Hi Michael,
On Thu, 2025-03-27 at 17:06 +0100, Michael Trapp wrote:
> Use MHD_OPTION_SOCK_ADDR to bind the http listen socket to a single address.
> The address should be an IPv4 or IPv6 address configured on the system:
> --listen-address=127.0.0.1
> --listen-address=::1
> --listen-ad
Use MHD_OPTION_SOCK_ADDR to bind the http listen socket to a single address.
The address should be an IPv4 or IPv6 address configured on the system:
--listen-address=127.0.0.1
--listen-address=::1
--listen-address='LOCAL_IPv4|IPv6_ADDRESS'
As debuginfod does not include any security fea
Hi Michael,
On Thu, Mar 27, 2025 at 04:43:49PM +0100, Michael Trapp wrote:
> Use MHD_OPTION_SOCK_ADDR to bind the http listen socket to a single address.
> The address should be an IPv4 or IPv6 address configured on the system:
> --listen-address=127.0.0.1
> --listen-address=::1
> --li
Use MHD_OPTION_SOCK_ADDR to bind the http listen socket to a single address.
The address should be an IPv4 or IPv6 address configured on the system:
--listen-address=127.0.0.1
--listen-address=::1
--listen-address='LOCAL_IPv4|IPv6_ADDRESS'
As debuginfod does not include any security fea
Hi Frank,
On Thu, Feb 20, 2025 at 02:18:04PM -0500, Frank Ch. Eigler wrote:
> > [...]
> > This does sounds like a bug in glibc sscanf. I cannot find a
> > description of what exactly happens with 'm' modifier allocated
> > buffers on error. So I can imagine a double free if sscanf frees the
> > bu
Hi -
> [...]
> This does sounds like a bug in glibc sscanf. I cannot find a
> description of what exactly happens with 'm' modifier allocated
> buffers on error. So I can imagine a double free if sscanf frees the
> buffer on error. But returning a bogus pointer? That seems a bug. If
> we aren't gu
Hi Frank,
On Tue, Feb 18, 2025 at 10:30:44PM -0500, Frank Ch. Eigler wrote:
> Planning to commit this shortly:
>
> commit a71bac67f4705b84368b71f5ece54deedaa1abf1 (HEAD -> master1)
> Author: Frank Ch. Eigler
> Date: Tue Feb 18 22:09:12 2025 -0500
>
> debuginfod-client: correct invalid fre
Planning to commit this shortly:
commit a71bac67f4705b84368b71f5ece54deedaa1abf1 (HEAD -> master1)
Author: Frank Ch. Eigler
Date: Tue Feb 18 22:09:12 2025 -0500
debuginfod-client: correct invalid free() in failed ima path
debuginfod-find with a failed signature configuration was f
Hi Aaron,
On Fri, Jan 24, 2025 at 08:32:48PM -0500, Aaron Merey wrote:
> debuginfod_validate_imasig might call free on an uninitialized sig_buf
> due to a goto that can occur before sig_buf is set to NULL.
>
> Fix this by setting sig_buf to NULL before the goto.
The first thing after exit_valida
debuginfod_validate_imasig might call free on an uninitialized sig_buf
due to a goto that can occur before sig_buf is set to NULL.
Fix this by setting sig_buf to NULL before the goto.
Signed-off-by: Aaron Merey
---
debuginfod/debuginfod-client.c | 2 +-
1 file changed, 1 insertion(+), 1 deletio
"Frank Ch. Eigler"
Date: Sat, 7 Dec 2024 15:01:54 -0500
Subject: [PATCH] debuginfod: in --cors mode, add CORS response headers and
OPTIONS method
From: Henning Meyer
CORS is the Cross-Origin-Resource-Sharing mechanism explained at
https://developer.mozilla.org/en-US/docs/Web/HTTP/CORS
g
On 08.12.24 12:06, Florian Weimer wrote:
* Frank Ch. Eigler:
Hi -
I'm planning to commit this shortly:
From 4ebefc8f3b4b8fb68a55c960e70122fda50a0fb9 Mon Sep 17 00:00:00 2001
From: "Frank Ch. Eigler"
Date: Sat, 7 Dec 2024 15:01:54 -0500
Subject: [PATCH] debuginfod: add CORS respon
Hi -
> [...]
