
- }
- }
- } else if (temp & SCSI_RESET_BUS_RESET) {
- for (SDpnt = host->host_queue; SDpnt; SDpnt = SDpnt->next) {
- SCpnt1 = SDpnt->device_queue;
- while (SCpnt1) {
- if (SCpnt1->request.rq_status != RQ_INACTIVE
- && SCpnt1 != SCpnt
- && SCpnt1->channel == SCpnt->channel)
- scsi_request_sense(SCpnt);
- SCpnt1 = SCpnt1->next;
- }
- }
- }
- return 0;
- case SCSI_RESET_SNOOZE:
- /* In this case, we set the timeout field to 0
- * so that this command does not time out any more,
- * and we return 1 so that we get a message on the
- * screen.
- */
- SCpnt->internal_timeout &= ~(IN_RESET | IN_RESET2 | IN_RESET3);
- update_timeout(SCpnt, 0);
- /* If you snooze, you lose... */
- case SCSI_RESET_ERROR:
- default:
- return 1;
- }
-
- return temp;
- }
- }
-}
-
-/*
- * The strategy is to cause the timer code to call scsi_times_out()
- * when the soonest timeout is pending.
- * The arguments are used when we are queueing a new command, because
- * we do not want to subtract the time used from this time, but when we
- * set the timer, we want to take this value into account.
- */
-
-int update_timeout(Scsi_Cmnd * SCset, int timeout)
-{
- int rtn;
-
- /*
- * We are using the new error handling code to actually register/deregister
- * timers for timeout.
- */
-
- if (!timer_pending(&SCset->eh_timeout)) {
- rtn = 0;
- } else {
- rtn = SCset->eh_timeout.expires - jiffies;
- }
-
- if (timeout == 0) {
- scsi_delete_timer(SCset);
- } else {
- scsi_add_timer(SCset, timeout, scsi_old_times_out);
- }
-
- return rtn;
-}
-
-
-/*
- * Overrides for Emacs so that we follow Linus's tabbing style.
- * Emacs will notice this stuff at the end of the file and automatically
- * adjust the settings for this buffer only. This must remain at the end
- * of the file.
- * ---------------------------------------------------------------------------
- * Local variables:
- * c-indent-level: 4
- * c-brace-imaginary-offset: 0
- * c-brace-offset: -4
- * c-argdecl-indent: 4
- * c-label-offset: -4
- * c-continued-statement-offset: 4
- * c-continued-brace-offset: 0
- * indent-tabs-mode: nil
- * tab-width: 8
- * End:
- */
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/scsi_obsolete.h linux/drivers/scsi/scsi_obsolete.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/scsi_obsolete.h Sat Sep 4 19:48:46 1999
+++ linux/drivers/scsi/scsi_obsolete.h Thu Jan 1 01:00:00 1970
@@ -1,106 +0,0 @@
-/*
- * scsi_obsolete.h Copyright (C) 1997 Eric Youngdale
- *
- */
-
-#ifndef _SCSI_OBSOLETE_H
-#define _SCSI_OBSOLETE_H
-
-/*
- * These are the return codes for the abort and reset functions. The mid-level
- * code uses these to decide what to do next. Each of the low level abort
- * and reset functions must correctly indicate what it has done.
- * The descriptions are written from the point of view of the mid-level code,
- * so that the return code is telling the mid-level drivers exactly what
- * the low level driver has already done, and what remains to be done.
- */
-
-/* We did not do anything.
- * Wait some more for this command to complete, and if this does not work,
- * try something more serious. */
-#define SCSI_ABORT_SNOOZE 0
-
-/* This means that we were able to abort the command. We have already
- * called the mid-level done function, and do not expect an interrupt that
- * will lead to another call to the mid-level done function for this command */
-#define SCSI_ABORT_SUCCESS 1
-
-/* We called for an abort of this command, and we should get an interrupt
- * when this succeeds. Thus we should not restore the timer for this
- * command in the mid-level abort function. */
-#define SCSI_ABORT_PENDING 2
-
-/* Unable to abort - command is currently on the bus. Grin and bear it. */
-#define SCSI_ABORT_BUSY 3
-
-/* The command is not active in the low level code. Command probably
- * finished. */
-#define SCSI_ABORT_NOT_RUNNING 4
-
-/* Something went wrong. The low level driver will indicate the correct
- * error condition when it calls scsi_done, so the mid-level abort function
- * can simply wait until this comes through */
-#define SCSI_ABORT_ERROR 5
-
-/* We do not know how to reset the bus, or we do not want to. Bummer.
