Pradeep Satyanarayana wrote:
Roland Dreier wrote:
Maybe we should use the /proc/pid/pagemap and /proc/kpageflags files?
These files let a userspace process find out which physical frame
each virtual page is mapped to and to get the properties of each page
frame including the page
Roland Dreier Wrote:
In order to avoid adding additional dependency to libibverbs, maybe
we should just to enhance the get_huge_page_size() so it will support
multiple huge page sizes?
I think that does make sense. However see my reply to Alexander
Schmidt -- maybe there is something
Maybe we should use the /proc/pid/pagemap and /proc/kpageflags files?
These files let a userspace process find out which physical frame
each virtual page is mapped to and to get the properties of each page
frame including the page size.
Sounds interesting... the only problem is that
Roland Dreier wrote:
Maybe we should use the /proc/pid/pagemap and /proc/kpageflags files?
These files let a userspace process find out which physical frame
each virtual page is mapped to and to get the properties of each page
frame including the page size.
Sounds interesting... the
In order to avoid adding additional dependency to libibverbs, maybe
we should just to enhance the get_huge_page_size() so it will support
multiple huge page sizes?
I think that does make sense. However see my reply to Alexander
Schmidt -- maybe there is something clever we can do with
Roland Dreier Wrote:
ibv_reg_mr() fails to register a memory region allocated on huge page and
not
the default page size. This happens because ibv_madvise_range() aligns
memory
region to the default system page size before calling to madvise() which
fails
with EINVAL error.
But we couldn't find another one without changing the API or changing
the kernel.
Changing which API? Changing the kernel is a possibility of course.
- R.
--
To unsubscribe from this list: send the line unsubscribe linux-rdma in
the body of a message to majord...@vger.kernel.org
More
ibv_reg_mr() fails to register a memory region allocated on huge page and not
the default page size. This happens because ibv_madvise_range() aligns memory
region to the default system page size before calling to madvise() which
fails
with EINVAL error. madvise() fails because it
Hi Eli,
Sorry fot the late response.
If strtol() fails, you may return with an invalid, non zero value for
huge page size. Maybe use some temporary variable to hold the
intermediate result.
If strtol() fails, for any reason, we always return value = 0 for huge page
size.
This indicates that
Hi Roland,
I would like to ask you if you have an estimation for when you will review the
patches below?
Thanks,
AlexV
Alex Vainman wrote:
Hi Roland,
Have you had a chance to look at this patch and at the patch
I've sent with this email:
Subject: [PATCH] libibverbs: Add huge page
On Sun, Nov 29, 2009 at 07:08:08PM +0200, Alex Vainman wrote:
+ p_hpage_val = strstr(buf, label);
+ if (!p_hpage_val) {
+ errno = EINVAL;
+ return -1;
+ }
+ p_hpage_val += strlen(label);
+
+ errno = 0;
+ ret_val = strtol(p_hpage_val,
Hi Roland,
Have you had a chance to look at this patch and at the patch
I've sent with this email:
Subject: [PATCH] libibverbs: Add huge page support to ibv_madvise_range()
Sent: Sun, 29 Nov 2009 19:08:08 +0200?
Thanks,
Alexv
Alex Vainman wrote:
ibv_reg_mr() fails to register a memory region
12 matches
Mail list logo