> I think it will allow public web clients to exfiltrate debuginfo data
> from debuginfod servers on private intranets. Previously, the
> cross-origin restrictions on web content would have prevented that.
Yes, this is the flip side of the CORS default coin. ISTM the
convenience is
* Frank Ch. Eigler:
> Hi -
>
> I'm planning to commit this shortly:
>
> From 4ebefc8f3b4b8fb68a55c960e70122fda50a0fb9 Mon Sep 17 00:00:00 2001
> From: "Frank Ch. Eigler"
> Date: Sat, 7 Dec 2024 15:01:54 -0500
> Subject: [PATCH] debuginfod: add CORS respon
Hi Henning,
On Fri, Dec 06, 2024 at 01:01:34AM +0100, Henning Meyer wrote:
> it is the Cross-Origin-Resource-Sharing mechanism explained at
> https://developer.mozilla.org/en-US/docs/Web/HTTP/CORS
Thanks, this might be a good URL to include in a comment so others
know what the extra headers are t
Hi -
I'm planning to commit this shortly:
>From 4ebefc8f3b4b8fb68a55c960e70122fda50a0fb9 Mon Sep 17 00:00:00 2001
From: "Frank Ch. Eigler"
Date: Sat, 7 Dec 2024 15:01:54 -0500
Subject: [PATCH] debuginfod: add CORS response headers and OPTIONS method
From: Henning Meyer
Hi Mark,
it is the Cross-Origin-Resource-Sharing mechanism explained at
https://developer.mozilla.org/en-US/docs/Web/HTTP/CORS 1. by default
JavaScript code from Website A cannot request arbitrary resources from
website B, these are called cross-origin-requests 2. The browser
performs what
Hi Frank, Hi Henning,
On Wed, 2024-12-04 at 19:40 -0500, Frank Ch. Eigler wrote:
> > This is my first attempt at implementing CORS support in debuginfod.
>
> Looks good to me really; will wait for a glance from others.
So the code looks correct if you want to handle the OPTION command and
add a
Hi -
> This is my first attempt at implementing CORS support in debuginfod.
Looks good to me really; will wait for a glance from others.
> I should probably add or change tests. [...]
I wouldn't overthink it - just add a "-i" to any random curl command
and look for ACAO:*.
I reconfigured the d
This is my first attempt at implementing CORS support in debuginfod.
I should probably add or change tests. Since debuginfod-find does not
need this functionality, it would be another test done with curl. I had
a look at the existing tests,
run-debuginfod-webapi-concurrency.sh looks like it wou
Hi -
On Fri, Nov 01, 2024 at 07:37:03PM +0300, Anton Moryakov wrote:
> Previously, the handle debuginfod_so created by dlopen was not closed
> in all cases, leading to a potential resource leak. This commit adds an
> additional dlclose call to ensure that the handle is always properly
> release
Previously, the handle debuginfod_so created by dlopen was not closed
in all cases, leading to a potential resource leak. This commit adds an
additional dlclose call to ensure that the handle is always properly
released, whether the symbols are successfully loaded or not.
Found by RASU JSC.
Si
Hi -
Proposed patch for a problem we spotted recently in unreleased code:
commit b853004e723d058f97362dbfdc40d1c49ea2ef51 (HEAD -> main)
Author: Frank Ch. Eigler
Date: Tue Sep 3 11:27:36 2024 -0400
debuginfod: service metadata queries in separate, timed-out connections
The --meta
Hi,
On Wed, Aug 21, 2024 at 03:32:59PM +0200, Mark Wielaard wrote:
> When doing a --enable-gcov build we link all binaries static.