- * Anyway, just wait a little more for the command in question, and hope that
- * it eventually finishes. If it never finishes, the SCSI device could
- * hang, so use this with caution. */
-#define SCSI_RESET_SNOOZE 0
-
-/* We do not know how to reset the bus, or we do not want to. Bummer.
- * We have given up on this ever completing. The mid-level code will
- * request sense information to decide how to proceed from here. */
-#define SCSI_RESET_PUNT 1
-
-/* This means that we were able to reset the bus. We have restarted all of
- * the commands that should be restarted, and we should be able to continue
- * on normally from here. We do not expect any interrupts that will return
- * DID_RESET to any of the other commands in the host_queue, and the mid-level
- * code does not need to do anything special to keep the commands alive.
- * If a hard reset was performed then all outstanding commands on the
- * bus have been restarted. */
-#define SCSI_RESET_SUCCESS 2
-
-/* We called for a reset of this bus, and we should get an interrupt
- * when this succeeds. Each command should get its own status
- * passed up to scsi_done, but this has not happened yet.
- * If a hard reset was performed, then we expect an interrupt
- * for *each* of the outstanding commands that will have the
- * effect of restarting the commands.
- */
-#define SCSI_RESET_PENDING 3
-
-/* We did a reset, but do not expect an interrupt to signal DID_RESET.
- * This tells the upper level code to request the sense info, and this
- * should keep the command alive. */
-#define SCSI_RESET_WAKEUP 4
-
-/* The command is not active in the low level code. Command probably
- finished. */
-#define SCSI_RESET_NOT_RUNNING 5
-
-/* Something went wrong, and we do not know how to fix it. */
-#define SCSI_RESET_ERROR 6
-
-#define SCSI_RESET_SYNCHRONOUS 0x01
-#define SCSI_RESET_ASYNCHRONOUS 0x02
-#define SCSI_RESET_SUGGEST_BUS_RESET 0x04
-#define SCSI_RESET_SUGGEST_HOST_RESET 0x08
-/*
- * This is a bitmask that is ored with one of the above codes.
- * It tells the mid-level code that we did a hard reset.
- */
-#define SCSI_RESET_BUS_RESET 0x100
-/*
- * This is a bitmask that is ored with one of the above codes.
- * It tells the mid-level code that we did a host adapter reset.
- */
-#define SCSI_RESET_HOST_RESET 0x200
-/*
- * Used to mask off bits and to obtain the basic action that was
- * performed.
- */
-#define SCSI_RESET_ACTION 0xff
-
-#endif /* SCSI_OBSOLETE_H */
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/scsi_syms.c linux/drivers/scsi/scsi_syms.c
--- ../master/linux-2.5.1-pre2/drivers/scsi/scsi_syms.c Tue Jul 3 00:27:56 2001
+++ linux/drivers/scsi/scsi_syms.c Wed Nov 28 13:39:16 2001
@@ -54,7 +54,6 @@
EXPORT_SYMBOL(scsi_release_command);
EXPORT_SYMBOL(print_Scsi_Cmnd);
EXPORT_SYMBOL(scsi_block_when_processing_errors);
-EXPORT_SYMBOL(scsi_mark_host_reset);
EXPORT_SYMBOL(scsi_ioctl_send_command);
#if defined(CONFIG_SCSI_LOGGING) /* { */
EXPORT_SYMBOL(scsi_logging_level);
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/sim710.h linux/drivers/scsi/sim710.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/sim710.h Fri Jul 20 06:07:45 2001
+++ linux/drivers/scsi/sim710.h Wed Nov 28 13:25:18 2001
@@ -38,7 +38,6 @@
sg_tablesize: 128, \
cmd_per_lun: 1, \
use_clustering: DISABLE_CLUSTERING, \
- use_new_eh_code: 1}
#ifndef HOSTS_C
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/sym53c8xx_2/sym53c8xx.h linux/drivers/scsi/sym53c8xx_2/sym53c8xx.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/sym53c8xx_2/sym53c8xx.h Wed Nov 28 13:20:39 2001
