Bug#1017426: libnss3: Uninitialised value was created by a stack allocation
On Tue, Aug 16, 2022 at 10:17 AM Mike Hommey wrote: > On Tue, Aug 16, 2022 at 09:59:30AM +0300, Alexey Kuznetsov wrote: > > On Tue, Aug 16, 2022 at 9:50 AM Mike Hommey wrote: > > > > > On Tue, Aug 16, 2022 at 09:06:20AM +0300, Alexey Kuznetsov wrote: > > > > On Tue, Aug 16, 2022 at 9:00 AM Mike Hommey wrote: > > > > > > > > > On Tue, Aug 16, 2022 at 08:30:07AM +0300, a...@me.com wrote: > > > > > > Package: libnss3 > > > > > > Version: 2:3.79-1 > > > > > > Severity: normal > > > > > > > > > > > > Dear Maintainer, > > > > > > > > > > > > debuging valgrind pidgin with result: > > > > > > > > > > > > ==804198== Uninitialised value was created by a stack allocation > > > > > > ==804198==at 0xB089DC0: ssl3_MACEncryptRecord > (ssl3con.c:2104) > > > > > > > > > > > > line correspopnds to the ssl3_MACEncryptRecord > > > > > > > > > > Looking at the code, it would seem to be a false positive, but I > might > > > > > have overlooked something, but you haven't pasted the most > interesting > > > > > parts of the valgrind output... > > > > > > > > > > Mike > > > > > > > > > > > > > This output comes exactly from valgrind. No usual stack trace. > Before and > > > > below are different issues. > > > > > > > > BTW pidgin crashing sometimes, and only issues I can record points > to the > > > > nss library. > > > > > > Usually, "Uninitialised value was created by a stack allocation" is the > > > reason for the error, with a stack trace, that comes above it. That's > > > the most crucial information. Without that, we don't know what is > trying > > > to use that unitialized value. > > > > > > > Ok .Let me restart pidgin. It 100% reproducible. Only thing you need is > to > > install dbgsym for glibc, nss3, pidgin and add frew irc and jabber > accounts > > (I also using matrix plugin). Command would be: > > > > G_SLICE=always-malloc valgrind --num-callers=30 --track-origins=yes > pidgin > > 2>&1 | tee 123.log > > > > https://paste.debian.net/1250580/ > > Can you reproduce with 3.81-1 in unstable? > > For posterity, the useful information: > > ==837133== Syscall param socketcall.sendto(msg) points to uninitialised > byte(s) > ==837133==at 0x5A153D6: __libc_send (send.c:28) > ==837133==by 0x5A153D6: send (send.c:23) > ==837133==by 0xB083527: pt_Send (ptio.c:2002) > ==837133==by 0xB01DFF7: ssl_DefSend (ssldef.c:105) > ==837133==by 0xB0229C0: ssl_SendSavedWriteData (sslsecur.c:452) > ==837133==by 0xB006839: ssl3_SendRecord (ssl3con.c:2568) > ==837133==by 0xB006C2C: ssl3_FlushHandshakeMessages (ssl3con.c:2774) > ==837133==by 0xB006C2C: ssl3_FlushHandshake (ssl3con.c:2747) > ==837133==by 0xB00F5E4: ssl3_SendFinished (ssl3con.c:11944) > ==837133==by 0xB00FB79: ssl3_SendClientSecondRound (ssl3con.c:8191) > ==837133==by 0xB011A7A: ssl3_HandleServerHelloDone (ssl3con.c:8061) > ==837133==by 0xB011A7A: ssl3_HandlePostHelloHandshakeMessage > (ssl3con.c:12568) > ==837133==by 0xB011A7A: ssl3_HandleHandshakeMessage (ssl3con.c:12479) > ==837133==by 0xB014A74: ssl3_HandleHandshake (ssl3con.c:12653) > ==837133==by 0xB014A74: ssl3_HandleNonApplicationData (ssl3con.c:13188) > ==837133==by 0xB0153C0: ssl3_HandleRecord (ssl3con.c:13529) > ==837133==by 0xB01B500: ssl3_GatherCompleteHandshake (ssl3gthr.c:561) > ==837133==by 0xB01B500: ssl3_GatherCompleteHandshake (ssl3gthr.c:449) > ==837133==by 0xB022A80: SSL_ForceHandshake (sslsecur.c:382) > ==837133==by 0xADCC8D6: ssl_nss_handshake_cb (ssl-nss.c:371) > ==837133==by 0x1824B1: pidgin_io_invoke (gtkeventloop.c:73) > ==837133==by 