> libdebuginfod.so now depends on crypto an jsonc. So also add those
> when linking against libdebuginfod.a
>
> debuginfod/Makefile.am (libdebuginfod): Add $(cr
When doing a --enable-gcov build we link all binaries static.
libdebuginfod.so now depends on crypto an jsonc. So also add those
when linking against libdebuginfod.a
debuginfod/Makefile.am (libdebuginfod): Add $(crypto_LIBS)
$(jsonc_LIBS) when BUILD_STATIC.
Signed-off-by: Mark Wielaar
Hi -
A few buildbots are unhappy with the debuginfod configury from
yesterday. It's right: it's rather convoluted. This patch
reworks and harmonizes all of that stuff.
Author: Frank Ch. Eigler
Date: Mon Jun 3 18:14:52 2024 -0400
rework debuginfod configury
Rework the top level
Hi Frank,
LGTM, please go ahead and merge the patch with these changes.
Aaron
On Sat, Jun 1, 2024 at 11:26 AM Frank Ch. Eigler wrote:
>
> Hi -
>
> > Thanks for the patch, some suggestions below.
>
> Thanks, adding the following 'git diff -w' to the patch you reviewed:
>
> > [...]
> > I was expe
Hi -
> Thanks for the patch, some suggestions below.
Thanks, adding the following 'git diff -w' to the patch you reviewed:
> [...]
> I was experimenting with metadata queries to a local server that indexed
> a single rpm (binutils-2.41-8.fc40.x86_64.rpm). The following commands
> produced JSON w
Khem Raj writes:
> Recent commit acd9525e9 has removed all references to max_fds
> therefore remove it, moreover clang18 is happier
Thanks, merged as obvious.
- FChE
Recent commit acd9525e9 has removed all references to max_fds
therefore remove it, moreover clang18 is happier
| ../../elfutils-0.191/debuginfod/debuginfod.cxx:1448:8: error: private field
'max_fds' is not used [-Werror,-Wunused-private-field]
| 1448 | long max_fds;
| |^
| 1 erro
commit 06e0aacb4b288403bd02a0820dd0f87c7f017a2b (HEAD -> master)
Author: Frank Ch. Eigler
Date: Tue Nov 14 14:09:40 2023 -0500
doc/debuginfod.8: clarify source file handling
Added text about the archive common-prefix heuristic, mentioned
the new eu-srcfiles tool, and gave som
libcurl passes an empty line to header_callback indicating the end
of the response's HTTP headers.
Currently this empty line is printed to the debuginfod_client's
verbose_fd with a "header" prefix:
$ echo $DEBUGINFOD_URLS
https://debuginfod.fedoraproject.org/
$ debuginfod-find -vv deb
Hi -
This patch, mostly the work of Ryan Goldberg, has undergone some
decent testing alongside its systemtap consumer (which is already
merged). If there's interest, I could install it speculatively on one
of the fedora debuginfod servers, the one that's already running its
prerequisite RPM-IMA p
Committing as obvious.
sscanf is used to get the value of x-debuginfod-size from the http
headers. The format string used assumes that the header field name
is entirely lower case. However mixed-case field names are possible,
resulting in the value not being read.
Fix this by removing "x-debugi
On Wed, Mar 29, 2023 at 10:57:47PM +0800, lilydjwg wrote:
> On Wed, Mar 29, 2023 at 08:28:35AM -0400, Frank Ch. Eigler wrote:
> > Hi -
> >
> > > [...]
> > > The reason is that when Content-Length is unavailable, cl is set to -1
> > > by curl
> >
> > Is that behaviour from new versions of curl?
>
On Wed, Mar 29, 2023 at 08:28:35AM -0400, Frank Ch. Eigler wrote:
> Hi -
>
> > [...]
> > The reason is that when Content-Length is unavailable, cl is set to -1
> > by curl
>
> Is that behaviour from new versions of curl?