+++ linux/drivers/scsi/sym53c8xx_2/sym53c8xx.h Wed Nov 28 13:36:10 2001
@@ -109,7 +109,6 @@
release: sym53c8xx_release, \
info: sym53c8xx_info, \
queuecommand: sym53c8xx_queue_command, \
- use_new_eh_code: 1, \
eh_abort_handler: sym53c8xx_eh_abort_handler, \
eh_device_reset_handler:sym53c8xx_eh_device_reset_handler, \
eh_bus_reset_handler: sym53c8xx_eh_bus_reset_handler, \
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/u14-34f.h linux/drivers/scsi/u14-34f.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/u14-34f.h Sun May 20 02:43:06 2001
+++ linux/drivers/scsi/u14-34f.h Wed Nov 28 13:25:12 2001
@@ -30,7 +30,6 @@
this_id: 7, \
unchecked_isa_dma: 1, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 1 /* Enable new error code */ \
}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/scsi/wd7000.h linux/drivers/scsi/wd7000.h
--- ../master/linux-2.5.1-pre2/drivers/scsi/wd7000.h Fri Jul 20 06:08:46 2001
+++ linux/drivers/scsi/wd7000.h Wed Nov 28 13:25:04 2001
@@ -57,6 +57,5 @@
cmd_per_lun: 1, \
unchecked_isa_dma: 1, \
use_clustering: ENABLE_CLUSTERING, \
- use_new_eh_code: 0 \
}
#endif
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/usb/hpusbscsi.h linux/drivers/usb/hpusbscsi.h
--- ../master/linux-2.5.1-pre2/drivers/usb/hpusbscsi.h Fri Oct 5 21:04:51 2001
+++ linux/drivers/usb/hpusbscsi.h Wed Nov 28 13:33:51 2001
@@ -73,7 +73,6 @@
present: 0,
unchecked_isa_dma: FALSE,
use_clustering: TRUE,
- use_new_eh_code: TRUE,
emulated: TRUE
};
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/usb/microtek.c linux/drivers/usb/microtek.c
--- ../master/linux-2.5.1-pre2/drivers/usb/microtek.c Fri Oct 5 21:04:51 2001
+++ linux/drivers/usb/microtek.c Wed Nov 28 13:33:37 2001
@@ -760,7 +760,6 @@
present: 0,
unchecked_isa_dma: FALSE,
use_clustering: TRUE,
- use_new_eh_code: TRUE,
emulated: TRUE
};
diff -uNr -Xdontdiff ../master/linux-2.5.1-pre2/drivers/usb/storage/scsiglue.c linux/drivers/usb/storage/scsiglue.c
--- ../master/linux-2.5.1-pre2/drivers/usb/storage/scsiglue.c Sun Nov 11 19:01:32 2001
+++ linux/drivers/usb/storage/scsiglue.c Wed Nov 28 13:33:57 2001
@@ -388,7 +388,6 @@
present: 0,
unchecked_isa_dma: FALSE,
use_clustering: TRUE,
- use_new_eh_code: TRUE,
emulated: TRUE
};
On Tue, Nov 27 2001, Jeff V. Merkey wrote:
> On Tue, Nov 27, 2001 at 05:04:46PM -0800, Linus Torvalds wrote:
> >
> > On Wed, 28 Nov 2001, Paul Mackerras wrote:
> > >
> > > Is there a description of the new block layer and its interface to
> > > block device drivers somewhere? That would be helpful, since Ben
> > > Herrenschmidt and I are going to have to convert several
> > > powermac-specific drivers.
> >
> > Jens has something written up, which he sent to me as an introduction to
> > the patch. I'll send that out unless he does a cleaned-up version, but I'd
> > actually prefer for him to do the sending. Jens?
> >
> > Linus
> >
>
>
> Linus/Jens,
>
> I've just completed my review of submit_bio and the changes to
> generic_make_request and I have some questions for whomever
> can answer.
>
> 1. The changes made to submit_bh indicate I can now send long
> chains of variable block size requests to the I/O layer similiar
> to the capability of Windows 2000 and NetWare I/O subsystems.
Yes, you can build generically a single I/O unit that spans up to 256
pages. If you bypass the bio_alloc/bvec_alloc mechanism, the sky is the
limit. Beware that a really big bio may need to be split up in the end
for devices that can't handle them that big.
> 2. The elevator layer is merging these requests, and making a
> single sweep request for contiguous sector runs.
Like always, yes.