0x54BBA9E: g_main_dispatch (gmain.c:3417) > ==837133==by 0x54BBA9E: g_main_context_dispatch (gmain.c:4135) > ==837133==by 0x54BBE57: g_main_context_iterate.constprop.0 > (gmain.c:4211) > ==837133==by 0x54BC10E: g_main_loop_run (gmain.c:4411) > ==837133==by 0x4C57B29: gtk_main (in > /usr/lib/x86_64-linux-gnu/libgtk-x11-2.0.so.0.2400.33) > ==837133==by 0x145D7B: main (gtkmain.c:948) > ==837133== Address 0x1b82e246 is 534 bytes inside a block of size 1,553 > alloc'd > ==837133==at 0x484582F: realloc (vg_replace_malloc.c:1437) > ==837133==by 0xB2114A1: PORT_Realloc_Util (secport.c:101) > ==837133==by 0xB01E1E4: sslBuffer_Grow (sslencode.c:50) > ==837133==by 0xB01E1E4: sslBuffer_Grow (sslencode.c:31) > ==837133==by 0xB01E42B: sslBuffer_Append (sslencode.c:82) > ==837133==by 0xB006817: ssl3_SendRecord (ssl3con.c:2559) > ==837133==by 0xB006C2C: ssl3_FlushHandshakeMessages (ssl3con.c:2774) > ==837133==by 0xB006C2C: ssl3_FlushHandshake (ssl3con.c:2747) > ==837133==by 0xB00F5E4: ssl3_SendFinished (ssl3con.c:11944) > ==837133==by 0xB00FB79: ssl3_SendClientSecondRound (ssl3con.c:8191) > ==837133==by 0xB011A7A: ssl3_HandleServerHelloDone (ssl3con.c:8061) > ==837133==by 0xB011A7A: ssl3_HandlePostHelloHandshakeMessage > (ssl3con.c:12568) > ==837133==by 0xB011A7A: ssl3_HandleHandshakeMessage (ssl3con.c:12479) > ==837133==
Bug#1017426: libnss3: Uninitialised value was created by a stack allocation
On Tue, Aug 16, 2022 at 09:59:30AM +0300, Alexey Kuznetsov wrote: > On Tue, Aug 16, 2022 at 9:50 AM Mike Hommey wrote: > > > On Tue, Aug 16, 2022 at 09:06:20AM +0300, Alexey Kuznetsov wrote: > > > On Tue, Aug 16, 2022 at 9:00 AM Mike Hommey wrote: > > > > > > > On Tue, Aug 16, 2022 at 08:30:07AM +0300, a...@me.com wrote: > > > > > Package: libnss3 > > > > > Version: 2:3.79-1 > > > > > Severity: normal > > > > > > > > > > Dear Maintainer, > > > > > > > > > > debuging valgrind pidgin with result: > > > > > > > > > > ==804198== Uninitialised value was created by a stack allocation > > > > > ==804198==at 0xB089DC0: ssl3_MACEncryptRecord (ssl3con.c:2104) > > > > > > > > > > line correspopnds to the ssl3_MACEncryptRecord > > > > > > > > Looking at the code, it would seem to be a false positive, but I might > > > > have overlooked something, but you haven't pasted the most interesting > > > > parts of the valgrind output... > > > > > > > > Mike > > > > > > > > > > This output comes exactly from valgrind. No usual stack trace. Before and > > > below are different issues. > > > > > > BTW pidgin crashing sometimes, and only issues I can record points to the > > > nss library. > > > > Usually, "Uninitialised value was created by a stack allocation" is the > > reason for the error, with a stack trace, that comes above it. That's > > the most crucial information. Without that, we don't know what is trying > > to use that unitialized value. > > > > Ok .Let me restart pidgin. It 100% reproducible. Only thing you need is to > install dbgsym for glibc, nss3, pidgin and add frew irc and jabber accounts > (I also using matrix plugin). Command would be: > > G_SLICE=always-malloc valgrind --num-callers=30 --track-origins=yes pidgin > 2>&1 | tee 123.log > > https://paste.debian.net/1250580/ Can you reproduce with 3.81-1 in unstable? For posterity, the useful information: ==837133== Syscall param socketcall.sendto(msg) points to uninitialised byte(s) ==837133==at 0x5A153D6: __libc_send (send.c:28) ==837133==by 0x5A153D6: send (send.c:23) ==837133==by 0xB083527: pt_Send (ptio.c:2002) ==837133==by 0xB01DFF7: ssl_DefSend (ssldef.c:105) ==837133==by 0xB0229C0: ssl_SendSavedWriteData (sslsecur.c:452) ==837133==by 0xB006839: ssl3_SendRecord (ssl3con.c:2568) ==837133==by 0xB006C2C: ssl3_FlushHandshakeMessages (ssl3con.c:2774) ==837133==by 0xB006C2C: ssl3_FlushHandshake (ssl3con.c:2747) ==837133==by 0xB00F5E4: ssl3_SendFinished (ssl3con.c:11944) ==837133==by 0xB00FB79: ssl3_SendClientSecondRound (ssl3con.c:8191) ==837133==by 0xB011A7A: ssl3_HandleServerHelloDone (ssl3con.c:8061) ==837133==by 0xB011A7A: ssl3_HandlePostHelloHandshakeMessage (ssl3con.c:12568) ==837133==by 0xB011A7A: ssl3_HandleHandshakeMessage (ssl3con.c:12479) ==837133==by 0xB014A74: ssl3_HandleHandshake (ssl3con.c:12653) ==837133==by 0xB014A74: ssl3_HandleNonApplicationData (ssl3con.c:13188) ==837133==by 0xB0153C0: ssl3_HandleRecord (ssl3con.c:13529) ==837133==by 0xB01B500: ssl3_GatherCompleteHandshake (ssl3gthr.c:561) ==837133==by 0xB01B500: ssl3_GatherCompleteHandshake (ssl3gthr.c:449) ==837133==by 0xB022A80: SSL_ForceHandshake (sslsecur.c:382) ==837133==by 0xADCC8D6: ssl_nss_handshake_cb (ssl-nss.c:371) ==837133==by 0x1824B1: pidgin_io_invoke (gtkeventloop.c:73) ==837133==by 0x54BBA9E: g_main_dispatch (gmain.c:3417) ==837133==by 0x54BBA9E: g_main_context_dispatch (gmain.c:4135) ==837133==by 0x54BBE57: g_main_context_iterate.constprop.0 (gmain.c:4211) ==837133==by 0x54BC10E: g_main_loop_run (gmain.c:4411) ==837133==by 0x4C57B29: gtk_main (in /usr/lib/x86_64-linux-gnu/libgtk-x11-2.0.so.0.2400.33) ==837133==by 0x145D7B: main (gtkmain.c:948) ==837133== Address 0x1b82e246 is 534 bytes inside a block of size 1,553 alloc'd ==837133==at 0x484582F: realloc (vg_replace_malloc.c:1437) ==837133==by 0xB2114A1: PORT_Realloc_Util (secport.c:101) ==837133==by 0xB01E1E4: sslBuffer_Grow (sslencode.c:50) ==837133==by 0xB01E1E4: sslBuffer_Grow (sslencode.c:31) ==837133==by 0xB01E42B: sslBuffer_Append (sslencode.c:82) ==837133==by 0xB006817: ssl3_SendRecord (ssl3con.c:2559) ==837133==by 0xB006C2C: ssl3_FlushHandshakeMessages (ssl3con.c:2774) ==837133==by 0xB006C2C: ssl3_FlushHandshake (ssl3con.c:2747) ==837133==by 0xB00F5E4: ssl3_SendFinished (ssl3con.c:11944) ==837133==by 0xB00FB79: ssl3_SendClientSecondRound (ssl3con.c:8191) ==837133==by 0xB011A7A: ssl3_HandleServerHelloDone (ssl3con.c:8061) ==837133==by 0xB011A7A: ssl3_HandlePostHelloHandshakeMessage (ssl3con.c:12568) ==837133==by 0xB011A7A: ssl3_HandleHandshakeMessage (ssl3con.c:12479) ==837133==by 0xB014A74: ssl3_HandleHandshake (ssl3con.c:12653) ==837133==by 0xB014A74: ssl3_HandleNonApplicationData (ssl3con.c:13188) ==837133==by 0xB0153C0: ssl3_HandleRecord (ssl3con.c:13529) ==837133==by 0xB01B500: ssl3_GatherCompleteHandshake (ssl3gthr.c:56
Bug#1017426: libnss3: Uninitialised value was created by a stack allocation
On Tue, Aug 16, 2022 at 9:50 AM Mike Hommey wrote: > On Tue, Aug 16, 2022 at 09:06:20AM +0300, Alexey Kuznetsov wrote: > > On Tue, Aug 16, 2022 at 9:00 AM Mike Hommey wrote: > > > > > On Tue, Aug 16, 2022 at 08:30:07AM +0300, a...@me.com wrote: > > > > Package: libnss3 > > > > Version: 2:3.79-1 > > > > Severity: normal > > > > > > > > Dear Maintainer, > > > > > > > > debuging valgrind pidgin with result: > > > > > > > > ==804198== Uninitialised value was created by a stack allocation > > > > ==804198==at 0xB089DC0: ssl3_MACEncryptRecord (ssl3con.c:2104) > > > > > > > > line correspopnds to the ssl3_MACEncryptRecord > > > > > > Looking at the code, it would seem to be a false positive, but I might > > > have overlooked something, but you haven't pasted the most interesting > > > parts of the valgrind output... > > > > > > Mike > > > > > > > This output