Yes. curl 8.0.1 to be exact.
> > but dl_size remains as 0, but later dl_s
Hi -
> [...]
> The reason is that when Content-Length is unavailable, cl is set to -1
> by curl
Is that behaviour from new versions of curl?
> but dl_size remains as 0, but later dl_size == -1 is checked.
Or perhaps dl_size needs to be initialized to -1 ("unknown") vs 0
("known to be zero"), a
From: lilydjwg
Currently when the debuginfod server doesn't set Content-Length but set
x-debuginfod-size header, it's ignored and the progressfn doesn't know
the total size.
This happens for me with Arch Linux's debuginfod server and gdb.
The reason is that when Content-Length is unavailable, c
Committing as obvious.
Before attempting to download a section, cache_find_section tries to
extract the section from existing files in the cache. If it's determined
that the section must not exist, cache_find_section returns -ENOENT to
indicate that the download should be skipped.
This patch fixe
doc/ChangeLog:
* debuginfod-find.1: add a note to DESCRIPTION section for the 'source'
subcommand, clarifying where to find the CU compilation-directory.
I'm looking at how to improve and document the workflow for using
debuginfod-find to study the behaviour of packaged software on a system, e.
Hi,
On Tue, 2022-11-15 at 17:55 +0100, Mark Wielaard wrote:
> On error going to out2, the response_data is freed. So initialize the
> response_data to NULL immediately after allocation or when going back
> to query_in_parallel.
Frank on irc said this looked fine.
Pushed,
Mark
On error going to out2, the response_data is freed. So initialize the
response_data to NULL immediately after allocation or when going back
to query_in_parallel.
Signed-off-by: Mark Wielaard
---
debuginfod/ChangeLog | 5 +
debuginfod/debuginfod-client.c | 4 ++--
2 files changed, 7
Hi -
Showing diff -w to omit reindentation whitespace noise. Maybe worth
backporting to 0.188 releases, could affect federation intermediate
servers with burst workload.
commit ec166cf3c8d825a2f02aca448a0823de12e78991 (HEAD -> master)
Author: Frank Ch. Eigler
Date: Thu Nov 3 10:07:31 2022 -0
Hi Daniel,
On Tue, 2022-10-18 at 14:21 -0700, Daniel Thornburgh via Elfutils-devel wrote:
> This DEBUGINFOD_HEADERS_FILE environment variable names a file to supply
> HTTP headers to outgoing requests. Notably, this allows for
> Authorization headers to be added from a file under OS access control
Hi Mark,
On Wed, Oct 26, 2022 at 11:06 AM Mark Wielaard wrote:
> On Mon, 2022-10-24 at 14:38 -0400, Frank Ch. Eigler via Elfutils-devel
> wrote:
> > - not sure I understand why the code worries about dots in or not in
> > section names. Why not just pass them verbatim throughout the code
> >
Hi Frank,
On Mon, Oct 24, 2022 at 2:38 PM Frank Ch. Eigler wrote:
> - use of write(2) to put files onto disk is not quite right; write(2) can
> be partial, so you need a loop (or a macro wrapping a loop)
Fixed.
> - not sure I understand why the code worries about dots in or not in
> section
Hi -
> Is/should the section name be URL-encoded?
Yes!
> I would drop the maybe_debuginfo_section heuristics. There are some
> sections like .strtab/.symtab that are probably in the debug file, but
> might be in the executable. I would assume that a named section can
> normally be found in the d
Hi,
On Mon, 2022-10-24 at 14:38 -0400, Frank Ch. Eigler via Elfutils-devel
wrote:
> - use of write(2) to put files onto disk is not quite right; write(2)
> can
> be partial, so you need a loop (or a macro wrapping a loop)
Since debuginfod-client.c already includes system.h it can use:
static i
Hi,
On Mon, 2022-10-24 at 14:38 -0400, Frank Ch. Eigler via Elfutils-devel
wrote:
> - not sure I understand why the code worries about dots in or not in
> section names. Why not just pass them verbatim throughout the code
> base, and not worry about whether or not there's a dot? Does the
>
Hi -
Generally looks fine, thanks a lot.