> 3. In theory, I should be able to support page cache capability
> for NWFS and possibly NTFS in Linux the way these wierd non-Unix
> OS's work.
Maybe :-)
> 4. This interface may **NOT** support non-block aligned requests
> across all the drivers. I also need to be able to submit a
> request chain 512-2048-512-1024-4096 where the first IO requested
> may by on a non-block aligned boundry. i.e. Device is configured
> for 1024 byte blocks, I start the request as 512 @ LBA 1 -> 1024 @ LBA 2,
> etc. The code looks like it will work.
As long as the smallest unit above is at least the size of the hardware
sector on the target, it should be ok.
> I would love to test this wonderful code and will hopefully this evening,
> however, all the SCSI drivers appear to be broken, as well as the
> 3Ware. :-)
Well not all, but many. I only converted stuff I could personally test,
basically, plus a bit more. Usually converting a SCSI driver is not a
lot of work, please see the changes to sym/sym2 etc in the pre2 patch.
--
Jens Axboe
On Nov 28, 2001 13:55 +0100, Christoph Hellwig wrote:
> > > While we are at breaking scsi, would you take a patch to remove the
> > > old-style (2.0) scsi error handling completly, forcing drivers still
> > > using it to be fixed? Early 2.5 looks like a good time for that to me..
>
> --- ../master/linux-2.5.1-pre2/drivers/scsi/aic7xxx_old/aic7xxx.h Sun Mar 4 23:30:18 2001
> +++ linux/drivers/scsi/aic7xxx_old/aic7xxx.h Wed Nov 28 13:35:21 2001
> @@ -55,7 +55,6 @@
> present: 0, /* number of 7xxx's present */\
> unchecked_isa_dma: 0, /* no memory DMA restrictions */\
> use_clustering: ENABLE_CLUSTERING, \
> - use_new_eh_code: 0 \
> }
>
> extern int aic7xxx_queue(Scsi_Cmnd *, void (*)(Scsi_Cmnd *));
What would be nice in the case of drivers that don't use the new error
handling code is to add something like:
#warning "Uses obsolete SCSI error code, see Documentation/2.5/scsi-error.txt"
for a hint as to the reason why it no longer compiles, and a short guide
on how to update the drivers.
The same would be good for Jens' changes - his document could be put into
2.5/bio.txt or something, so any currently out-of-kernel coders can find it.
Cheers, Andreas
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/
On Wed, Nov 28, 2001 at 09:26:00AM -0700, Andreas Dilger wrote:
> What would be nice in the case of drivers that don't use the new error
> handling code is to add something like:
>
> #warning "Uses obsolete SCSI error code, see Documentation/2.5/scsi-error.txt"
>
> for a hint as to the reason why it no longer compiles, and a short guide
> on how to update the drivers.
I already thought about that - as the old error handling code is selected
by setting a member in a struct to '1' I don't see any easy way to do so...
Christoph
--
Of course it doesn't work. We've performed a software upgrade.
Christoph Hellwig wrote:
>
> On Wed, Nov 28, 2001 at 09:26:00AM -0700, Andreas Dilger wrote:
> > What would be nice in the case of drivers that don't use the new error
> > handling code is to add something like:
> >
> > #warning "Uses obsolete SCSI error code, see Documentation/2.5/scsi-error.txt"
> >
> > for a hint as to the reason why it no longer compiles, and a short guide
> > on how to update the drivers.
>
> I already thought about that - as the old error handling code is selected
> by setting a member in a struct to '1' I don't see any easy way to do so...
>
> Christoph
Please note that this selection is static with regard to the driver.
It only happens in the initialization code for the driver!
The #warning doesn't have to at the same place where the filed in
the struct was.
On Nov 28, 2001 17:42 +0100, Christoph Hellwig wrote:
> On Wed, Nov 28, 2001 at 09:26:00AM -0700, Andreas Dilger wrote:
> > What would be nice in the case of drivers that don't use the new error
> > handling code is to add something like:
> >
> > #warning "Uses obsolete SCSI error code, see Documentation/2.5/scsi-error.txt"
> >
> > for a hint as to the reason why it no longer compiles, and a short guide
> > on how to update the drivers.
>
> I already thought about that - as the old error handling code is selected
> by setting a member in a struct to '1' I don't see any easy way to do so...