comes exactly from valgrind. No usual stack trace. Before and > > below are different issues. > > > > BTW pidgin crashing sometimes, and only issues I can record points to the > > nss library. > > Usually, "Uninitialised value was created by a stack allocation" is the > reason for the error, with a stack trace, that comes above it. That's > the most crucial information. Without that, we don't know what is trying > to use that unitialized value. > Ok .Let me restart pidgin. It 100% reproducible. Only thing you need is to install dbgsym for glibc, nss3, pidgin and add frew irc and jabber accounts (I also using matrix plugin). Command would be: G_SLICE=always-malloc valgrind --num-callers=30 --track-origins=yes pidgin 2>&1 | tee 123.log https://paste.debian.net/1250580/
Bug#1017426: libnss3: Uninitialised value was created by a stack allocation
On Tue, Aug 16, 2022 at 09:06:20AM +0300, Alexey Kuznetsov wrote: > On Tue, Aug 16, 2022 at 9:00 AM Mike Hommey wrote: > > > On Tue, Aug 16, 2022 at 08:30:07AM +0300, a...@me.com wrote: > > > Package: libnss3 > > > Version: 2:3.79-1 > > > Severity: normal > > > > > > Dear Maintainer, > > > > > > debuging valgrind pidgin with result: > > > > > > ==804198== Uninitialised value was created by a stack allocation > > > ==804198==at 0xB089DC0: ssl3_MACEncryptRecord (ssl3con.c:2104) > > > > > > line correspopnds to the ssl3_MACEncryptRecord > > > > Looking at the code, it would seem to be a false positive, but I might > > have overlooked something, but you haven't pasted the most interesting > > parts of the valgrind output... > > > > Mike > > > > This output comes exactly from valgrind. No usual stack trace. Before and > below are different issues. > > BTW pidgin crashing sometimes, and only issues I can record points to the > nss library. Usually, "Uninitialised value was created by a stack allocation" is the reason for the error, with a stack trace, that comes above it. That's the most crucial information. Without that, we don't know what is trying to use that unitialized value.
Bug#1017426: libnss3: Uninitialised value was created by a stack allocation
On Tue, Aug 16, 2022 at 9:00 AM Mike Hommey wrote: > On Tue, Aug 16, 2022 at 08:30:07AM +0300, a...@me.com wrote: > > Package: libnss3 > > Version: 2:3.79-1 > > Severity: normal > > > > Dear Maintainer, > > > > debuging valgrind pidgin with result: > > > > ==804198== Uninitialised value was created by a stack allocation > > ==804198==at 0xB089DC0: ssl3_MACEncryptRecord (ssl3con.c:2104) > > > > line correspopnds to the ssl3_MACEncryptRecord > > Looking at the code, it would seem to be a false positive, but I might > have overlooked something, but you haven't pasted the most interesting > parts of the valgrind output... > > Mike > This output comes exactly from valgrind. No usual stack trace. Before and below are different issues. BTW pidgin crashing sometimes, and only issues I can record points to the nss library.
Bug#1017426: libnss3: Uninitialised value was created by a stack allocation
On Tue, Aug 16, 2022 at 08:30:07AM +0300, a...@me.com wrote: > Package: libnss3 > Version: 2:3.79-1 > Severity: normal > > Dear Maintainer, > > debuging valgrind pidgin with result: > > ==804198== Uninitialised value was created by a stack allocation > ==804198==at 0xB089DC0: ssl3_MACEncryptRecord (ssl3con.c:2104) > > line correspopnds to the ssl3_MACEncryptRecord Looking at the code, it would seem to be a false positive, but I might have overlooked something, but you haven't pasted the most interesting parts of the valgrind output... Mike