A few nits:
- use of write(2) to put files onto disk is not quite right; write(2) can
be partial, so you need a loop (or a macro wrapping a loop)
- not sure I understand why the code worries about dots in or not in
section names. Why not just pass th
I'm resending this patch with a small modification. I added a new
field "progressfn_cancel" to debuginfod_client that indicates whether
the most recent query was cancelled due to progressfn returning 1.
If a server doesn't support section queries and the client begins
downloading a debuginfo or e
This DEBUGINFOD_HEADERS_FILE environment variable names a file to supply
HTTP headers to outgoing requests. Notably, this allows for
Authorization headers to be added from a file under OS access control.
Signed-off-by: Daniel Thornburgh
---
NEWS| 3 +++
debuginfod/Ch
Hi Frank,
On Wed, Sep 28, 2022 at 10:28 AM Frank Ch. Eigler wrote:
> On Tue, Sep 27, 2022 at 10:10:52PM -0400, Aaron Merey via Elfutils-devel
> wrote:
>
> > [...] In order to distinguish between debuginfo and executable
> > files with the same build-id, this function includes a bool
> > paramet
Hi, Aaron -
On Tue, Sep 27, 2022 at 10:10:52PM -0400, Aaron Merey via Elfutils-devel wrote:
> [...] In order to distinguish between debuginfo and executable
> files with the same build-id, this function includes a bool
> parameter use_debuginfo. If true, attempt to retrieve the section
> from t
Committing as obvious:
Switch a couple error codes from positive to negative so they aren't
interpreted as file descriptors by the caller.
Signed-off-by: Aaron Merey
---
debuginfod/ChangeLog | 5 +
debuginfod/debuginfod-client.c | 4 ++--
2 files changed, 7 insertions(+), 2 deleti
This patch adds a new function debuginfod_find_section which queries
debuginfod servers for the binary contents of the specified ELF/DWARF
section in a file matching the given build-id.
In order to distinguish between debuginfo and executable files with the
same build-id, this function includes a
Pushed, thanks for the server!
Martin
On 9/15/22 03:26, Sergio Durigan Junior via Elfutils-devel wrote:
> Signed-off-by: Sergio Durigan Junior
> ---
> Debuginfod.html | 8
> 1 file changed, 8 insertions(+)
>
> diff --git a/Debuginfod.html b/Debuginfod.html
> index 64fef86c..71bc7c9b 1
Signed-off-by: Sergio Durigan Junior
---
Debuginfod.html | 8
1 file changed, 8 insertions(+)
diff --git a/Debuginfod.html b/Debuginfod.html
index 64fef86c..71bc7c9b 100644
--- a/Debuginfod.html
+++ b/Debuginfod.html
@@ -170,6 +170,14 @@
all
+
Hi -
> [...]
> Fix this by using an auto-sized thread pool when '-C' is not given, just
> as we do when it's given with no argument. Update the doc's description
> of '-C'.
Thanks, pushed.
- FChE
Since commit 4b42d9ad, libmicrohttpd's epoll event loop is used when
available in which case we must disable its setting for spawning a thread
per request. This contradicts the debuginfod doc's description of '-C',
which indicates that if this command line option is not given then the
thread pool
Hi -
> When dwarf_extract_source_paths is called, it can call handle_buildid
> when a rpm file used dwz. Ignore such internal request in
> http_requests_total statistics.
Noble goal:
> @@ -1906,7 +1906,7 @@ handle_buildid (MHD_Connection* conn,
> const string& buildid /* unsafe
Hi -
> Report the file that has such a problem so that one can inspect it.