I'm thinking about a compile-time #warning/#error, that at least tells an
interested party where to start looking for how to fix this. It could be
anywhere in the affected driver source file (preferrably at the top, so it
is hit before the compiler exits because of too many other errors). It
would just be a matter of cut-n-paste a single line into every file which
previously used the old error-handling code.
I presume that the nature of this change will either cause compile errors
anyways (because of missing struct definitions), or linker errors (because
of missing functions). I think that having a #warning at compile time will
at least avoid a whole bunch of "this driver doesn't compile, help me"
emails, and may also help someone to actually update the driver rather than
giving up because they have no idea where to begin looking for the fix.
Actually, what would also be nice is to include the kernel version where
this change happened, so that interested parties could also see what changes
where necessary to bring about this fix. In this case (removing the old
error handling support), it probably doesn't help much, but in Jens' BIO
changes it _would_ be very helpful to know when a major change was made
so potential fixers can "follow along" with the changes rather than having
to reverse-engineer it 50 releases from now (along with the 33 other major
changes that have been made along the way). Even now, with the BIO +
SCSI error handling changes there are a lot of SCSI drivers with problems
and it won't be "obvious" how to fix them later on.
Cheers, Andreas
PS - are you sure the old error handling code was selected with a '1' and
not with a '0' for "use_new_eh_code"?
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/
----- Original Message -----
From: "Jens Axboe" <[email protected]>
To: "Jeff V. Merkey" <[email protected]>
Cc: "Linus Torvalds" <[email protected]>;
<[email protected]>; <[email protected]>
Sent: Wednesday, November 28, 2001 6:35 AM
Subject: Re: Block I/O Enchancements, 2.5.1-pre2
> On Tue, Nov 27 2001, Jeff V. Merkey wrote:
> > On Tue, Nov 27, 2001 at 05:04:46PM -0800, Linus Torvalds wrote:
> > >
> > > On Wed, 28 Nov 2001, Paul Mackerras wrote:
> > > >
> > > > Is there a description of the new block layer and its interface to
> > > > block device drivers somewhere? That would be helpful, since Ben
> > > > Herrenschmidt and I are going to have to convert several
> > > > powermac-specific drivers.
> > >
> > > Jens has something written up, which he sent to me as an introduction
to
> > > the patch. I'll send that out unless he does a cleaned-up version, but
I'd
> > > actually prefer for him to do the sending. Jens?
> > >
> > > Linus
> > >
> >
> >
> > Linus/Jens,
> >
> > I've just completed my review of submit_bio and the changes to
> > generic_make_request and I have some questions for whomever
> > can answer.
> >
> > 1. The changes made to submit_bh indicate I can now send long
> > chains of variable block size requests to the I/O layer similiar
> > to the capability of Windows 2000 and NetWare I/O subsystems.
>
> Yes, you can build generically a single I/O unit that spans up to 256
> pages. If you bypass the bio_alloc/bvec_alloc mechanism, the sky is the
> limit. Beware that a really big bio may need to be split up in the end
> for devices that can't handle them that big.
I got your docs and tried this.
>
> > 2. The elevator layer is merging these requests, and making a
> > single sweep request for contiguous sector runs.
>
> Like always, yes.
>
> > 3. In theory, I should be able to support page cache capability
> > for NWFS and possibly NTFS in Linux the way these wierd non-Unix
> > OS's work.
>
> Maybe :-)
The next posting of NWFS will have the page cache enabled. I am testing the
code
this morning. I also have reduced the memory usage to about 1/100 of what I
was doing
with the in-memory name hashes. I never liked the way NetWare did this.
It's fast
as hell for lookups, but sure uses a lot of memory.
>
> > 4. This interface may **NOT** support non-block aligned requests
> > across all the drivers. I also need to be able to submit a
> > request chain 512-2048-512-1024-4096 where the first IO requested
> > may by on a non-block aligned boundry. i.e. Device is configured
> > for 1024 byte blocks, I start the request as 512 @ LBA 1 -> 1024 @ LBA
2,
> > etc. The code looks like it will work.
>
> As long as the smallest unit above is at least the size of the hardware
> sector on the target, it should be ok.
Some of the Adaptec drivers will be busted. I remmeber trying this once
before at
Linus' suggestion about a year ago, and I saw significant breakage. Oh
well, this
is 2.5. I guess folks will just have to fix their drivers if this breaks
them.