> Signed-off-by: Martin Liska
The complication with this type of change is that the exception text
objects also make it into the prometheus metrics. When file names and
such variables show up, they will bifurcate all the
When dwarf_extract_source_paths is called, it can call handle_buildid
when a rpm file used dwz. Ignore such internal request in
http_requests_total statistics.
Signed-off-by: Martin Liska
---
debuginfod/debuginfod.cxx | 10 ++
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a
Report the file that has such a problem so that one can inspect it.
Signed-off-by: Martin Liska
---
debuginfod/debuginfod.cxx | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/debuginfod/debuginfod.cxx b/debuginfod/debuginfod.cxx
index 9245be53..67683354 100644
--- a/debugi
As explained in detail in the PR, the problem happens when multiple
requests lead to a single RPM and one of the threads does prefetching.
In that case, the requested file is added to fdcache and thus skip
as interned.
---
debuginfod/debuginfod.cxx | 69 ---
1 f
Hi -
> Subject: [PATCH] debuginfod: optimize regular expressions in groom()
Thanks, pushed.
- FChE
>From 2b377704f46081f2348dfc5650820ac9b5485758 Mon Sep 17 00:00:00 2001
From: Josef Cejka
Date: Wed, 3 Aug 2022 13:33:21 +0200
Subject: [PATCH] debuginfod: optimize regular expressions in groom()
Check if applying of -I and -X during grooming is enabled
before the regular expressions are matc
Hi -
> --disable-source-scan disables scanning of the dwarf source info
> of debuginfo sections. The source info is not required in setups
> without source code access.
Thanks, merged (with a bit of ChangeLoggery added afterwards).
- FChE
--disable-source-scan disables scanning of the dwarf source info
of debuginfo sections. The source info is not required in setups
without source code access.
Signed-off-by: Michael Trapp
---
This option should save some scan cycles and DB space. If there
is no access to source code, the source i
Hi,
On Mon, May 09, 2022 at 11:10:44PM +0200, Mark Wielaard wrote:
> When getting the connection info getnameinfo is called getting the
> hostname and servname except when the sockaddr is a pure ipv6
> address. In that last case only hostname is requested. Since servname
> is stack allocated and n
Hi,
On Mon, May 09, 2022 at 06:22:29PM +0200, Mark Wielaard wrote:
> curl_easy_setup can fail for various reasons. Add a curl_easy_setopt_ck
> macro to check all curl_easy_setopt calls and provides a human readable
> error message in verbose mode.
Pushed.
Cheers,
Mark
On Fri, May 06, 2022 at 12:38:40AM +0200, Mark Wielaard wrote:
> Also disable MHD_USE_THREAD_PER_CONNECTION when using MHD_USE_EPOLL.
Pushed after OK from Frank on irc and some testing in Fedora.
Cheers,
Mark
Hi,
On Fri, May 06, 2022 at 12:37:35AM +0200, Mark Wielaard wrote:
> On a systems that have ipv6 disabled debuginfod doesn't start up
> anymore because libhttpd MHD_USE_DUAL_STACK only works if it can
> open an ipv6 socket. If MHD_start_daemon with MHD_USE_DUAL_STACK
> fails try again without that
When getting the connection info getnameinfo is called getting the
hostname and servname except when the sockaddr is a pure ipv6
address. In that last case only hostname is requested. Since servname
is stack allocated and not initialized it might contain garbage which
is then put in the log. Just a
curl_easy_setup can fail for various reasons. Add a curl_easy_setopt_ck
macro to check all curl_easy_setopt calls and provides a human readable
error message in verbose mode.
Signed-off-by: Mark Wielaard
---
debuginfod/ChangeLog | 5 +++
debuginfod/debuginfod-client.c | 61 +++
This was the only place in debuginfod-client.c where we didn't check
the result of curl_easy_getinfo. Just check it to make things consistent.
Signed-off-by: Mark Wielaard
---
debuginfod/ChangeLog | 5 +
1 file changed, 5 insertions(+)
diff --git a/debuginfod/ChangeLog b/debuginfod/ChangeLo
Also disable MHD_USE_THREAD_PER_CONNECTION when using MHD_USE_EPOLL.
https://sourceware.org/bugzilla/show_bug.cgi?id=29123
Signed-off-by: Mark Wielaard
---
debuginfod/ChangeLog | 6 ++
debuginfod/debuginfod.cxx | 15 ++-
2 files changed, 16 insertions(+), 5 deletions(-)
d
On a systems that have ipv6 disabled debuginfod doesn't start up
anymore because libhttpd MHD_USE_DUAL_STACK only works if it can
open an ipv6 socket. If MHD_start_daemon with MHD_USE_DUAL_STACK
fails try again without that flag set.
https://sourceware.org/bugzilla/show_bug.cgi?id=29122
Signed-of
Hi Frank,
On Fri, Apr 22, 2022 at 7:27 PM Frank Ch. Eigler wrote:
> > - add_mhd_response_header (r, "X-DEBUGINFOD-SIZE",
> > - to_string(fs.st_size).c_str());
>
> > + rc = fstat (fd, &fs);
> > + if (rc == 0)
> > +add_mhd_res
On Sun, Apr 24, 2022 at 11:05 AM Mark Wielaard wrote:
> Looks good. Pleas commit.
Thanks, pushed as:
commit 55fee962676fbff60c6b0469305bcb077910d64f
Author: Aaron Merey
Date: Tue Jan 11 22:07:55 2022 -0500
debuginfod: Use the debuginfod-size response header
In some cases the content
On Fri, Apr 22, 2022 at 11:53:43PM +0200, Mark Wielaard wrote:
> We used to go out of our way to initialize libcurl early before any other
> thread/code was running. But this meant that we might pay startup cost,
> which under FIPS is significant, even for code that never uses libdebuginfod
> or TL
Hi Aaron,
On Fri, Apr 22, 2022 at 06:56:41PM -0400, Aaron Merey via Elfutils-devel wrote:
> I've updated the patch below with the changes Mark recommended.
>
> A couple X-DEBUGINFOD-SIZE tests were added in another patch I recently
> posted [1] that also fixes a bug when computing this header's v
I've updated the patch below with the changes Mark recommended.
A couple X-DEBUGINFOD-SIZE tests were added in another patch I recently
posted [1] that also fixes a bug when computing this header's value for
an archived file.
Aaron
[1] https://sourceware.org/pipermail/elfutils-devel/2022q2/0049
We used to go out of our way to initialize libcurl early before any other
thread/code was running. But this meant that we might pay startup cost,
which under FIPS is significant, even for code that never uses libdebuginfod
or TLS libcurl connections. Although curl_global_init itself isn't thread-sa
On Thu, Mar 31, 2022 at 1:44 PM Mark Wielaard wrote:
> Just a question about this part:
>
> > + /* If Content-Length is -1, try to get the size from
> > + X-Debuginfod-Size */
> > + if (dl_size == -1 && c->winning_headers != NULL)
> > +{
> > +
Hi -
> That IN6_IS_ADDR_V4MAPPED is funny. Is that actually used in practice?
Yeah, actually all the time. On a dual-stack socket, an ipv4 connection
gets an ipv6 peer-address that renders like :::1.2.3.4.
> Too bad this cannot be merged with conninfo above.
> But this calculates less inf
Hi Frank,
On Sun, 2022-04-03 at 19:49 -0400, Frank Ch. Eigler via Elfutils-devel wrote:
> commit 3a00f412b6554ba70f7b7e3d3aa6f67f278868af (HEAD -> master)
> Author: Frank Ch. Eigler
> Date: Sun Apr 3 19:42:48 2022 -0400
>
> debuginfod: use single ipv4+ipv6 microhttpd daemon configuration
>
Hi -
This little improvement reduces wasted memory from duplicated
worker threads for ipv4 vs ipv6 stacks.