>
> > I would love to test this wonderful code and will hopefully this
evening,
> > however, all the SCSI drivers appear to be broken, as well as the
> > 3Ware. :-)
>
> Well not all, but many. I only converted stuff I could personally test,
> basically, plus a bit more. Usually converting a SCSI driver is not a
> lot of work, please see the changes to sym/sym2 etc in the pre2 patch.
I am testing on IDE systems -- no IDE breakage at present. Andre's IDE
drivers are
some of the fastest implementations out there and are comparable in
performance to
SCSI.
Jeff
>
> --
> Jens Axboe
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
Jens,
I just booted my Athlon VIA KT133 chipset box with 2.5.1-pre2 only to
discover it dropped me into single user mode because /dev/hda2 could not be
mounted. (Rebooting into 2.5.0+viro patch everything is ok, back into
2.5.1-pre2 is broken...)
Looking with hexedit /dev/hda2 when booted into 2.5.1-pre2 the first sector
contains junk, the second sector contains the real data that I see as the
first sector when booted into 2.5.0+viro fix.
That suggests to me there is an off by one error in the VIA IDE driver in
the 2.5.10pre2 kernel causing the partition to start one sector earlier
than it should.
If you would like any further information / patch testing / whatever let me
know.
Best regards,
Anton
--
"I've not lost my mind. It's backed up on tape somewhere." - Unknown
--
Anton Altaparmakov <aia21 at cam.ac.uk> (replace at with @)
Linux NTFS Maintainer / WWW: http://linux-ntfs.sf.net/
ICQ: 8561279 / WWW: http://www-stu.christs.cam.ac.uk/~aia21/
On Nov 28, 2001 23:31 +0000, Anton Altaparmakov wrote:
> I just booted my Athlon VIA KT133 chipset box with 2.5.1-pre2 only to
> discover it dropped me into single user mode because /dev/hda2 could not be
> mounted. (Rebooting into 2.5.0+viro patch everything is ok, back into
> 2.5.1-pre2 is broken...)
>
> Looking with hexedit /dev/hda2 when booted into 2.5.1-pre2 the first sector
> contains junk, the second sector contains the real data that I see as the
> first sector when booted into 2.5.0+viro fix.
>
> That suggests to me there is an off by one error in the VIA IDE driver in
> the 2.5.10pre2 kernel causing the partition to start one sector earlier
> than it should.
It may be an issue with your particular partition table having /dev/hda1
being an odd number of 512-byte sectors long, but Jens' code only doing
math on 1kB blocks. Just speculation of course. What does "fdisk -ul"
tell you under the two kernels?
Cheers, Andreas
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/
At 23:55 28/11/01, Andreas Dilger wrote:
>On Nov 28, 2001 23:31 +0000, Anton Altaparmakov wrote:
> > I just booted my Athlon VIA KT133 chipset box with 2.5.1-pre2 only to
> > discover it dropped me into single user mode because /dev/hda2 could
> not be
> > mounted. (Rebooting into 2.5.0+viro patch everything is ok, back into
> > 2.5.1-pre2 is broken...)
> >
> > Looking with hexedit /dev/hda2 when booted into 2.5.1-pre2 the first
> sector
> > contains junk, the second sector contains the real data that I see as the
> > first sector when booted into 2.5.0+viro fix.
> >
> > That suggests to me there is an off by one error in the VIA IDE driver in
> > the 2.5.10pre2 kernel causing the partition to start one sector earlier
> > than it should.
>
>It may be an issue with your particular partition table having /dev/hda1
>being an odd number of 512-byte sectors long, but Jens' code only doing
>math on 1kB blocks. Just speculation of course. What does "fdisk -ul"
>tell you under the two kernels?
Both exactly the same:
Disk /dev/hda: 255 heads, 63 sectors, 5005 cylinders
Units = sectors of 1 * 512 bytes
Device Boot Start End Blocks Id System
/dev/hda1 * 63 31471334 15735636 83 Linux
/dev/hda2 31471335 35680364 2104515 83 Linux
/dev/hda3 35680365 39889394 2104515 82 Linux swap
/dev/hda4 39889395 80405324 20257965 f Win95 Ext'd (LBA)
/dev/hda5 39889458 71360729 15735636 83 Linux
/dev/hda6 71360793 75264524 1951866 7 HPFS/NTFS
So everything always starts on same sector. And yes, hda2 starts on an odd
sector.