commit 3a00f412b6554ba70f7b7e3d3aa6f67f278868af (HEAD -> master)
Author: Frank Ch. Eigler
Date: Sun Apr 3 19:42:48 2022 -0400
debuginfod: use single ipv4+ipv6 microhttpd daemon conf
Hi Aaron,
On Tue, 2022-01-11 at 22:07 -0500, Aaron Merey via Elfutils-devel
wrote:
> In some cases the content-length header may not be available in order
> to pass to a progressfn. If content-length isn't available then attempt
> to get the size of the download from the debuginfod-size header in
commit c4726913b7766634cc5d8a97624e1b4708452d73 (HEAD -> master)
Author: Frank Ch. Eigler
Date: Mon Jan 31 18:13:40 2022 -0500
man debuginfod-client-config.7: Elaborate $DEBUGINFOD_URLS
Add reference to /etc/profile.d and /etc/debuginfod/*.urls as possible
source of default. (
In some cases the content-length header may not be available in order
to pass to a progressfn. If content-length isn't available then attempt
to get the size of the download from the debuginfod-size header instead.
It should be mentioned that if a compressed file (ex. gzip) is being
transferred,
Hi,
On Thu, Dec 09, 2021 at 06:35:14AM -0500, Frank Ch. Eigler via Elfutils-devel
wrote:
> > x32, riscv32, arc use 64bit time_t even while they are 32bit
> > architectures, therefore directly using integer printf formats will not
> > work portably.
>
> > Use a plain long everywhere as the interv
Hi -
> x32, riscv32, arc use 64bit time_t even while they are 32bit
> architectures, therefore directly using integer printf formats will not
> work portably.
> Use a plain long everywhere as the intervals are small enough
> that it will not be problematic.
lgtm!
- FChE
Alright, I just sent a patch that replaces time_t with long instead. I
tested it on x86, x86-x32, x86-64.
Alex
On Wed, 8 Dec 2021 at 16:31, Mark Wielaard wrote:
> Hi Alexander,
>
> On Sun, 2021-12-05 at 21:45 +0100, Alexander Kanavin wrote:
> > I'm not sure; the point of this patch is simply to
time_t is platform dependent and some of architectures e.g.
x32, riscv32, arc use 64bit time_t even while they are 32bit
architectures, therefore directly using integer printf formats will not
work portably.
Use a plain long everywhere as the intervals are small enough
that it will not be problema
Hi -
> [...]
> We seem to not expect these intervals to be much bigger than a week,
> so an int should always be big enough (even when stretched up to a
> whole year).
Yes, ints are fine for these humane-number-of-seconds kinds of values
in the cache configuration. There's no need for maximum le
On Sat, Dec 04, 2021 at 10:33:35PM +0100, Mark Wielaard wrote:
> gcc address sanitizer detected a dangling debuginfod_client handler
> when debuginfod exits. Make sure to groom the debuginfod client pool
> before exit after all threads are done.
Pushed,
Mark
Hi -
Committing as obvious.
Author: Frank Ch. Eigler
Date: Wed Dec 8 10:20:58 2021 -0500
debuginfod: correct concurrency bug in fdcache metrics
The intern() function called set_metrics() outside a necessary lock
being held. helgrind identified this race condition. No QA imp
Hi Alexander,
On Sun, 2021-12-05 at 21:45 +0100, Alexander Kanavin wrote:
> I'm not sure; the point of this patch is simply to ensure debuginfod builds
> everywhere without errors. Making the types consistent is perhaps better
> done as a followup?
I think the issue of the code not compiling in s
Hi Frank,
On Mon, 2021-12-06 at 13:41 -0500, Frank Ch. Eigler wrote:
> gcc address sanitizer detected a leak of the debuginfod_client
> > winning_headers when the handle was reused. Make sure to free and
> > reset the winning_headers field before reuse.
>
> This is good.
Thanks, pushed.
Cheers,
1 - 100 of 244 matches
Mail list logo