However, that is not the explanation as the ntfs partition hda6 also starts
with an odd sector but that works fine (using NTFS TNG to mount it and also
looking at start of partition with hexedit).
Best regards,
Anton
--
"I've not lost my mind. It's backed up on tape somewhere." - Unknown
--
Anton Altaparmakov <aia21 at cam.ac.uk> (replace at with @)
Linux NTFS Maintainer / WWW: http://linux-ntfs.sf.net/
ICQ: 8561279 / WWW: http://www-stu.christs.cam.ac.uk/~aia21/
On November 27, 2001 11:02 pm, Paul Mackerras wrote:
> Is there a description of the new block layer and its interface to
> block device drivers somewhere? That would be helpful, since Ben
> Herrenschmidt and I are going to have to convert several
> powermac-specific drivers.
http://lse.sourceforge.net/io/bionotes.txt
(Writeup by Suparna on Jens' work and design issues, +5, insightful)
--
Daniel
On November 28, 2001 04:38 am, Linus Torvalds wrote:
> We will probably _not_ get 64-bit page index numbers, though. I don't want
> to make the page structure bigger/slower for very little gain. So the page
> cache is probably going to be limited to about 44 bits (45+ if people
> start doing large pages, which is probably worth it). So there would still
> be partition/file limits on the order of 16-64 TB in the next few years.
The Ext2+ crowd is actively working on preparing for the world of larger
blocks, so that 64KB blocks will be entirely practical. This will get us to
1/4 Petabyte, still with 32 bit block pointers internally. The main problem
that has to be solved is internal fragmentation.
Going beyond 64KB blocks with mimimal changes is possible too, basically just
working around the 16 bit pointers in directory blocks. However, it's not
clear it's a pressing need.
> (In a longer timeframe, assuming RAM keeps getting cheaper and cheaper,
> and 64-bit computing starts hppening on PC's, a few years down the line we
> can re-visit this - that particular transition is not going to be too
> painful).
>
> And yes, I realize that you can already build big arrays and use LVM etc
> to make them be more than 16TB. I just do not think it's a problem yet,
> and I'd rather cater to "normal" people than to peopel who can't bother
> to partition their data at all.
Right, there's still a lot more scalability to be squeezed out of good old 32
bits. If we do run out of something it's likely to be the 32 bits of inodes,
after all, who can get by on a mere 4 billion files these days?
--
Daniel
Daniel Phillips wrote:
>On November 28, 2001 04:38 am, Linus Torvalds wrote:
>
>>We will probably _not_ get 64-bit page index numbers, though. I don't want
>>to make the page structure bigger/slower for very little gain. So the page
>>cache is probably going to be limited to about 44 bits (45+ if people
>>start doing large pages, which is probably worth it). So there would still
>>be partition/file limits on the order of 16-64 TB in the next few years.
>>
>
>The Ext2+ crowd is actively working on preparing for the world of larger
>blocks, so that 64KB blocks will be entirely practical. This will get us to
>1/4 Petabyte, still with 32 bit block pointers internally. The main problem
>that has to be solved is internal fragmentation.
>
This is a bad idea for general purpose usage patterns. The problem is
that if you cure the internal fragmentation problem like reiserfs did,
you are still going to be left with a memory bandwidth consumption that
is proportional to the size of nodes, because inserting an 8 byte item
is likely to cause a shift of contents proportional to node size.
We need to go to 64 bit blocknumbers and 64 bit inode numbers. If we
are not going to do this, well, I really need to change the design of
Reiser4, as I was just assuming it would be done. We get some nice
simplicities out of never reusing objectids (with 64 bits, you never run
out of them, the math for what it takes to run out of them assuming you
use a trillion a second is a bit amusing.)
I recommend the use of 64 bits for inode numbers and blocknumbers plus
the use of compression (effective for filesystems small enough to not
use all 64 bits) for those who dislike the byte wastage.
Hans
Nikita Danilov wrote:
>Hans Reiser writes:
>
>
>Having 64 bit inode numbers is good, but we *can* live without them:
>currently inode hash table can store inodes with identical inode
>numbers, provided they can be distinguished by find_actor. Inode numbers
>are just first fast guess during table scan, if they coincide,
>find_actor is used.
>
Nikita is entirely correct, 32 bits of inode number hash is plenty.
64k blocks remain a solution whose potential drain on memory bandwidth
worries me.
Hans